From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa01487; 1 Nov 2000 9:34 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id IAA08443; Wed, 1 Nov 2000 08:34:50 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTJH7T>; Wed, 1 Nov 2000 08:34:51 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B2C@umr-mail02> From: "Neulinger, Nathan R." To: "'openafs-info@penn.central.org'" Date: Wed, 1 Nov 2000 08:34:49 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] build status/testing... Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Just thought I'd pass this along - some things I needed to do, sucesses/failures/etc. i386_linux22 - Edited linux kernel version list in: src/config/Makefile.i386_linux22 works fine on 2.2.16 UP had problems with it on 2.2.17 SMP... afsd just starts sucking 99% cpu sun4x_57 - built ok, not tested sun4x_56 - build failed - gtx needed initscr32 and other curses symbols, which are not defined in libcurses. most of it built with make -k. Edited various Makefile.sun4x_56,Makefile.sun4x_57, */MakefileProto.SOLARIS files to point cc at our install of sunwspro cc (linked in afs) any reason not to just comment the CC lines in these? they should inherit from src/config/Makefile.sun4x_5x hp_ux102 - unsup, I'd imagine because of lack of pthreads hp_ux110 - doesn't build yet, src/rx/HPUX directory missing, copied the contents of src/rx/UKERNEL/ as hpux, building so far, no idea if it will actually work though do not use recent binutils ranlib, it doesn't work right sgi_65 - haven't tried yet ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa01707; 1 Nov 2000 9:59 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id IAA14908; Wed, 1 Nov 2000 08:59:34 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTJ2NS>; Wed, 1 Nov 2000 08:59:34 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B2D@umr-mail02> From: "Neulinger, Nathan R." To: "'openafs-info@penn.central.org'" , "'info-afs@transarc.com'" Date: Wed, 1 Nov 2000 08:59:33 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] changes to openafs to get hp_ux110 to build Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: in src/libafs/MakefileProto.HPUX #$(MAKE) BITSUFFIX=$$bsuff CPU_KDEFS="$$bopts" DESTDIR=../${DESTDIR} libafs || exit $$?; \ $(MAKE) BITSUFFIX=$$bsuff CPU_KDEFS="$$bopts" DESTDIR=${DESTDIR} libafs || exit $$?; \ The src/rx/HPUX directory is missing. Copying one of the other dirs works for a while, but fails later on with some redefinition warnings, and still has an error about osi_vfs.h missing. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa02273; 1 Nov 2000 11:40 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id KAA08428; Wed, 1 Nov 2000 10:40:06 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTJK5D>; Wed, 1 Nov 2000 10:40:06 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B34@umr-mail02> From: "Neulinger, Nathan R." To: "'info-afs@transarc.com'" , "'openafs-info@penn.central.org'" Date: Wed, 1 Nov 2000 10:40:04 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] openafs CML/NOCML - what's the plan as far as getting the mkvers stuff turned into something reasonable? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: At the very least, I think it'd be good to just change it to set the version to a compile timestamp - that should be a simple enough change. Ideally, it would be good to have it do something that would get the current CVS revision #, and perhaps all tags associated with tag revision number. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa03747; 1 Nov 2000 14:26 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id NAA17207; Wed, 1 Nov 2000 13:26:14 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTJ318>; Wed, 1 Nov 2000 13:26:14 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B3F@umr-mail02> From: "Neulinger, Nathan R." To: "'gar@arrakisplanet.com'" , "'openafs-info@central.org'" Date: Wed, 1 Nov 2000 13:26:13 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] RE: OpenAFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > -----Original Message----- > From: Greg Rhoades [mailto:gar@arrakisplanet.com] > Sent: Wednesday, November 01, 2000 1:21 PM > To: Neulinger, Nathan R. > Subject: RE: OpenAFS > > > >>> Neulinger, Nathan R writes: > NNR> I didn't need it... Not sure why he did... The build worked > NNR> just fine for me without washtool present. > > did you have any trouble with the solaris 5.x kernel modules? if i > don't use -k on the make, the build failes and it's always in the > MODLOAD section. i went ahead and completed the compilation with -k > but don't know yet what all i'm missing. Only trouble I had was on 5.6 the build failed due to some curses symbols in gtx. It worked just fine on 5.7. In fact, I'm running the client on 5.7 and one linux 2.2.16 right now without any trouble. > also, were you concerned at all about the afs version of fsck not > being made available? i know why, but don't know what exactly to do > about it. Well, yes and no... It means it most likely isn't suitable as a file server at the moment - although I wonder if maybe they changed it to operate like the linux version of 3.6? The linux version of the 3.6 fileserver actually creates files that are visible on the filesystem, hence no need for a scary fsck replacement. (At least that's how it looked to me.) -- Nathan From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa03900; 1 Nov 2000 14:37 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id OAA21497; Wed, 1 Nov 2000 14:38:23 -0500 (EST) Date: Wed, 01 Nov 2000 14:39:55 -0500 From: Derrick J Brashear To: "'gar@arrakisplanet.com'" , "'openafs-info@central.org'" Subject: Re: [OpenAFS] RE: OpenAFS Message-ID: <3483958112.973107595@skittlebrau.trafford.dementia.org> In-Reply-To: <9DA8D24B915BD1118911006094516EAF053E5B3F@umr-mail02> X-Mailer: Mulberry/2.0.5a7 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On 11/01/2000 01:26:13 PM -0600 "Neulinger, Nathan R." wrote: >> -----Original Message----- >> From: Greg Rhoades [mailto:gar@arrakisplanet.com] >> Sent: Wednesday, November 01, 2000 1:21 PM >> To: Neulinger, Nathan R. >> Subject: RE: OpenAFS >> >> >> >>> Neulinger, Nathan R writes: >> NNR> I didn't need it... Not sure why he did... The build worked >> NNR> just fine for me without washtool present. >> >> did you have any trouble with the solaris 5.x kernel modules? if i >> don't use -k on the make, the build failes and it's always in the >> MODLOAD section. i went ahead and completed the compilation with -k >> but don't know yet what all i'm missing. > > Only trouble I had was on 5.6 the build failed due to some curses symbols > in gtx. It worked just fine on 5.7. In fact, I'm running the client on > 5.7 and one linux 2.2.16 right now without any trouble. > >> also, were you concerned at all about the afs version of fsck not >> being made available? i know why, but don't know what exactly to do >> about it. > > Well, yes and no... It means it most likely isn't suitable as a file > server at the moment - although I wonder if maybe they changed it to > operate like the linux version of 3.6? The linux version of the 3.6 > fileserver actually creates files that are visible on the filesystem, > hence no need for a scary fsck replacement. (At least that's how it > looked to me.) bsd vfsck works just fine on solaris machines; I use it on mine that are running "Transarc AFS". The only problem is exit statuses don't necessarily match what the rc scripts expect. I assume a near-term patch will address that, but 35 production fileservers in the andrew.cmu.edu cell all run the bsd vfsck and are all solaris machines. (the bsd vfsck is provided; only the modified vendor fscks are not) -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa04136; 1 Nov 2000 14:47 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id OAA21561; Wed, 1 Nov 2000 14:48:35 -0500 (EST) Date: Wed, 01 Nov 2000 14:50:08 -0500 From: Derrick J Brashear To: Greg Rhoades cc: "'openafs-info@central.org'" Subject: Re: [OpenAFS] RE: OpenAFS Message-ID: <3550398112.973108208@skittlebrau.trafford.dementia.org> In-Reply-To: <14848.29201.857096.981917@melange.arrakisplanet.com> X-Mailer: Mulberry/2.0.5a7 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On 11/01/2000 02:42:09 PM -0500 Greg Rhoades wrote: >>>> Derrick J Brashear writes: > > DJB> bsd vfsck works just fine on solaris machines; I use it on mine > DJB> that are running "Transarc AFS". The only problem is exit > DJB> statuses don't necessarily match what the rc scripts expect. I > DJB> assume a near-term patch will address that, but 35 production > DJB> fileservers in the andrew.cmu.edu cell all run the bsd vfsck > DJB> and are all solaris machines. > > DJB> (the bsd vfsck is provided; only the modified vendor fscks are > DJB> not) > > thanks for the info. have you tried to build the solaris 5.6 version > of the software? did you have any trouble with the kernel modules > (specifically the absence the file called afs_nfsdisp.c)? Thus far I have not, but given that the AFS-NFS translator is "known" to be not included, I assume if it doesn't build changes will need to be made to accomodate that. -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.arrakisplanet.com by penn.central.org id aa04027; 1 Nov 2000 14:42 EST Received: by sunfw1.arrakisplanet.com; id OAA07439; Wed, 1 Nov 2000 14:42:50 -0500 (EST) Received: from exchange.arrakisplanet.com(169.121.23.26) by sunfw1.arrakisplanet.com via smap (V4.2) id xma007435; Wed, 1 Nov 00 14:42:10 -0500 Received: from melange.arrakisplanet.com ([169.121.22.101]) by cam-1cc-mail01.arrakisplanet.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2650.21) id VTF7SJKG; Wed, 1 Nov 2000 14:42:06 -0500 Received: (from gar@localhost) by melange.arrakisplanet.com (8.9.3+Sun/8.9.3) id OAA11692; Wed, 1 Nov 2000 14:42:10 -0500 (EST) From: Greg Rhoades MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <14848.29201.857096.981917@melange.arrakisplanet.com> Date: Wed, 1 Nov 2000 14:42:09 -0500 (EST) To: Derrick J Brashear Cc: "'gar@arrakisplanet.com'" , "'openafs-info@central.org'" Subject: Re: [OpenAFS] RE: OpenAFS In-Reply-To: <3483958112.973107595@skittlebrau.trafford.dementia.org> References: <9DA8D24B915BD1118911006094516EAF053E5B3F@umr-mail02> <3483958112.973107595@skittlebrau.trafford.dementia.org> X-Mailer: VM 6.71 under 21.1 (patch 4) "Arches" XEmacs Lucid Reply-To: gar@arrakisplanet.com (Greg Rhoades) X-Attribution: gar Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>> Derrick J Brashear writes: DJB> bsd vfsck works just fine on solaris machines; I use it on mine DJB> that are running "Transarc AFS". The only problem is exit DJB> statuses don't necessarily match what the rc scripts expect. I DJB> assume a near-term patch will address that, but 35 production DJB> fileservers in the andrew.cmu.edu cell all run the bsd vfsck DJB> and are all solaris machines. DJB> (the bsd vfsck is provided; only the modified vendor fscks are DJB> not) thanks for the info. have you tried to build the solaris 5.6 version of the software? did you have any trouble with the kernel modules (specifically the absence the file called afs_nfsdisp.c)? thanks, -greg -- --------------------------------------------------------------------------- Greg Rhoades Email:gar@arrakisplanet.com Arrakis Phone: (617) 768-3219 One Cambridge Center Fax: (617) 768-3001 Cambridge, MA 02142 --------------------------------------------------------------------------- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SCULLY.WV.CC.CMU.EDU by penn.central.org id aa15582; 2 Nov 2000 15:25 EST Received: from localhost (root@localhost) by scully.wv.cc.cmu.edu (8.11.0/8.11.0) with SMTP id eA2KPte02133; Thu, 2 Nov 2000 15:25:55 -0500 Date: Thu, 2 Nov 2000 15:25:55 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@scully.wv.cc.cmu.edu To: openafs-info@central.org cc: openafs-devel@central.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] patches and enhancements (and bug reports) solicited Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: If you have patches or enhancements to the code, please submit them to openafs-gatekeepers@central.org for inclusion; Please (unless they're huge) also copy openafs-devel@central.org so others can see and comment on your code. Two groups have started working on porting OpenAFS to new platforms, namely FreeBSD and NetBSD. Those groups have mailing lists devoted to them, namely port-netbsd and port-freebsd, both at central.org until the openafs.org domain name correctly points to the machine in question. We are now waiting on the registrar for openafs.org to process the submitted change. If there is additional interest in working on new ports, more lists can be created. Thanks -- Derrick J Brashear From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa15759; 2 Nov 2000 15:34 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id OAA31128; Thu, 2 Nov 2000 14:34:15 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTKAJ1>; Thu, 2 Nov 2000 14:34:15 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B4A@umr-mail02> From: "Neulinger, Nathan R." To: "'Derrick J Brashear'" , openafs-info@central.org Cc: openafs-devel@central.org Date: Thu, 2 Nov 2000 14:34:14 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] RE: [OpenAFS-devel] patches and enhancements (and bug reports) so licited Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > If there is additional interest in working on new ports, more > lists can be > created. Do you have any knowledge of what the status of Win2k is? If that isn't going to be forthcoming from transarc, it'd probably be a worthwhile list to have, presuming that the port isn't easy, if it is, probably not. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SCULLY.WV.CC.CMU.EDU by penn.central.org id aa16185; 2 Nov 2000 16:04 EST Received: from localhost (root@localhost) by scully.wv.cc.cmu.edu (8.11.0/8.11.0) with SMTP id eA2L4tl02188; Thu, 2 Nov 2000 16:04:55 -0500 Date: Thu, 2 Nov 2000 16:04:55 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@scully.wv.cc.cmu.edu To: openafs-info@central.org cc: info-afs@transarc.com Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] seeking CellServDB entry submissions for OpenAFS inclusion Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: As IBM does not wish to have a the CellServDB distributed with OpenAFS seeded from the one Transarc distributes, I volunteered to compile a new CellServDB to be distributed with OpenAFS.(*) Please send any submissions of CellServDB entries for organizations which you are affiliated and authorized to submit to cellservdb@central.org. Please include a contact email address so that if your cell has "apparently vanished" at some point in the future you can be contacted for updates. Please do not submit entries for other organizations. While having a cell merely unpublished does not protect the cell from being discovered, this is intended to be a voluntary process for people who wish to advertise their cell to the users of OpenAFS. If any cell you submit is turned off or moved please also submit an update or delete request. Thanks -- Derrick J Brashear * In the future OpenAFS presumably have DNS AFSDB RR support enabling more dynamic configuration of the client, but in the meantime having a useful CellServDB to distribute seemed like a good idea. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ginger.cmf.nrl.navy.mil by penn.central.org id aa22442; 3 Nov 2000 10:51 EST Received: from borg.cmf.nrl.navy.mil (borg.cmf.nrl.navy.mil [134.207.10.148]) by ginger.cmf.nrl.navy.mil (8.10.1/8.10.1) with ESMTP id eA3Fp9H29517; Fri, 3 Nov 2000 10:51:10 -0500 (EST) From: Chas Williams Received: (from chas@localhost) by borg.cmf.nrl.navy.mil (8.8.5/8.8.5) id KAA131316; Fri, 3 Nov 2000 10:51:09 -0500 (EST) Date: Fri, 3 Nov 2000 10:51:09 -0500 (EST) Message-Id: <200011031551.KAA131316@borg.cmf.nrl.navy.mil> To: openafs-info@WWW-OPENAFS.CENTRAL.ORG, info-afs@transarc.com Subject: [OpenAFS] update available to linux 2.4 patch for openafs Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: i updated symlink handling (since it was broken and only worked with relative links really) ftp://ftp.cmf.nrl.navy.mil/pub/chas/linux-openafs/symlink-fix.diff [you should apply this after applying afs.diffs] From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa05569; 5 Nov 2000 14:15 EST Received: (qmail 4862 invoked from network); 5 Nov 2000 19:15:23 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 5 Nov 2000 19:15:23 -0000 Message-ID: <3A05B1CB.ABAA7E64@umr.edu> Date: Sun, 05 Nov 2000 13:15:23 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.18pre17 i686) X-Accept-Language: en MIME-Version: 1.0 To: openafs-info@central.org Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Are the patches in the 3.6.2.5 release included in openafs? or will they be? In particular, I'm referring to the changes in i386_nt40 for various bugs, and the support of win2k. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from VPN-100-230.CC.CMU.EDU by penn.central.org id aa05654; 5 Nov 2000 14:55 EST Received: from localhost (root@localhost) by scully.wv.cc.cmu.edu (8.11.0/8.11.0) with SMTP id eA5Jtfl07732 for ; Sun, 5 Nov 2000 14:55:42 -0500 Date: Sun, 5 Nov 2000 14:55:41 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@scully.wv.cc.cmu.edu cc: openafs-info@central.org Subject: Re: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release In-Reply-To: <3A05B1CB.ABAA7E64@umr.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Sun, 5 Nov 2000, Nathan Neulinger wrote: > Are the patches in the 3.6.2.5 release included in openafs? or will they > be? In particular, I'm referring to the changes in i386_nt40 for various > bugs, and the support of win2k. Presumably not, and not any time soon. The problem is the amount of time to pass things past attorneys to clear release is apparently mind-boggling, with the attendant price tag for people's time. I can't, and won't, speak for IBM, but all indications were that nothing would soon be being released. -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa06028; 5 Nov 2000 15:48 EST Received: (qmail 24118 invoked from network); 5 Nov 2000 20:48:51 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 5 Nov 2000 20:48:51 -0000 Message-ID: <3A05C7B3.2FF29CB4@umr.edu> Date: Sun, 05 Nov 2000 14:48:51 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.18pre17 i686) X-Accept-Language: en MIME-Version: 1.0 To: Derrick J Brashear CC: openafs-info@central.org Subject: Re: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Derrick J Brashear wrote: > > On Sun, 5 Nov 2000, Nathan Neulinger wrote: > > > Are the patches in the 3.6.2.5 release included in openafs? or will they > > be? In particular, I'm referring to the changes in i386_nt40 for various > > bugs, and the support of win2k. > > Presumably not, and not any time soon. The problem is the amount of time > to pass things past attorneys to clear release is apparently > mind-boggling, with the attendant price tag for people's time. > > I can't, and won't, speak for IBM, but all indications were that nothing > would soon be being released. > > -D > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info Bummer. Let's home someone that signs onto the list is an NT/Win2k expert and duplicates ibm's effort. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dingo.sta.cathedral.org by penn.central.org id aa06126; 5 Nov 2000 16:49 EST Received: by dingo.sta.cathedral.org (STA/mail, from userid 500) id 0BD13206188; Sun, 5 Nov 2000 16:48:58 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by dingo.sta.cathedral.org (STA/mail) with ESMTP id 0252B20614F for ; Sun, 5 Nov 2000 16:48:57 -0500 (EST) Date: Sun, 5 Nov 2000 16:48:57 -0500 (EST) From: Michael Nelson X-Sender: mikenel@dingo.sta.cathedral.org To: openafs-info@lists-openafs.central.org Subject: Re: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Return-Path: Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I've gotten OpenAFS to build on Win2K -- and just like on Linux (for me at least), certain files weren't copied into the right places, and the Makefiles needed some minor tweaks. I've narrowed down roughly why OpenAFS isn't working, but I haven't figured out how to fix it (yet). Rather than write an NT file system driver, Transarc implemented a mini SMB (Microsoft's file sharing protocol) implementation that runs locally. Their SMB server registers itself as if it were another fileserver on a fictitious machine called "-afs". It acts as a gateway to AFS and also provides a way for the various tools (e.g. fs) to send ioctl's to the AFS cache manager. All of this is done via the NetBIOS system (yuck). From another computer I can actually do a "dir \\\all" (all is a special internal share used for AFS IOCTL's) and get an error message that no connections are allowed (by default AFS doesn't allow remote machines to connect to an AFS cache manager's SMB server). But, I can't connect to it locally (using dir). I've fiddled with LANA IDs (in NetBIOS, you send/receive datagrams on a specific LAN adapter), scoured the Microsoft Knowledge Base, but haven't come up with anything interesting. I've been thinking about installing Win2k SP1, but I'm having modem problems today. Because I don't have the commercial version of AFS or the InstallShield kit, I had to (painstakingly) figure out which registry settings and files had to be created. The cache manager does make contact with my cell when starting, and klog does work, so I am pretty confident that I have configure the (minimal) settings. -mike From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa06167; 5 Nov 2000 17:20 EST Received: (qmail 11163 invoked from network); 5 Nov 2000 22:20:10 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 5 Nov 2000 22:20:10 -0000 Message-ID: <3A05DD1A.50E590A4@umr.edu> Date: Sun, 05 Nov 2000 16:20:10 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.18pre17 i686) X-Accept-Language: en MIME-Version: 1.0 To: Michael Nelson CC: openafs-info@WWW-OPENAFS.central.org Subject: Re: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Interesting... probably just minor bugs at this point. For reference, on the commercial one, the only registry keys that I saw it install were in software/transarccorp/afs client/current version/3.6, or 'afs server'. Although we haven't looked around much. As for the share names, I wasn't aware of 'all', but the other ones should be configured in C:\WINNT\afsdsbmt.ini. There is also a toolbar/systray thing that should allow you to configure mounts, klog, etc. There are a couple other files like that as well. Once you have a mount configured in that file, you should be able to attached \\hostname-afs\submountname to a drive without any user/password, and it will attach it. Which files didn't copy for you on linux? I haven't had any trouble. -- Nathan Michael Nelson wrote: > > I've gotten OpenAFS to build on Win2K -- and just like on Linux (for me at > least), certain files weren't copied into the right places, and the > Makefiles needed some minor tweaks. I've narrowed down roughly why > OpenAFS isn't working, but I haven't figured out how to fix it (yet). > > Rather than write an NT file system driver, Transarc implemented a mini > SMB (Microsoft's file sharing protocol) implementation that runs locally. > Their SMB server registers itself as if it were another fileserver on a > fictitious machine called "-afs". It acts as a gateway to > AFS and also provides a way for the various tools (e.g. fs) to send > ioctl's to the AFS cache manager. All of this is done via the NetBIOS > system (yuck). > > From another computer I can actually do a "dir \\\all" (all > is a special internal share used for AFS IOCTL's) and get an error message > that no connections are allowed (by default AFS doesn't allow remote > machines to connect to an AFS cache manager's SMB server). But, I can't > connect to it locally (using dir). > > I've fiddled with LANA IDs (in NetBIOS, you send/receive datagrams on a > specific LAN adapter), scoured the Microsoft Knowledge Base, but haven't > come up with anything interesting. I've been thinking about installing > Win2k SP1, but I'm having modem problems today. > > Because I don't have the commercial version of AFS or the InstallShield > kit, I had to (painstakingly) figure out which registry settings and files > had to be created. The cache manager does make contact with my cell when > starting, and klog does work, so I am pretty confident that I have > configure the (minimal) settings. > > -mike > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dingo.sta.cathedral.org by penn.central.org id aa06189; 5 Nov 2000 17:37 EST Received: by dingo.sta.cathedral.org (STA/mail, from userid 500) id 475EB206188; Sun, 5 Nov 2000 17:37:20 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by dingo.sta.cathedral.org (STA/mail) with ESMTP id 41ED320614F; Sun, 5 Nov 2000 17:37:19 -0500 (EST) Date: Sun, 5 Nov 2000 17:37:19 -0500 (EST) From: Michael Nelson X-Sender: mikenel@dingo.sta.cathedral.org To: Nathan Neulinger Cc: openafs-info@WWW-OPENAFS.central.org Subject: Re: [OpenAFS] openafs and win2k? wrt transarc 3.6.2.5 release In-Reply-To: <3A05DD1A.50E590A4@umr.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Return-Path: Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > Interesting... probably just minor bugs at this point. Okay, I've got it working now -- sort of. I installed NetBEUI and then using a program I wrote that dumps out the valid LANA IDs, tried 6 (which comes after 0) to see what would happen, and it works. Without NetBEUI, none of the LANA IDs work. > For reference, on the commercial one, the only registry keys that I saw > it install were in software/transarccorp/afs client/current version/3.6, > or 'afs server'. Although we haven't looked around much. There is stuff in HKLM\CurrentControlSet\Control\Services\TransarcAFSDaemon too. > As for the share names, I wasn't aware of 'all', but the other ones > should be configured in C:\WINNT\afsdsbmt.ini. There is also a > toolbar/systray thing that should allow you to configure mounts, klog, > etc. There are a couple other files like that as well. Once you have a > mount configured in that file, you should be able to attached > \\hostname-afs\submountname to a drive without any user/password, and > it will attach it. All appears to point to the root.afs volume. I was fooling with afscreds (the tray tool you refer to) before, but until I did the above, nothing was working. I was picking on "all" because it is built-in and would work if the basic functionality was working. Now it does. > Which files didn't copy for you on linux? I haven't had any trouble. My kernel's version # isn't one of the standard ones (I've been doing stuff to it and wanted to distinguish it from the baseline 2.2.14) and I think that confused the Makefiles. Some of the include file links (don't remember the path) didn't get setup properly (they pointed to /usr/src/linux-2.2.5-15 instead of the right place), and I seem to remember one of the template makefiles not getting copied properly. -mike From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa08909; 6 Nov 2000 13:36 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id MAA18813; Mon, 6 Nov 2000 12:36:28 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTL1YN>; Mon, 6 Nov 2000 12:36:27 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B6C@umr-mail02> From: "Neulinger, Nathan R." To: "'openafs-info@central.org'" Date: Mon, 6 Nov 2000 12:36:26 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] status report - openafs i386_linux22 server Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: It seems to work just fine. In fact, have it running against reiserfs as well and it seemed to handle that just fine. We're not actually planning on a server deployment with openafs just yet, but it's nice to see it working pretty well out of the box. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa08392; 7 Nov 2000 13:14 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id MAA10770; Tue, 7 Nov 2000 12:13:45 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id <4NTTL4S3>; Tue, 7 Nov 2000 12:13:45 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B77@umr-mail02> From: "Neulinger, Nathan R." To: "'Brian Kosick'" , info-afs@transarc.com Cc: "'openafs-info@central.org'" Date: Tue, 7 Nov 2000 12:13:44 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] RE: compiling on linux 2.2.x Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: This should be on openafs-devel@central.org or openafs-info@central.org. To fix, edit the file src/config/Makefile.i386_linux22 and change the kernel line to only list 2.2.14. Make sure that you have a link to the kernel sources as /usr/src/linux-2.2.14. Then, run the make again. -- Nathan > -----Original Message----- > From: Brian Kosick [mailto:briank@nacs.net] > Sent: Tuesday, November 07, 2000 11:35 AM > To: info-afs@transarc.com > Subject: compiling on linux 2.2.x > > > Hello all, forgive me if this is the wrong place to be > posting this, I've > browsed over the archives, and have really not found anything on > openafs-1.0.0 compile problems. I have two boxes that I have > tried on, > 1) Intel P200MMX 128MB RAM 10G IDE HD running fresh install of RH7.0 > kernel 2.2.16-22 > 2) AMD k6-2 500 128MB RAM 4G IDE, 6G IDE, 2 x 3G SCSI Raid 0 > running RH6.2 > kernel 2.2.15 > > I have downloaded, and compiled kernel 2.2.14 on both, so I > could get some > valid headers for open afs, the kernel is compiled, but NOT running. > > I start the compile and everything seems to work > beautifully...then it > starts horking out errors. (I've included the problem output > from the > compiler at the end of this email) > > If I've left any info out that would be of help, let me know, > I will gladly > provide it. > > Compile errors: > In file included from ../afs/afs_sysnames.h:17, > from ../afs/param.h:24, > from ../afs/afs_analyze.c:11: > ../afs/stds.h:152:25: ../h/string.h: No such file or directory > In file included from /usr/include/linux/affs_fs_i.h:5, > from /usr/include/linux/fs.h:276, > from /usr/include/linux/capability.h:17, > from /usr/include/linux/binfmts.h:5, > from /usr/include/linux/sched.h:9, > from /usr/include/asm/uaccess.h:8, > from ../afs/sysincludes.h:26, > from ../afs/afs_analyze.c:13: > /usr/include/linux/time.h:69:15: warning: "FD_SET" redefined > /usr/include/sys/select.h:63:1: warning: this is the location of the > previous definition > /usr/include/linux/time.h:70:15: warning: "FD_CLR" redefined > /usr/include/sys/select.h:64:1: warning: this is the location of the > previous definition > /usr/include/linux/time.h:71:17: warning: "FD_ISSET" redefined > /usr/include/sys/select.h:65:1: warning: this is the location of the > previous definition > /usr/include/linux/time.h:72:16: warning: "FD_ZERO" redefined > /usr/include/sys/select.h:66:1: warning: this is the location of the > previous definition > In file included from /usr/include/linux/affs_fs_i.h:6, > from /usr/include/linux/fs.h:277, > from /usr/include/linux/capability.h:18, > from /usr/include/linux/binfmts.h:6, > from /usr/include/linux/sched.h:10, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/time.h:9: redefinition of `struct timespec' > /usr/include/linux/time.h:51: parse error before `suseconds_t' > /usr/include/linux/time.h:51: warning: no semicolon at end of > struct or union > /usr/include/linux/time.h:88: field `it_interval' has incomplete type > /usr/include/linux/time.h:89: field `it_value' has incomplete type > In file included from /usr/include/linux/fs.h:277, > from /usr/include/linux/capability.h:18, > from /usr/include/linux/binfmts.h:6, > from /usr/include/linux/sched.h:10, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/affs_fs_i.h:11: field `kc_lru_time' has > incomplete type > In file included from /usr/include/linux/fs.h:279, > from /usr/include/linux/capability.h:18, > from /usr/include/linux/binfmts.h:6, > from /usr/include/linux/sched.h:10, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/efs_fs_i.h:13: parse error before `efs_ino_t' > /usr/include/linux/efs_fs_i.h:13: warning: data definition > has no type or > storage class > /usr/include/linux/efs_fs_i.h:48: parse error before `uint32_t' > /usr/include/linux/efs_fs_i.h:48: warning: no semicolon at > end of struct or > union > /usr/include/linux/efs_fs_i.h:56: parse error before `}' > In file included from /usr/include/linux/sched.h:15, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/times.h:5: parse error before `clock_t' > /usr/include/linux/times.h:5: warning: no semicolon at end of > struct or union > /usr/include/linux/times.h:6: warning: data definition has no type or > storage class > /usr/include/linux/times.h:7: parse error before `tms_cutime' > /usr/include/linux/times.h:7: warning: data definition has no type or > storage class > /usr/include/linux/times.h:8: parse error before `tms_cstime' > /usr/include/linux/times.h:8: warning: data definition has no type or > storage class > In file included from /usr/include/linux/sched.h:16, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/timex.h:159: field `time' has incomplete type > In file included from /usr/include/linux/signal.h:6, > from /usr/include/linux/sched.h:26, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/asm/siginfo.h:48: parse error before `clock_t' > /usr/include/asm/siginfo.h:48: warning: no semicolon at end > of struct or union > /usr/include/asm/siginfo.h:48: warning: no semicolon at end > of struct or union > /usr/include/asm/siginfo.h:49: warning: no semicolon at end > of struct or union > /usr/include/asm/siginfo.h:50: warning: data definition has > no type or > storage class > /usr/include/asm/siginfo.h:62: parse error before `}' > /usr/include/asm/siginfo.h:62: warning: data definition has > no type or > storage class > /usr/include/asm/siginfo.h:63: parse error before `}' > /usr/include/asm/siginfo.h:63: warning: data definition has > no type or > storage class > /usr/include/asm/siginfo.h:219: parse error before `*' > /usr/include/asm/siginfo.h: In function `copy_siginfo': > /usr/include/asm/siginfo.h:221: `from' undeclared (first use > in this function) > /usr/include/asm/siginfo.h:221: (Each undeclared identifier > is reported > only once > /usr/include/asm/siginfo.h:221: for each function it appears in.) > /usr/include/asm/siginfo.h:222: `to' undeclared (first use in > this function) > /usr/include/asm/siginfo.h: At top level: > /usr/include/asm/siginfo.h:228: parse error before `*' > In file included from /usr/include/linux/sched.h:26, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/signal.h:15: parse error before `siginfo_t' > /usr/include/linux/signal.h:15: warning: no semicolon at end > of struct or union > In file included from /usr/include/linux/sched.h:76, > from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/resource.h:22: field `ru_utime' has incomplete type > /usr/include/linux/resource.h:23: field `ru_stime' has incomplete type > In file included from /usr/include/asm/uaccess.h:9, > from ../afs/sysincludes.h:27, > from ../afs/afs_analyze.c:14: > /usr/include/linux/sched.h:333: field `times' has incomplete type > /usr/include/linux/sched.h:552: parse error before `siginfo_t' > In file included from ../afs/afsincludes.h:23, > from ../afs/afs_analyze.c:29: > ../afs/afs_osi.h:11:24: ../h/types.h: No such file or directory > ../afs/afs_osi.h:12:24: ../h/param.h: No such file or directory > ../afs/afs_osi.h:20:21: ../h/fs.h: No such file or directory > ../afs/afs_osi.h:21:21: ../h/mm.h: No such file or directory > In file included from ../afs/afs_osi.h:370, > from ../afs/afsincludes.h:23, > from ../afs/afs_analyze.c:29: > ../afs/osi_machdep.h:25:24: ../h/sched.h: No such file or directory > In file included from ../rx/rx_packet.h:35, > from ../rx/rx.h:50, > from ../afs/afsincludes.h:24, > from ../afs/afs_analyze.c:29: > /usr/include/sys/param.h:58:12: warning: "MIN" redefined > ../afs/param.h:60:1: warning: this is the location of the > previous definition > /usr/include/sys/param.h:59:12: warning: "MAX" redefined > ../afs/param.h:63:1: warning: this is the location of the > previous definition > In file included from ../afs/afsincludes.h:24, > from ../afs/afs_analyze.c:29: > ../rx/rx.h:52:27: ../netinet/in.h: No such file or directory > In file included from ../afs/afsincludes.h:22, > from ../afs/afs_analyze.c:30: > ../afs/osi_vfs.h:144: field `va_atime' has incomplete type > ../afs/osi_vfs.h:145: field `va_mtime' has incomplete type > ../afs/osi_vfs.h:146: field `va_ctime' has incomplete type > In file included from ../rx/rx.h:46, > from ../afs/afsincludes.h:25, > from ../afs/afs_analyze.c:30: > ../rx/rx_kmutex.h: In function `afs_mutex_init': > ../rx/rx_kmutex.h:45: `MUTEX' undeclared (first use in this function) > ../rx/rx_kmutex.h: In function `CV_WAIT': > ../rx/rx_kmutex.h:94: warning: passing arg 1 of > `interruptible_sleep_on' > from incompatible pointer type > ../rx/rx_kmutex.h: In function `CV_TIMEDWAIT': > ../rx/rx_kmutex.h:110: warning: passing arg 1 of > `interruptible_sleep_on_timeout' from incompatible pointer type > In file included from /usr/include/sys/uio.h:30, > from ../rx/rx_packet.h:8, > from ../rx/rx.h:51, > from ../afs/afsincludes.h:25, > from ../afs/afs_analyze.c:30: > /usr/include/bits/uio.h: At top level: > /usr/include/bits/uio.h:43: redefinition of `struct iovec' > In file included from ../afs/afsincludes.h:38, > from ../afs/afs_analyze.c:29: > ../afsint/afsint.h:15:24: ../h/types.h: Bad file descriptor > ../afsint/afsint.h:31:27: ../netinet/in.h: Bad file descriptor > ../afsint/afsint.h:34:23: ../h/time.h: No such file or directory > In file included from ../afs/afsincludes.h:26, > from ../afs/afs_analyze.c:30: > ../afs/lock.h:117: field `time_waiting' has incomplete type > In file included from ../afs/afsincludes.h:40, > from ../afs/afs_analyze.c:29: > ../afsint/vldbint.h:15:24: ../h/types.h: No such file or directory > ../afsint/vldbint.h:31:27: ../netinet/in.h: No such file or directory > ../afsint/vldbint.h:34:23: ../h/time.h: No such file or directory > In file included from ../afs/afsincludes.h:42, > from ../afs/afs_analyze.c:30: > ../afs/afs.h:356: field `host' has incomplete type > In file included from ../afs/afs_analyze.c:31: > ../afs/afs_stats.h:859: field `sumTime' has incomplete type > ../afs/afs_stats.h:860: field `sqrTime' has incomplete type > ../afs/afs_stats.h:861: field `minTime' has incomplete type > ../afs/afs_stats.h:862: field `maxTime' has incomplete type > ../afs/afs_stats.h:885: field `sumTime' has incomplete type > ../afs/afs_stats.h:886: field `sqrTime' has incomplete type > ../afs/afs_stats.h:887: field `minTime' has incomplete type > ../afs/afs_stats.h:888: field `maxTime' has incomplete type > ../afs/afs_analyze.c: In function `afs_Analyze': > ../afs/afs_analyze.c:334: storage size of `opStartTime' isn't known > ../afs/afs_analyze.c:334: storage size of `opStopTime' isn't known > ../afs/afs_analyze.c:334: storage size of `elapsedTime' isn't known > make[4]: *** [afs_analyze.o] Error 1 > make[4]: Leaving directory > `/usr/local/afs/i386_linux22/obj/libafs/MODLOAD-2.2.5-15-MP' > make[3]: *** [linux_compdirs] Error 2 > make[3]: Leaving directory `/usr/local/afs/i386_linux22/obj/libafs' > make[2]: *** [libafs] Error 2 > make[2]: Leaving directory `/usr/local/afs' > make[1]: *** [install] Error 2 > make[1]: Leaving directory `/usr/local/afs' > make: *** [all] Error 2 > Brian Kosick > Web Programmer > New Age Consulting Service, Inc. > 216-619-2000 > briank@nacs.net > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from devel.nacs.net by penn.central.org id aa10823; 7 Nov 2000 18:43 EST Received: from nine.nacs.net (tawny.nacs.net [207.166.192.127]) by devel.nacs.net (8.9.3/8.9.3) with ESMTP id PAA15856 for ; Tue, 7 Nov 2000 15:18:42 -0500 Message-Id: <4.3.1.2.20001107150431.00b44940@mail.nacs.net> X-Sender: bkosick@mail.nacs.net X-Mailer: QUALCOMM Windows Eudora Version 4.3.1 Date: Tue, 07 Nov 2000 15:14:49 -0500 To: openafs-info@central.org From: Brian Kosick Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Subject: [OpenAFS] apenafs 1.0.0 linux compile problems Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hello all, I'm trying to compile open afs on two linux boxes: 1) Intel P200MMX 128MB RAM 10G IDE HD running fresh install of RH7.0 kernel 2.2.16-22 2) AMD k6-2 500 128MB RAM 4G IDE, 6G IDE, 2 x 3G SCSI Raid 0 running RH6.2 kernel 2.2.15 I have downloaded, and compiled kernel 2.2.14 on both, so I could get some valid headers for open afs, the kernel is compiled, but NOT running. I've gotten past a few other small obsticles, and am now stuck these errror messages for the rh6.2 box ------------------------------------------------------------------------------------------------------ make[3]: Entering directory `/usr/local/afs/i386_linux22/obj/gtx' cc -o gtxtest gtxtest.o libgtx.a /usr/local/afs/i386_linux22/dest/lib/librx.a /usr/local/afs/i386_linux22/dest/lib/liblwp.a /usr/local/afs/i386_linux22/dest/lib/afs/libsys.a /usr/local/afs/i386_linux22/dest/lib/afs/libcmd.a /usr/local/afs/i386_linux22/dest/lib/afs/libkauth.a /usr/local/afs/i386_linux22/dest/lib/afs/libauth.a /usr/local/afs/i386_linux22/dest/lib/librxkad.a /usr/local/afs/i386_linux22/dest/lib/libdes.a /usr/local/afs/i386_linux22/dest/lib/afs/libcom_err.a /usr/local/afs/i386_linux22/dest/lib/afs/util.a /usr/lib/libncurses.so /usr/local/afs/i386_linux22/dest/lib/afs/libtermlib.a cc: /usr/lib/libncurses.so: No such file or directory make[3]: *** [gtxtest] Error 1 make[3]: Leaving directory `/usr/local/afs/i386_linux22/obj/gtx' make[2]: *** [gtx] Error 2 make[2]: Leaving directory `/usr/local/afs' make[1]: *** [install] Error 2 make[1]: Leaving directory `/usr/local/afs' make: *** [all] Error 2 --------------------------------------------------------------------------- /usr/lib/libncurses.so IS installed ls /usr/lib/libncurses. libncurses.a libncurses.so libncurses.so.5 libncurses.so.5.0 ============================================================= and these errors for box 2 (rh7.0) ============================================================= --------------------------------------------------------------------------------------------------------------------- In file included from ../afs/sysincludes.h:125, from ../afs/afs_analyze.c:14: /usr/include/pthread.h:270: parse error before `0' make[4]: *** [../UAFS/afs_analyze.o] Error 1 make[4]: Leaving directory `/usr/local/afs/i386_linux22/obj/libuafs/UAFS' make[3]: *** [UAFS/libuafs.a] Error 2 make[3]: Leaving directory `/usr/local/afs/i386_linux22/obj/libuafs' make[2]: *** [libuafs] Error 2 make[2]: Leaving directory `/usr/local/afs' make[1]: *** [install] Error 2 make[1]: Leaving directory `/usr/local/afs' make: *** [all] Error 2 ------------------------------------------------------------------------------ Any ideas? I'm really looking foward to trying afs, I've heard many Good Things about it. Thanks in advance Brian Kosick From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from devel.nacs.net by penn.central.org id aa10824; 7 Nov 2000 18:43 EST Received: from nine.nacs.net (tawny.nacs.net [207.166.192.127]) by devel.nacs.net (8.9.3/8.9.3) with ESMTP id PAA16361 for ; Tue, 7 Nov 2000 15:22:57 -0500 Message-Id: <4.3.1.2.20001107151859.00c5a320@mail.nacs.net> X-Sender: bkosick@mail.nacs.net X-Mailer: QUALCOMM Windows Eudora Version 4.3.1 Date: Tue, 07 Nov 2000 15:19:04 -0500 To: openafs-info@central.org From: Brian Kosick Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Subject: [OpenAFS] apenafs 1.0.0 linux compile problems Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hello all, I'm trying to compile open afs on two linux boxes: 1) Intel P200MMX 128MB RAM 10G IDE HD running fresh install of RH7.0 kernel 2.2.16-22 2) AMD k6-2 500 128MB RAM 4G IDE, 6G IDE, 2 x 3G SCSI Raid 0 running RH6.2 kernel 2.2.15 I have downloaded, and compiled kernel 2.2.14 on both, so I could get some valid headers for open afs, the kernel is compiled, but NOT running. I've gotten past a few other small obsticles, and am now stuck these errror messages for the rh6.2 box ------------------------------------------------------------------------------------------------------ make[3]: Entering directory `/usr/local/afs/i386_linux22/obj/gtx' cc -o gtxtest gtxtest.o libgtx.a /usr/local/afs/i386_linux22/dest/lib/librx.a /usr/local/afs/i386_linux22/dest/lib/liblwp.a /usr/local/afs/i386_linux22/dest/lib/afs/libsys.a /usr/local/afs/i386_linux22/dest/lib/afs/libcmd.a /usr/local/afs/i386_linux22/dest/lib/afs/libkauth.a /usr/local/afs/i386_linux22/dest/lib/afs/libauth.a /usr/local/afs/i386_linux22/dest/lib/librxkad.a /usr/local/afs/i386_linux22/dest/lib/libdes.a /usr/local/afs/i386_linux22/dest/lib/afs/libcom_err.a /usr/local/afs/i386_linux22/dest/lib/afs/util.a /usr/lib/libncurses.so /usr/local/afs/i386_linux22/dest/lib/afs/libtermlib.a cc: /usr/lib/libncurses.so: No such file or directory make[3]: *** [gtxtest] Error 1 make[3]: Leaving directory `/usr/local/afs/i386_linux22/obj/gtx' make[2]: *** [gtx] Error 2 make[2]: Leaving directory `/usr/local/afs' make[1]: *** [install] Error 2 make[1]: Leaving directory `/usr/local/afs' make: *** [all] Error 2 --------------------------------------------------------------------------- /usr/lib/libncurses.so IS installed ls /usr/lib/libncurses. libncurses.a libncurses.so libncurses.so.5 libncurses.so.5.0 ============================================================= and these errors for box 2 (rh7.0) ============================================================= --------------------------------------------------------------------------------------------------------------------- In file included from ../afs/sysincludes.h:125, from ../afs/afs_analyze.c:14: /usr/include/pthread.h:270: parse error before `0' make[4]: *** [../UAFS/afs_analyze.o] Error 1 make[4]: Leaving directory `/usr/local/afs/i386_linux22/obj/libuafs/UAFS' make[3]: *** [UAFS/libuafs.a] Error 2 make[3]: Leaving directory `/usr/local/afs/i386_linux22/obj/libuafs' make[2]: *** [libuafs] Error 2 make[2]: Leaving directory `/usr/local/afs' make[1]: *** [install] Error 2 make[1]: Leaving directory `/usr/local/afs' make: *** [all] Error 2 ------------------------------------------------------------------------------ Any ideas? I'm really looking foward to trying afs, I've heard many Good Things about it. Thanks in advance Brian Kosick From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa19697; 9 Nov 2000 10:27 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id JAA23030; Thu, 9 Nov 2000 09:27:30 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 9 Nov 2000 09:27:30 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5B92@umr-mail02> From: "Neulinger, Nathan R." To: "'info-afs@transarc.com'" Cc: "'openafs-info@central.org'" Date: Thu, 9 Nov 2000 09:27:28 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Subject: [OpenAFS] file and vl server prefs - do "same preference" servers get load balanced/round robin? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I've noticed recently that many of our clients are choosing the same file server to be the lowest preference server for both FS and VL traffic. Needless to say, this seems to be causing a dramatic effect on the loading of the various servers. Will setting the server prefs all to the same value cause the loading to balance across the servers, or will it result in all traffic going to the first server in the list? We have never messed with hand setting the server prefs before, so don't have much experience with how this will affect things. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from grace.speakeasy.org by penn.central.org id aa20513; 9 Nov 2000 12:11 EST Received: (qmail 13426 invoked by uid 27365); 9 Nov 2000 17:11:07 -0000 Date: 9 Nov 2000 17:11:07 -0000 Message-ID: <20001109171107.13425.qmail@grace.speakeasy.org> From: aeneous@speakeasy.org To: openafs-info@central.org Cc: nneul@umr.edu Subject: [OpenAFS] server preferences Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: the cm adds a small random value to each preference, so setting them all to the same value should effectively balance them out. After you do an fs sets, do fs gets, and you'll see the randomized settings. The defaults are calculated based on subnet masks, so apparently one of your servers is on the same subnet as many of your clients? From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from pilsener.srv.ualberta.ca by penn.central.org id aa20815; 9 Nov 2000 14:04 EST Received: from speedbump.ucs.ualberta.ca (speedbump.ucs.ualberta.ca [129.128.11.44]) by pilsener.srv.ualberta.ca (8.10.2/8.10.2) with ESMTP id eA9J4nM05065; Thu, 9 Nov 2000 12:04:49 -0700 (MST) Received: (from cb@localhost) by speedbump.ucs.ualberta.ca (8.9.3/8.9.3/Debian 8.9.3-21) id MAA25238; Thu, 9 Nov 2000 12:04:48 -0700 X-Authentication-Warning: speedbump.ucs.ualberta.ca: cb set sender to cb@speedbump.ucs.ualberta.ca using -f To: "Neulinger, Nathan R." Cc: "'Brian Kosick'" , "'openafs-info@central.org'" References: <9DA8D24B915BD1118911006094516EAF053E5B77@umr-mail02> Reply-to: chris.bayly@ualberta.ca From: Chris Date: 09 Nov 2000 12:04:48 -0700 In-Reply-To: "Neulinger, Nathan R."'s message of "Tue, 7 Nov 2000 12:13:44 -0600" Message-ID: <5066lx0xwv.fsf@speedbump.ucs.ualberta.ca> Lines: 45 User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: [OpenAFS] Re: compiling on linux 2.2.x Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "NNR" == Neulinger, Nathan R writes: NNR> This should be on openafs-devel@central.org or NNR> openafs-info@central.org. To fix, edit the file NNR> src/config/Makefile.i386_linux22 and change the kernel line NNR> to only list 2.2.14. Make sure that you have a link to the NNR> kernel sources as /usr/src/linux-2.2.14. NNR> Then, run the make again. Or just run make with the proper options :). (The README is incorrect and SHOULD have the following: make SYS="i386_linux22" LINUX_VERS="2.2.14" instead of the make SYS_NAME="i386_linux22" LINUX_VERSION="2.2.14" >> -----Original Message----- From: Brian Kosick >> [mailto:briank@nacs.net] Sent: Tuesday, November 07, 2000 11:35 >> AM To: info-afs@transarc.com Subject: compiling on linux 2.2.x >> >> >> Hello all, forgive me if this is the wrong place to be posting >> this, I've browsed over the archives, and have really not found >> anything on openafs-1.0.0 compile problems. I have two boxes >> that I have tried on, 1) Intel P200MMX 128MB RAM 10G IDE HD >> running fresh install of RH7.0 kernel 2.2.16-22 2) AMD k6-2 500 >> 128MB RAM 4G IDE, 6G IDE, 2 x 3G SCSI Raid 0 running RH6.2 >> kernel 2.2.15 >> >> I have downloaded, and compiled kernel 2.2.14 on both, so I >> could get some valid headers for open afs, the kernel is >> compiled, but NOT running. -- Chris Bayly Email: Chris.Bayly@UAlberta.CA | CNS, UNIX Support Phone: (780) 492-9343 | 151 General Services Building Fax: (780) 492-1729 | University of Alberta Web: http://www.ualberta.ca/~cbayly/ | Edmonton, Alberta | Canada T6G 2S7 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from READ-EVAL-PRINT-LOOP.MIT.EDU by penn.central.org id aa01539; 13 Nov 2000 1:09 EST Received: from read-eval-print-loop.mit.edu (localhost [127.0.0.1]) by read-eval-print-loop.mit.edu (8.8.8/8.8.8) with ESMTP id BAA04046 for ; Mon, 13 Nov 2000 01:09:35 -0500 (EST) Message-Id: <200011130609.BAA04046@read-eval-print-loop.mit.edu> To: openafs-info@central.org Subject: Re: [OpenAFS] apenafs 1.0.0 linux compile problems In-Reply-To: Message from Brian Kosick of "Tue, 07 Nov 2000 15:14:49 EST." <4.3.1.2.20001107150431.00b44940@mail.nacs.net> References: <4.3.1.2.20001107150431.00b44940@mail.nacs.net> Date: Mon, 13 Nov 2000 01:09:34 -0500 From: Shantonu Sen Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > I'm trying to compile open afs on two linux boxes: > 1) Intel P200MMX 128MB RAM 10G IDE HD running fresh install of RH7.0 > kernel 2.2.16-22 > ============================================================= > and these errors for box 2 (rh7.0) > ============================================================= > ----------------------------------------------------------------------------- > ---------------------------------------- > In file included from ../afs/sysincludes.h:125, > from ../afs/afs_analyze.c:14: > /usr/include/pthread.h:270: parse error before `0' The following patch allows me to compile on redhat 7 with 2.4.0-test10. Although it's a different kernel, the root of your error is a messy glibc interaction, addressed by the second half of the patch. It removes a questionable pthread macro (as discussed briefly on openafs-devel). After that, my compiler had a seg fault while compiling venus, which is the first half of the patch. If you do not have that problem, feel free to disregard it. Good luck, Shantonu Index: src/Makefile =================================================================== RCS file: /cvs/openafs/src/Makefile,v retrieving revision 1.3 diff -r1.3 Makefile 440c440 < venus update xstat afsmonitor dauth tests libafsrpc \ --- > update xstat afsmonitor dauth tests libafsrpc \ Index: src/afs/UKERNEL/sysincludes.h =================================================================== RCS file: /cvs/openafs/src/afs/UKERNEL/sysincludes.h,v retrieving revision 1.2 diff -r1.2 sysincludes.h 115,119d114 < /* this is necessary for client programs as well as the library itself */ < #if defined(AFS_LINUX22_ENV) || defined(AFS_USR_LINUX22_ENV) < #define pthread_attr_setstacksize(a,b) 0 < #endif < From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa01569; 13 Nov 2000 1:42 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id BAA16216 for ; Mon, 13 Nov 2000 01:42:51 -0500 (EST) Date: Mon, 13 Nov 2000 01:47:17 -0500 From: Derrick J Brashear To: openafs-info@central.org Subject: Re: [OpenAFS] apenafs 1.0.0 linux compile problems Message-ID: <118178112.974098037@skittlebrau.trafford.dementia.org> In-Reply-To: <200011130609.BAA04046@read-eval-print-loop.mit.edu> X-Mailer: Mulberry/2.0.6a8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Monday, November 13, 2000 01:09:34 AM -0500 Shantonu Sen wrote: >> I'm trying to compile open afs on two linux boxes: >> 1) Intel P200MMX 128MB RAM 10G IDE HD running fresh install of RH7.0 >> kernel 2.2.16-22 > >> ============================================================= >> and these errors for box 2 (rh7.0) >> ============================================================= >> ------------------------------------------------------------------------ >> ----- ---------------------------------------- >> In file included from ../afs/sysincludes.h:125, >> from ../afs/afs_analyze.c:14: >> /usr/include/pthread.h:270: parse error before `0' > > The following patch allows me to compile on redhat 7 with > 2.4.0-test10. > > Although it's a different kernel, the root of your error > is a messy glibc interaction, addressed by the second half > of the patch. It removes a questionable pthread macro (as > discussed briefly on openafs-devel). > > After that, my compiler had a seg fault while compiling venus, > which is the first half of the patch. If you do not have that > problem, feel free to disregard it. kgcc won't segfault but will still fail to build kdump (on your redhat 7 system). From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ibm-1.MPA-Garching.MPG.DE by penn.central.org id aa02113; 13 Nov 2000 4:32 EST Received: from ibm-0.MPA-Garching.MPG.DE (ibm-0.MPA-Garching.MPG.DE [130.183.83.30]) by ibm-1.MPA-Garching.MPG.DE (AIX4.3/8.9.3/8.9.1) with ESMTP id KAA53816; Mon, 13 Nov 2000 10:32:56 +0100 Received: (from nog@localhost) by ibm-0.MPA-Garching.MPG.DE (8.9.1/8.9.1) id KAA19534; Mon, 13 Nov 2000 10:32:55 +0100 Date: Mon, 13 Nov 2000 10:32:55 +0100 From: Norbert Gruener To: Derrick J Brashear Cc: openafs-info@central.org Message-ID: <20001113103255.A16638@ibm-0.MPA-Garching.MPG.DE> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2i In-Reply-To: ; from shadow@dementia.org on Tue, Oct 31, 2000 at 05:29:28PM -0500 Organization: Max-Planck-Institute for Astrophysics, Garching, Germany Subject: [OpenAFS] Re: OpenAFS lists, cvs to be available at openafs.org Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Tue, Oct 31 2000, Derrick J Brashear wrote: > As soon as the relevant DNS changes happen, lists devoted to openafs > development will be available at openafs.org, and a cvs archive will also I am sure I have seen this archive already. But now I can't find it anymore :-(( Where is this cvs archive located now ? Cheers, Norbert -- +-------------------------------------------------------------------+ | Norbert E. Gruener http://www.MPA-Garching.MPG.DE/~nog/ | | Max-Planck-Institut | | fuer Astrophysik PGP encrypted mail preferred | | PGPprint(RSA): 66 64 C0 D0 6F 1A 16 02 C2 C6 37 83 3A 5F 88 9B | | PGPprint(DH): EBBF 02ED 1B91 39C3 5FB9 4D2F 9478 E224 334C 34CD | | PGPprint(DSS): A3F0 BDFA EAC0 128D A6C5 B11D 6EEC 0A88 47CD A77D | +-------------------------------------------------------------------+ From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from devel.nacs.net by penn.central.org id aa03814; 13 Nov 2000 20:04 EST Received: from nine.nacs.net (tawny.nacs.net [207.166.192.127]) by devel.nacs.net (8.9.3/8.9.3) with ESMTP id UAA15040 for ; Mon, 13 Nov 2000 20:03:59 -0500 Message-Id: <4.3.1.2.20001113193754.00cd49b0@mail.nacs.net> X-Sender: briank@mail.nacs.net X-Mailer: QUALCOMM Windows Eudora Version 4.3.1 Date: Mon, 13 Nov 2000 19:59:55 -0500 To: openafs-info@central.org From: Brian Kosick Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Subject: [OpenAFS] a few open afs questions... Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hello all, First of all I'd like to thank all those who helped me to get up and running with it. Second of all I'd like to know how the stability is with it. Once its compiled, I have had my machines lock up once or twice because of it. (User error perhaps ;-) ) But over all it seems pretty solid to me. The file caching really surprised me, when I first setup it up, I copied a relatively large user directory over to it, and logged in... I had to check twice to see if I was really on the remote box or not.;-) I really was not prepared for it. What I'm trying to do is set up a server cluster, and mount the file systems of the nodes using afs. (I'm currently utilizing a nfs solution and not too impressed). If any one has some experiences with the pros and cons of something like this could email me I'd appreciate it. Also, what if two people modify the same document at once? Ideally, I would like to put things like "/var/spool/mail", web pages and perhaps databases on it. Oh yeah, does play nice with RAID 5? Thanks, Brian Kosick Web Programmer New Age Consulting Service, Inc. 216-619-2000 briank@nacs.net From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa12951; 14 Nov 2000 9:09 EST Received: (qmail 27692 invoked from network); 14 Nov 2000 14:08:38 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 14 Nov 2000 14:08:38 -0000 Message-ID: <3A114764.55EF5F8D@umr.edu> Date: Tue, 14 Nov 2000 08:08:36 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18pre17 i686) X-Accept-Language: en MIME-Version: 1.0 To: Brian Kosick CC: openafs-info@central.org Subject: Re: [OpenAFS] a few open afs questions... References: <4.3.1.2.20001113193754.00cd49b0@mail.nacs.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > Also, what if two people modify the same document at once? Ideally, I > would like to put things like "/var/spool/mail", web pages and perhaps > databases on it. Oh yeah, does play nice with RAID 5? I don't know about the block level changes to files that would occur with databases, I'd be inclined to avoid that. Mail - go with a maildir format spool and you should have no problems at all. You can use a standard mbox style spool, but you're much safer with a maildir spool as it doesn't need to do any locking, and it is safe to be accessed from multiple clients simultaneously. As for web pages, I'd say that based on how most editors save files (write out the whole new file as opposed to changed pieces), it will be the last person to close the file. Now, if they are on the same afs client box, that behavior will likely be different, since they'll be accessing the same instance of the file in the cache, but I don't know enough about the low level behavior to know for sure. As for raid 5, that's purely a filesystem issue. AFS shouldn't care. We use hardware raid 5 on all of our transarc-afs file servers. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17672; 15 Nov 2000 17:07 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id RAA18102; Wed, 15 Nov 2000 17:07:15 -0500 (EST) Date: Wed, 15 Nov 2000 17:07:15 -0500 (EST) Message-Id: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman To: openafs-info@central.org Subject: [OpenAFS] Directory layout for new cells Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: So, I'm working on Debian packages for Openafs. Partially because of Debian policy and partially because it is useful to the user community , I'm trying to make it easy to set up AFS cells. As such, I'm working on a script to create and populate root.afs, root.cell and the top volumes in the cell. I'm interested in whether anyone else is doing something similar for other operating systems/environments. If so, we should work together. I'm also interested in established practice for what belongs in an initial cell. I'm mostly ignoring the Transarc documentation because I believe it is a bit dated and seems to focus more on how your clients find their AFS binaries than on how to set up a cell to get useful work done. If the point of a cell organization is only to make the lives of the AFS administrators (as apposed to the users) easier, then perhaps that cell is not yet a critical part of the community it serves. I tend to be a bit MIT-centric in my cell organization; I think cell/system/@sys is cleaner than cell/@sys, etc. However, I'm interested in discussion of general cell hierarchy. --Sam From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from FORT-POINT-STATION.MIT.EDU by penn.central.org id aa17723; 15 Nov 2000 17:17 EST Received: from grand-central-station.MIT.EDU (GRAND-CENTRAL-STATION.MIT.EDU [18.69.0.34]) by fort-point-station.mit.edu (8.9.2/8.9.2) with ESMTP id RAA03781 for ; Wed, 15 Nov 2000 17:17:18 -0500 (EST) Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id RAA25441; Wed, 15 Nov 2000 17:17:18 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id RAA19021; Wed, 15 Nov 2000 17:17:18 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id RAA14020; Wed, 15 Nov 2000 17:17:13 -0500 (EST) To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Derek Atkins Date: 15 Nov 2000 17:17:13 -0500 In-Reply-To: Sam Hartman's message of "Wed, 15 Nov 2000 17:07:15 -0500 (EST)" Message-ID: Lines: 55 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: When you install your server package, how does it know whether it's installing a database server or fileserver? And how does it know whether it needs to generate a new cell? This seems like somewhere that might be very challenging to do in a general way. I can certainly see creating a bunch of scripts that: 1) configure a new new on this new primary db/fileserver 2) configure a new db server and add it to the cell 3) configure a new fileserver and add it to the cell FTR, I don't do any of this in my RedHat scripts; I let the user setup servers and cells on their own. However, having such scripts would be nice, and I would certainly include them in my packages. -derek Sam Hartman writes: > So, I'm working on Debian packages for Openafs. Partially because of > Debian policy and partially because it is useful to the user community > , I'm trying to make it easy to set up AFS cells. > > As such, I'm working on a script to create and populate root.afs, > root.cell and the top volumes in the cell. > > I'm interested in whether anyone else is doing something similar for > other operating systems/environments. If so, we should work together. > > I'm also interested in established practice for what belongs in an > initial cell. I'm mostly ignoring the Transarc documentation because > I believe it is a bit dated and seems to focus more on how your > clients find their AFS binaries than on how to set up a cell to get > useful work done. If the point of a cell organization is only to make > the lives of the AFS administrators (as apposed to the users) easier, > then perhaps that cell is not yet a critical part of the community it > serves. > > I tend to be a bit MIT-centric in my cell organization; I think > cell/system/@sys is cleaner than cell/@sys, etc. However, I'm > interested in discussion of general cell hierarchy. > > > > --Sam > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17743; 15 Nov 2000 17:25 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id RAA18115; Wed, 15 Nov 2000 17:25:00 -0500 (EST) To: Derek Atkins Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman Date: 15 Nov 2000 17:25:00 -0500 In-Reply-To: Derek Atkins's message of "15 Nov 2000 17:17:13 -0500" Message-ID: Lines: 34 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Derek" == Derek Atkins writes: Derek> When you install your server package, how does it know Derek> whether it's installing a database server or fileserver? Derek> And how does it know whether it needs to generate a new Derek> cell? This seems like somewhere that might be very Derek> challenging to do in a general way. If they don't install openafs-dbserver they are probably not setting up a dbserver. I don't think I'll do the server setup just from package install, but if I do, I'll ask the user using debconf. Derek> I can certainly see creating a bunch of scripts that: 1) Derek> configure a new new on this new primary db/fileserver 2) Derek> configure a new db server and add it to the cell 3) Derek> configure a new fileserver and add it to the cell I'm looking at writing afs-newcell - set up primary db server and set up initial prdb (using ptutil or something not involving noauth) afs-addserver - add a file or db server and sync the equivelent of /usr/afs/etc (/etc/openafs/server in my case) among them afs-rootvol - create root.cell and populate root.afs I am not quite sure where to integrate root.afs creation, as you want to do that before starting an AFS client on the server. Annoyingly you cannot get tokens before loading the kernel module. I don't want to depend on a particular aklog/klog/whatever. I may add root.afs creation to afs-newcell, asserting that at least initially the primary db server is also a fileserver. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ginger.cmf.nrl.navy.mil by penn.central.org id aa17763; 15 Nov 2000 17:35 EST Received: from cmf.nrl.navy.mil (borg.cmf.nrl.navy.mil [134.207.10.148]) (authenticated) by ginger.cmf.nrl.navy.mil (8.10.1/8.10.1) with ESMTP id eAFMYwH20916; Wed, 15 Nov 2000 17:34:59 -0500 (EST) Message-Id: <200011152234.eAFMYwH20916@ginger.cmf.nrl.navy.mil> To: Sam Hartman cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells In-reply-to: Your message of "Wed, 15 Nov 2000 17:07:15 EST." <200011152207.RAA18102@sweet-transvestite.mit.edu> X-url: http://www.nrl.navy.mil/CCS/people/chas/index.html X-mailer: nmh 1.0 Date: Wed, 15 Nov 2000 17:34:58 -0500 From: Chas Williams Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: In message <200011152207.RAA18102@sweet-transvestite.mit.edu>,Sam Hartman write s: >I tend to be a bit MIT-centric in my cell organization; I think >cell/system/@sys is cleaner than cell/@sys, etc. However, I'm >interested in discussion of general cell hierarchy. well... we use depot and cell/@sys points the appropriate release tree from the depot system. so people put /afs/cmf/@sys/bin in their paths. in the old days (when running sunos) it was rather important to keep the path to bin short so #! in shell scirpts would work correctly. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from FORT-POINT-STATION.MIT.EDU by penn.central.org id aa17769; 15 Nov 2000 17:38 EST Received: from grand-central-station.MIT.EDU (GRAND-CENTRAL-STATION.MIT.EDU [18.69.0.34]) by fort-point-station.mit.edu (8.9.2/8.9.2) with ESMTP id RAA08697 for ; Wed, 15 Nov 2000 17:37:57 -0500 (EST) Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id RAA01547; Wed, 15 Nov 2000 17:37:56 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id RAA25571; Wed, 15 Nov 2000 17:37:56 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id RAA14035; Wed, 15 Nov 2000 17:37:52 -0500 (EST) To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Derek Atkins Date: 15 Nov 2000 17:37:52 -0500 In-Reply-To: Sam Hartman's message of "15 Nov 2000 17:25:00 -0500" Message-ID: Lines: 61 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman writes: > If they don't install openafs-dbserver they are probably not setting > up a dbserver. I don't think I'll do the server setup just from > package install, but if I do, I'll ask the user using debconf. Ahh, ok. It was unclear whether you meant to do this at package install time or 'later'. It would be nice if we could come up with a set of scripts such that you could hook in debconf later, but they scripts themselves didn't depend on it :) Also, I didn't even think about splitting out db-server from file-server packages. That might be a good idea, I'll have to think about it. > Derek> I can certainly see creating a bunch of scripts that: 1) > Derek> configure a new new on this new primary db/fileserver 2) > Derek> configure a new db server and add it to the cell 3) > Derek> configure a new fileserver and add it to the cell > > I'm looking at writing > afs-newcell - set up primary db server and set up initial prdb > (using ptutil or something not involving noauth) > > afs-addserver - add a file or db server and > sync the equivelent of /usr/afs/etc (/etc/openafs/server in my case) > among them > > afs-rootvol - create root.cell and populate root.afs > > I am not quite sure where to integrate root.afs creation, as you want > to do that before starting an AFS client on the server. Annoyingly > you cannot get tokens before loading the kernel module. I don't want > to depend on a particular aklog/klog/whatever. I may add root.afs > creation to afs-newcell, asserting that at least initially the primary > db server is also a fileserver. I would think that afs-rootvol and afs-newcell would be one function, as I can't see why you could do one without the other. Are you particularly going to depend on Kerberos, or are you going to support KAServer? I honestly don't think that creating simple interactive scripts would be very challenging. Doing something that could be integrated into e.g. debconf might be more challenging (and isn't something that I personally would be interested in ;) To answer your original question, I like a minimal layout of: /afs/cellname/{service,system/@sys,user}/ Obviously some sites are going to want different setups. I don't think there is even a BCP. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17795; 15 Nov 2000 17:41 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id RAA18139; Wed, 15 Nov 2000 17:41:41 -0500 (EST) To: Chas Williams Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152234.eAFMYwH20916@ginger.cmf.nrl.navy.mil> From: Sam Hartman Date: 15 Nov 2000 17:41:41 -0500 In-Reply-To: Chas Williams's message of "Wed, 15 Nov 2000 17:34:58 -0500" Message-ID: Lines: 26 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Chas" == Chas Williams writes: Chas> In message Chas> <200011152207.RAA18102@sweet-transvestite.mit.edu>,Sam Chas> Hartman write s: >> I tend to be a bit MIT-centric in my cell organization; I think >> cell/system/@sys is cleaner than cell/@sys, etc. However, I'm >> interested in discussion of general cell hierarchy. Chas> well... we use depot and cell/@sys points the appropriate Chas> release tree from the depot system. so people put Chas> /afs/cmf/@sys/bin in their paths. in the old days (when Chas> running sunos) it was rather important to keep the path to Chas> bin short so #! in shell scirpts would work correctly. I understand why this was done and I'm not trying to convince existing cells to change. I'm also not really addressing large cells. They'll have their own requirements and if they have to do some/all of the cell creation on their own, well that's OK. I'm thinking about new small cells--many of them single-server cells running on one machine. Most such cells will not be running SunOS or Ultrix. If I were building Depot from scratch, would it be possible to reasonably change /afs/cell/@sys to /afs/cell/system/@sys in a few places in the source? From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17824; 15 Nov 2000 17:48 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id RAA18143; Wed, 15 Nov 2000 17:48:36 -0500 (EST) To: Derek Atkins Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman Date: 15 Nov 2000 17:48:35 -0500 In-Reply-To: Derek Atkins's message of "15 Nov 2000 17:37:52 -0500" Message-ID: Lines: 17 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: afs-rootvol and afs-newcell are separate because afs-rootvol requires tokens and afs-newcell runs before tokens can exist. I do not plan to package kaserver unless I get significant requests from the Debian community to do so. I would hate to accidentally encourage the spread of krb4 at this point in time. I'm currently working on interactive scripts but will add command line options. Doing debconf integration and not doing the cell setup at install time seems silly. Even if I do debconf integration, I'll have scripts that can run interactively so they can be run after install time because that will certainly need to be an option I need to support. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa17843; 15 Nov 2000 17:52 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA22766; Wed, 15 Nov 00 17:52:07 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id RAA05618; Wed, 15 Nov 2000 17:52:37 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id RAA29950; Wed, 15 Nov 2000 17:52:37 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id RAA14045; Wed, 15 Nov 2000 17:52:33 -0500 (EST) To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Derek Atkins Date: 15 Nov 2000 17:52:33 -0500 In-Reply-To: Sam Hartman's message of "15 Nov 2000 17:48:35 -0500" Message-Id: Lines: 35 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman writes: > afs-rootvol and afs-newcell are separate because afs-rootvol requires > tokens and afs-newcell runs before tokens can exist. That only implies that you need to setup your ptserver before you do the rest. Nothing says you can't already have libafs loaded in your kernel. So, I sstill don't see why they have to be two separate scripts. > I do not plan to package kaserver unless I get significant requests > from the Debian community to do so. I would hate to accidentally > encourage the spread of krb4 at this point in time. Fair enough. I would at least package the ka client software. > I'm currently working on interactive scripts but will add command line > options. Sounds like a reasonable approach. > Doing debconf integration and not doing the cell setup at install time > seems silly. Even if I do debconf integration, I'll have scripts that > can run interactively so they can be run after install time because > that will certainly need to be an option I need to support. Ok. :) -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from bodhi.src.uchicago.edu by penn.central.org id aa17862; 15 Nov 2000 17:59 EST Received: (from wngdn@localhost) by bodhi.src.uchicago.edu (8.9.3/8.9.3/Debian 8.9.3-21) id QAA32365 for openafs-info@central.org; Wed, 15 Nov 2000 16:59:14 -0600 From: Kelsang Wangden Date: Wed, 15 Nov 2000 16:59:06 -0600 To: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <20001115165906.K27538@src.uchicago.edu> References: <200011152207.RAA18102@sweet-transvestite.mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from warlord@mit.edu on Wed, Nov 15, 2000 at 05:37:52PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: To answer the question about layouts ... We built a new cell from scratch last December because our old cell had gotten quite crufty, and we were moving to new servers for Y2K anyway. I asked for some feedback on info-afs@transarc.org and got some really good responses - you might want to look in the archive for those. Anyway, here's the layout we settled on in /afs/src.uchicago.edu: admin/ administrative files arch/ sun4x_57/ one directory per @sys hp_ux102/ ... common/ stuff common to all system types data/ stuff belonging to our data library home/ symlinks to user home directories pkg/ opt-style software installations users/ actual user home directories www/ web pages Under arch/@sys and common/ are directories such as bin, lib, include, and man. In the arch/@sys ones the contents are generally symlinks into pkg//. In common/, these are generally scripts and config files which work across architechtures. In pkg/, there is a directory for each version of each software package - e.g. automake-1.4, bison-1.28, etc. Inside each of these are the standard dirctories such as bin, lib, and include. Some things are broken down further by architechture name (generaly bin and lib). Under users/ we have the user directories split up to keep directory contents low. If your username is foobar, for example, then the official mount point for your home directory is users/f/o/foobar/, but I wouldn't actually tell you that. home/foobar is a symlink to there, and that's the one I tell users. (Actually I tell them /afs/home/foobar, because /afs/home is a symlink to /afs/src.uchicago.edu/home/.) User web pages live in volumes which are mounted once inside www/ and once in their home directories. If they want multiple users editing the files, then multiple users get a mount of the volume in their home directories. Because of this, ordinary users don't need read access to the web tree, and permissions can be quite tight. I think that covers it. This is probably more complexity than the average user wants if they're just installing a package with dselect or apt-get. Good AFS cell structure takes some thinking, some planning, and (IMHO) some experience with good or bad structure to which you can compare your ideas. Regards, Wangden -- Kelsang Wangden (Buddhist monk) Technical Manager Social Science Research Computing, University of Chicago wngdn@src.uchicago.edu (773) 702-3792 PGP key: http://www.src.uchicago.edu/~wngdn/key.txt From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17864; 15 Nov 2000 17:59 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id RAA18163; Wed, 15 Nov 2000 17:59:36 -0500 (EST) To: Derek Atkins Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman Date: 15 Nov 2000 17:59:35 -0500 In-Reply-To: Derek Atkins's message of "15 Nov 2000 17:52:33 -0500" Message-ID: Lines: 26 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Derek" == Derek Atkins writes: Derek> Sam Hartman writes: >> afs-rootvol and afs-newcell are separate because afs-rootvol >> requires tokens and afs-newcell runs before tokens can exist. Derek> That only implies that you need to setup your ptserver Derek> before you do the rest. Nothing says you can't already Derek> have libafs loaded in your kernel. So, I sstill don't see Derek> why they have to be two separate scripts. I may get there. This is sort of annoying because I don't want to run afsd before root.afs exists. >> I do not plan to package kaserver unless I get significant >> requests from the Debian community to do so. I would hate to >> accidentally encourage the spread of krb4 at this point in >> time. Derek> Fair enough. I would at least package the ka client Derek> software. I package klog and unlog. I will package kpasswd but it is going to be in its own package because there are so many kpasswds. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17894; 15 Nov 2000 18:04 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id SAA18172; Wed, 15 Nov 2000 18:04:41 -0500 (EST) To: Kelsang Wangden Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> <20001115165906.K27538@src.uchicago.edu> From: Sam Hartman Date: 15 Nov 2000 18:04:41 -0500 In-Reply-To: Kelsang Wangden's message of "Wed, 15 Nov 2000 16:59:06 -0600" Message-ID: Lines: 20 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Kelsang" == Kelsang Wangden writes: Kelsang> To answer the question about layouts ... We built a new Kelsang> cell from scratch last December because our old cell had Kelsang> gotten quite crufty, and we were moving to new servers Kelsang> for Y2K anyway. I asked for some feedback on Kelsang> info-afs@transarc.org and got some really good responses Kelsang> - you might want to look in the archive for those. Kelsang> Anyway, here's the layout we settled on in Kelsang> /afs/src.uchicago.edu: Kelsang> admin/ administrative files arch/ sun4x_57/ one directory How common is the arch designation at top-level? MIT uses system for that at the top-level and arch within package-specific "lockers." If arch is very common, then I'll probably create arch as the directory with system as a symlink. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from bodhi.src.uchicago.edu by penn.central.org id aa17902; 15 Nov 2000 18:08 EST Received: (from wngdn@localhost) by bodhi.src.uchicago.edu (8.9.3/8.9.3/Debian 8.9.3-21) id RAA32436; Wed, 15 Nov 2000 17:08:06 -0600 From: Kelsang Wangden Date: Wed, 15 Nov 2000 17:08:06 -0600 To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <20001115170806.L27538@src.uchicago.edu> References: <200011152207.RAA18102@sweet-transvestite.mit.edu> <20001115165906.K27538@src.uchicago.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from hartmans@MIT.EDU on Wed, Nov 15, 2000 at 06:04:41PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman wrote: > >>>>> "Kelsang" == Kelsang Wangden writes: > > Kelsang> To answer the question about layouts ... We built a new > Kelsang> cell from scratch last December because our old cell had > Kelsang> gotten quite crufty, and we were moving to new servers > Kelsang> for Y2K anyway. I asked for some feedback on > Kelsang> info-afs@transarc.org and got some really good responses > Kelsang> - you might want to look in the archive for those. > > Kelsang> Anyway, here's the layout we settled on in > Kelsang> /afs/src.uchicago.edu: > > Kelsang> admin/ administrative files arch/ sun4x_57/ one directory > > How common is the arch designation at top-level? MIT uses system for > that at the top-level and arch within package-specific "lockers." > > > If arch is very common, then I'll probably create arch as the > directory with system as a symlink. IIRC, my co-worker Jon and I decided on "arch" in a brainstorming session, so for all I know it's unique to our site. Regards, Wangden -- Kelsang Wangden (Buddhist monk) Technical Manager Social Science Research Computing, University of Chicago wngdn@src.uchicago.edu (773) 702-3792 PGP key: http://www.src.uchicago.edu/~wngdn/key.txt From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa17904; 15 Nov 2000 18:08 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA26514; Wed, 15 Nov 00 18:07:45 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id SAA09307; Wed, 15 Nov 2000 18:08:15 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id SAA03911; Wed, 15 Nov 2000 18:08:15 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id SAA14056; Wed, 15 Nov 2000 18:08:15 -0500 (EST) To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Derek Atkins Date: 15 Nov 2000 18:08:14 -0500 In-Reply-To: Sam Hartman's message of "15 Nov 2000 17:59:35 -0500" Message-Id: Lines: 26 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman writes: > >>>>> "Derek" == Derek Atkins writes: > > Derek> Sam Hartman writes: > >> afs-rootvol and afs-newcell are separate because afs-rootvol > >> requires tokens and afs-newcell runs before tokens can exist. > > Derek> That only implies that you need to setup your ptserver > Derek> before you do the rest. Nothing says you can't already > Derek> have libafs loaded in your kernel. So, I sstill don't see > Derek> why they have to be two separate scripts. > > I may get there. This is sort of annoying because I don't want to run > afsd before root.afs exists. Hrm, yea. Why not just generate a minimal root.afs and root.cell using -localauth? -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa17942; 15 Nov 2000 18:11 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id SAA18180; Wed, 15 Nov 2000 18:11:21 -0500 (EST) To: Derek Atkins Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman Date: 15 Nov 2000 18:11:21 -0500 In-Reply-To: Derek Atkins's message of "15 Nov 2000 18:08:14 -0500" Message-ID: Lines: 9 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Derek" == Derek Atkins writes: Derek> Hrm, yea. Why not just generate a minimal root.afs and Derek> root.cell using -localauth? I can generate the root.afs volume. This is probably what I will do, but I cannot populate it until I have a client with tokens. That is why I see the script split. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from FORT-POINT-STATION.MIT.EDU by penn.central.org id aa17966; 15 Nov 2000 18:16 EST Received: from grand-central-station.MIT.EDU (GRAND-CENTRAL-STATION.MIT.EDU [18.69.0.34]) by fort-point-station.mit.edu (8.9.2/8.9.2) with ESMTP id SAA16705 for ; Wed, 15 Nov 2000 18:16:31 -0500 (EST) Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id SAA10983; Wed, 15 Nov 2000 18:16:30 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id SAA05689; Wed, 15 Nov 2000 18:16:30 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id SAA14067; Wed, 15 Nov 2000 18:16:29 -0500 (EST) To: Sam Hartman Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Derek Atkins Date: 15 Nov 2000 18:16:29 -0500 In-Reply-To: Sam Hartman's message of "15 Nov 2000 18:11:21 -0500" Message-ID: Lines: 29 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman writes: > >>>>> "Derek" == Derek Atkins writes: > > Derek> Hrm, yea. Why not just generate a minimal root.afs and > Derek> root.cell using -localauth? > > I can generate the root.afs volume. This is probably what I will do, > but I cannot populate it until I have a client with tokens. That is > why I see the script split. Well, you could: 0) generate ptdb 1) generate root.afs that only contains the new cell 2) generate root.cell 3) start afsd 4) get tokens 5) rebuild root.afs I don't see why you need tokens, tho.. You can keep the cell with security turned off for the short-time it takes to populate root.afs and root.cell. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from windlord.Stanford.EDU by penn.central.org id aa17986; 15 Nov 2000 18:29 EST Received: (qmail 10502 invoked by uid 50); 15 Nov 2000 23:29:26 -0000 To: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> <20001115165906.K27538@src.uchicago.edu> <20001115170806.L27538@src.uchicago.edu> In-Reply-To: Kelsang Wangden's message of "Wed, 15 Nov 2000 17:08:06 -0600" From: Russ Allbery Organization: The Eyrie Date: 15 Nov 2000 15:29:25 -0800 Message-ID: Lines: 46 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Kelsang Wangden writes: > Sam Hartman wrote: >> How common is the arch designation at top-level? MIT uses system for >> that at the top-level and arch within package-specific "lockers." >> If arch is very common, then I'll probably create arch as the >> directory with system as a symlink. > IIRC, my co-worker Jon and I decided on "arch" in a brainstorming > session, so for all I know it's unique to our site. We use systems instead. Our cell layout looks vaguely like: /afs// class Course directories data Large data volumes (statistical data sets, etc.) dept University department directories dev Local source trees and CVS repositories dist Distribution trees (ftp and web) group Student organizations and research groups pubsw Software packages (symlinked into system/@sys) service CellServDB and whatnot site University computing organizations systems/@sys Trees for each sysname users/u/s/username User volumes Our AFS installation instructions have people create a few symlinks, most notably /usr/pubsw -> /afs//systems/@sys/pubsw. That directory is a pure symlink tree, with all of the symlinks pointing into ../package, which is another mount for /afs//pubsw. In that directory, things are broken down by class of software (Data, Kerberos, Editors, etc.) and then by package name with version in the same sort of system that lots of people use with various different automated tools (I rolled my own). If we had it to do all over again, /afs//pubsw would probably be called package instead. /usr/pubsw is a good thing to call the software tree, though, since "pubsw" has the same number of letters as "local", allowing us to do binary edits of some vendor software to convince it to look in the right place even if we can't recompile it (and yeah, we've had to do this quite a few times). -- Russ Allbery (rra@stanford.edu) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SMTP1.ANDREW.CMU.EDU by penn.central.org id aa18009; 15 Nov 2000 18:30 EST Received: from LARIMER.ANDREW.CMU.EDU (LARIMER.ANDREW.CMU.EDU [128.2.122.210]) by smtp1.andrew.cmu.edu (8.9.3/8.9.3) with ESMTP id SAA24125; Wed, 15 Nov 2000 18:30:22 -0500 (EST) Date: Wed, 15 Nov 2000 18:30:24 -0500 (EST) From: Jeffrey Hutzelman X-Sender: jhutz@larimer.andrew.cmu.edu To: Sam Hartman cc: Derek Atkins , openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On 15 Nov 2000, Sam Hartman wrote: > afs-rootvol - create root.cell and populate root.afs > > I am not quite sure where to integrate root.afs creation, as you want > to do that before starting an AFS client on the server. Annoyingly > you cannot get tokens before loading the kernel module. I don't want > to depend on a particular aklog/klog/whatever. I may add root.afs > creation to afs-newcell, asserting that at least initially the primary > db server is also a fileserver. It would seem that the obvious way to do this would be to use a tool which allows you to create volume dumps, and then 'vos restare' such a dump onto the fileserver. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa18083; 15 Nov 2000 18:39 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id SAA18232; Wed, 15 Nov 2000 18:39:36 -0500 (EST) To: Derek Atkins Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <200011152207.RAA18102@sweet-transvestite.mit.edu> From: Sam Hartman Date: 15 Nov 2000 18:39:35 -0500 In-Reply-To: Derek Atkins's message of "15 Nov 2000 18:16:29 -0500" Message-ID: Lines: 8 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Derek" == Derek Atkins writes: Derek> Well, you could: 0) generate ptdb 1) generate root.afs that Derek> only contains the new cell 2) generate root.cell 3) start Derek> afsd 4) get tokens 5) rebuild root.afs As I said earlier, noauth considered harmful. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from bodhi.src.uchicago.edu by penn.central.org id aa20775; 16 Nov 2000 11:30 EST Received: (from wngdn@localhost) by bodhi.src.uchicago.edu (8.9.3/8.9.3/Debian 8.9.3-21) id KAA01771; Thu, 16 Nov 2000 10:30:41 -0600 From: Kelsang Wangden Date: Thu, 16 Nov 2000 10:30:27 -0600 To: Sam Hartman Cc: Derek Atkins , openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <20001116103027.A1707@src.uchicago.edu> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from hartmans@MIT.EDU on Wed, Nov 15, 2000 at 06:39:35PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sam Hartman wrote: > >>>>> "Derek" == Derek Atkins writes: > > > Derek> Well, you could: 0) generate ptdb 1) generate root.afs that > Derek> only contains the new cell 2) generate root.cell 3) start > Derek> afsd 4) get tokens 5) rebuild root.afs > > As I said earlier, noauth considered harmful. Why do you consider it harmful? It's the recommended way to build a new cell. Is it because apt, dpkg, and friends install things while there are potentially lots of normal things going on like users logging in remotely, web pages being served out, and so on? Wangden -- Kelsang Wangden (Buddhist monk) Technical Manager Social Science Research Computing, University of Chicago wngdn@src.uchicago.edu (773) 702-3792 PGP key: http://www.src.uchicago.edu/~wngdn/key.txt From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa20869; 16 Nov 2000 11:42 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id KAA17702; Thu, 16 Nov 2000 10:42:43 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 16 Nov 2000 10:42:43 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5BD0@umr-mail02> From: "Neulinger, Nathan R." To: "'Kelsang Wangden'" , Sam Hartman Cc: Derek Atkins , openafs-info@central.org Subject: RE: [OpenAFS] Directory layout for new cells Date: Thu, 16 Nov 2000 10:42:37 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I might suggest that something like a "fs" equivalent to -localauth would be the way to go. In fact, if you combined that with a non-kernel ftp-like client, you could start accessing the servers using -localauth prior to the client cache manager being really up. That would eliminate the need to use noauth, and would also let you do all the volume creation/etc. right away. However, I think alot of this is moot depending on how long it's going to take to get the dynamically created root.afs support. Once that support is created, then you really only have to start creating root.cell, and you wouldn't actually need to auto-populate that volume with anything. -- Nathan > -----Original Message----- > From: Kelsang Wangden [mailto:wngdn@src.uchicago.edu] > Sent: Thursday, November 16, 2000 10:30 AM > To: Sam Hartman > Cc: Derek Atkins; openafs-info@central.org > Subject: Re: [OpenAFS] Directory layout for new cells > > > Sam Hartman wrote: > > >>>>> "Derek" == Derek Atkins writes: > > > > > > Derek> Well, you could: 0) generate ptdb 1) generate > root.afs that > > Derek> only contains the new cell 2) generate root.cell 3) start > > Derek> afsd 4) get tokens 5) rebuild root.afs > > > > As I said earlier, noauth considered harmful. > > Why do you consider it harmful? It's the recommended way to build a > new cell. Is it because apt, dpkg, and friends install things while > there are potentially lots of normal things going on like users > logging in remotely, web pages being served out, and so on? > > Wangden > -- > Kelsang Wangden (Buddhist monk) Technical Manager > Social Science Research Computing, University of Chicago > wngdn@src.uchicago.edu (773) 702-3792 > PGP key: http://www.src.uchicago.edu/~wngdn/key.txt > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from PACIFIC-CARRIER-ANNEX.MIT.EDU by penn.central.org id aa20968; 16 Nov 2000 11:54 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA24869; Thu, 16 Nov 00 11:55:36 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id LAA22012; Thu, 16 Nov 2000 11:54:21 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id LAA14496; Thu, 16 Nov 2000 11:54:21 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id LAA14664; Thu, 16 Nov 2000 11:54:20 -0500 (EST) To: "Neulinger, Nathan R." Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <9DA8D24B915BD1118911006094516EAF053E5BD0@umr-mail02> From: Derek Atkins Date: 16 Nov 2000 11:54:20 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 16 Nov 2000 10:42:37 -0600" Message-Id: Lines: 19 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: "Neulinger, Nathan R." writes: > However, I think alot of this is moot depending on how long it's going to > take to get the dynamically created root.afs support. Once that support is > created, then you really only have to start creating root.cell, and you > wouldn't actually need to auto-populate that volume with anything. > > -- Nathan Well, is anyone currently working on AFSDB RR support? Wouldn't that be a pre-requisite? -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa21105; 16 Nov 2000 12:31 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id LAA29987; Thu, 16 Nov 2000 11:31:36 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 16 Nov 2000 11:31:36 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5BD1@umr-mail02> From: "Neulinger, Nathan R." To: "'Derek Atkins'" Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: RE: [OpenAFS] Directory layout for new cells Date: Thu, 16 Nov 2000 11:31:29 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Dunno... I wouldn't think that would have to be a prerequisite. I mean - you have truly-dynamic root.afs creation, and you have boot-time root.afs creation. If you do boot time (easy first-shot approach), just have it process the CellServDB at the client. Later iterations of it could add cells. Probably would want to have the client automatically add a mount to the pseudo volume if it gets a 'fs newcell' command. Second iteration of the code would be fully dynamic, using cellservdb and all the afsdb recs. (Do you really want to use the AFSDB records, or do you want to use SRV records? I would think the latter might be a better idea, since suport of TXT/SRV is going to be forced by the fact that microsoft uses them extensively with ADS. -- Nathan > -----Original Message----- > From: Derek Atkins [mailto:warlord@MIT.EDU] > Sent: Thursday, November 16, 2000 10:54 AM > To: Neulinger, Nathan R. > Cc: 'Kelsang Wangden'; Sam Hartman; openafs-info@central.org > Subject: Re: [OpenAFS] Directory layout for new cells > > > "Neulinger, Nathan R." writes: > > > However, I think alot of this is moot depending on how long > it's going to > > take to get the dynamically created root.afs support. Once > that support is > > created, then you really only have to start creating > root.cell, and you > > wouldn't actually need to auto-populate that volume with anything. > > > > -- Nathan > > Well, is anyone currently working on AFSDB RR support? Wouldn't > that be a pre-requisite? > > -derek > > -- > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > Member, MIT Student Information Processing Board (SIPB) > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH > warlord@MIT.EDU PGP key available > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa21140; 16 Nov 2000 12:46 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA10298; Thu, 16 Nov 00 12:45:46 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id MAA09349; Thu, 16 Nov 2000 12:46:16 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id MAA03562; Thu, 16 Nov 2000 12:46:16 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id MAA14761; Thu, 16 Nov 2000 12:46:15 -0500 (EST) To: "Neulinger, Nathan R." Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) References: <9DA8D24B915BD1118911006094516EAF053E5BD1@umr-mail02> From: Derek Atkins Date: 16 Nov 2000 12:46:15 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 16 Nov 2000 11:31:29 -0600" Message-Id: Lines: 43 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: "Neulinger, Nathan R." writes: > Dunno... I wouldn't think that would have to be a prerequisite. I mean - you > have truly-dynamic root.afs creation, and you have boot-time root.afs > creation. If you do boot time (easy first-shot approach), just have it > process the CellServDB at the client. Later iterations of it could add > cells. Probably would want to have the client automatically add a mount to > the pseudo volume if it gets a 'fs newcell' command. > > Second iteration of the code would be fully dynamic, using cellservdb and > all the afsdb recs. (Do you really want to use the AFSDB records, or do you > want to use SRV records? I would think the latter might be a better idea, > since suport of TXT/SRV is going to be forced by the fact that microsoft > uses them extensively with ADS. I think that the approaches for these two mechanisms need not be the same. In the former case, you just need to supply a mechanism to build root.afs from fs newcell commands. I don't think that would be terribly difficult to do. In the latter case, however, you would need a mechanism for the kernel to request the DNS records for the newly requested cell (be they AFSDB or SRV, I don't particularly care). I would certainly prefer not to include DNS directly in the kernel module, which implies a callback to some user-space process to perform the DNS request. I'm not sure how this would work, except, perhaps, another afsd thread that sits waiting for 'callbacks' from the kernel? Hrm, that might be a reasonable approach for the kernel to know whether to build root.afs or not -- if the process calls the 'wait for AFSDB request' afs_call, then the kernel knows it can make requests for a dynamic root.afs. When it wants to make a request, it wakes up that thread and returns to user-space; the process then wakes up, does the DNS request, and then responds back to the kernel with the DNS response, finally waiting for more requests. Hrm. I think I like this approach. :) -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa21158; 16 Nov 2000 12:51 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id LAA01712; Thu, 16 Nov 2000 11:51:11 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 16 Nov 2000 11:51:11 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5BD2@umr-mail02> From: "Neulinger, Nathan R." To: "'Derek Atkins'" Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: RE: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) Date: Thu, 16 Nov 2000 11:51:05 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > I would certainly prefer not to include DNS directly in the kernel > module, which implies a callback to some user-space process to perform > the DNS request. I'm not sure how this would work, except, perhaps, > another afsd thread that sits waiting for 'callbacks' from the kernel? Why get so complicated? Just have the existing 'afsd' trap requests for the 'root.afs' volume and handle them internally. If you make this an option - people can still choose to continue using a REAL root.afs volume, or they can have afsd build a virtual one instead. -- Nathan From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa21164; 16 Nov 2000 12:53 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA12030; Thu, 16 Nov 00 12:52:53 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id MAA11599; Thu, 16 Nov 2000 12:53:24 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id MAA06029; Thu, 16 Nov 2000 12:53:24 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id MAA14777; Thu, 16 Nov 2000 12:53:23 -0500 (EST) To: "Neulinger, Nathan R." Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: Re: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) References: <9DA8D24B915BD1118911006094516EAF053E5BD2@umr-mail02> From: Derek Atkins Date: 16 Nov 2000 12:53:23 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 16 Nov 2000 11:51:05 -0600" Message-Id: Lines: 21 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: "Neulinger, Nathan R." writes: > Why get so complicated? Just have the existing 'afsd' trap requests for the > 'root.afs' volume and handle them internally. If you make this an option - > people can still choose to continue using a REAL root.afs volume, or they > can have afsd build a virtual one instead. Huh? afsd is just a call into the kernel, where almost all the work is done. None of the real AFS work is done by afsd proper. So, how would "'afsd' trap requests for 'root.afs'" when none of the file operations pass through the afsd code? > -- Nathan -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa21191; 16 Nov 2000 12:58 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id LAA03059; Thu, 16 Nov 2000 11:58:22 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 16 Nov 2000 11:58:22 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5BD4@umr-mail02> From: "Neulinger, Nathan R." To: "'Derek Atkins'" Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: RE: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) Date: Thu, 16 Nov 2000 11:58:15 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Oh, never mind... I was still stuck in arla-land... In arla, most of the work is done in user space arlad. -- Nathan > -----Original Message----- > From: Derek Atkins [mailto:warlord@MIT.EDU] > Sent: Thursday, November 16, 2000 11:53 AM > To: Neulinger, Nathan R. > Cc: 'Kelsang Wangden'; Sam Hartman; openafs-info@central.org > Subject: Re: Dynamic root.afs (was Re: [OpenAFS] Directory layout for > new cells) > > > "Neulinger, Nathan R." writes: > > > Why get so complicated? Just have the existing 'afsd' trap > requests for the > > 'root.afs' volume and handle them internally. If you make > this an option - > > people can still choose to continue using a REAL root.afs > volume, or they > > can have afsd build a virtual one instead. > > Huh? afsd is just a call into the kernel, where almost all the work > is done. None of the real AFS work is done by afsd proper. So, how > would "'afsd' trap requests for 'root.afs'" when none of the file > operations pass through the afsd code? > > > -- Nathan > > -derek > > -- > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > Member, MIT Student Information Processing Board (SIPB) > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH > warlord@MIT.EDU PGP key available > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa21208; 16 Nov 2000 13:01 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA13853; Thu, 16 Nov 00 13:00:43 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id NAA14369; Thu, 16 Nov 2000 13:01:14 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id NAA09069; Thu, 16 Nov 2000 13:01:13 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id NAA14781; Thu, 16 Nov 2000 13:01:13 -0500 (EST) To: "Neulinger, Nathan R." Cc: "'Kelsang Wangden'" , Sam Hartman , openafs-info@central.org Subject: Re: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) References: <9DA8D24B915BD1118911006094516EAF053E5BD4@umr-mail02> From: Derek Atkins Date: 16 Nov 2000 13:01:13 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 16 Nov 2000 11:58:15 -0600" Message-Id: Lines: 51 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Ahh. Yea, in that case it would be easier to arla to do this. :) -derek "Neulinger, Nathan R." writes: > Oh, never mind... I was still stuck in arla-land... In arla, most of the > work is done in user space arlad. > > -- Nathan > > > -----Original Message----- > > From: Derek Atkins [mailto:warlord@MIT.EDU] > > Sent: Thursday, November 16, 2000 11:53 AM > > To: Neulinger, Nathan R. > > Cc: 'Kelsang Wangden'; Sam Hartman; openafs-info@central.org > > Subject: Re: Dynamic root.afs (was Re: [OpenAFS] Directory layout for > > new cells) > > > > > > "Neulinger, Nathan R." writes: > > > > > Why get so complicated? Just have the existing 'afsd' trap > > requests for the > > > 'root.afs' volume and handle them internally. If you make > > this an option - > > > people can still choose to continue using a REAL root.afs > > volume, or they > > > can have afsd build a virtual one instead. > > > > Huh? afsd is just a call into the kernel, where almost all the work > > is done. None of the real AFS work is done by afsd proper. So, how > > would "'afsd' trap requests for 'root.afs'" when none of the file > > operations pass through the afsd code? > > > > > -- Nathan > > > > -derek > > > > -- > > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > > Member, MIT Student Information Processing Board (SIPB) > > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH > > warlord@MIT.EDU PGP key available > > -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa21698; 16 Nov 2000 15:21 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id PAA24551; Thu, 16 Nov 2000 15:21:07 -0500 (EST) To: "Neulinger, Nathan R." Cc: "'Derek Atkins'" , "'Kelsang Wangden'" , openafs-info@central.org Subject: Re: Dynamic root.afs (was Re: [OpenAFS] Directory layout for new cells) References: <9DA8D24B915BD1118911006094516EAF053E5BD2@umr-mail02> From: Sam Hartman Date: 16 Nov 2000 15:21:06 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 16 Nov 2000 11:51:05 -0600" Message-ID: Lines: 21 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Neulinger," == Neulinger, Nathan R writes: >> I would certainly prefer not to include DNS directly in the >> kernel module, which implies a callback to some user-space >> process to perform the DNS request. I'm not sure how this >> would work, except, perhaps, another afsd thread that sits >> waiting for 'callbacks' from the kernel? Neulinger,> Why get so complicated? Just have the existing 'afsd' Neulinger,> trap requests for the 'root.afs' volume and handle Neulinger,> them internally. If you make this an option - people Neulinger,> can still choose to continue using a REAL root.afs Neulinger,> volume, or they can have afsd build a virtual one Neulinger,> instead. Even were this possible, you probably want to have support for kernel to callback to userspace. I suspect that RX3 will need to do this anyway, for example, if different servers have different keys. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa21728; 16 Nov 2000 15:29 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id PAA24575; Thu, 16 Nov 2000 15:29:19 -0500 (EST) To: Kelsang Wangden Cc: Derek Atkins , openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <20001116103027.A1707@src.uchicago.edu> From: Sam Hartman Date: 16 Nov 2000 15:29:19 -0500 In-Reply-To: Kelsang Wangden's message of "Thu, 16 Nov 2000 10:30:27 -0600" Message-ID: Lines: 33 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Kelsang" == Kelsang Wangden writes: Kelsang> Sam Hartman wrote: >> >>>>> "Derek" == Derek Atkins writes: >> >> Derek> Well, you could: 0) generate ptdb 1) generate root.afs that Derek> only contains the new cell 2) generate root.cell 3) start Derek> afsd 4) get tokens 5) rebuild root.afs >> As I said earlier, noauth considered harmful. Kelsang> Why do you consider it harmful? It's the recommended way Kelsang> to build a new cell. Recommended by whom? Transarc? Yes, they recommend it because it's all the software they ship supports. Also, note that they (I believe) still recommend using kaserver. Kerberos4 is a bit dated. I consider noauth harmful because it is an unnecessary network exposure that allows anyone who can send packets to your machine to execute arbitrary AFS requests. While bos exec exists, this is a remote root compromise potential. Even without bos exec, it seems significant unnecessary effort to go back and revalidate the cell to make sure that no unauthorized changes happened when you come out of noauth. All you need to avoid ever using noauth is a tool to construct a prdb and to modify it if things get broken. MIT has such a tool. Sadly, it is designed to build against AFS 3.2. Gary Zacheiss has been trying to port to Openafs but is running into some trouble. I suspect he will eventually succeed. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa21804; 16 Nov 2000 15:52 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id PAA25109 for ; Thu, 16 Nov 2000 15:52:22 -0500 (EST) Date: Thu, 16 Nov 2000 15:52:30 -0500 From: Derrick J Brashear cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <173528928.974407950@skittlebrau.trafford.dementia.org> In-Reply-To: X-Mailer: Mulberry/2.0.6a8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Thursday, November 16, 2000 03:29:19 PM -0500 Sam Hartman wrote: > All you need to avoid ever using noauth is a tool to construct a prdb > and to modify it if things get broken. MIT has such a tool. Sadly, > it is designed to build against AFS 3.2. Gary Zacheiss has been > trying to port to Openafs but is running into some trouble. I suspect > he will eventually succeed. I heard a proposal that ptserver honor UserList when system:administrators has no members, which could also deal with this. Comments? -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa21827; 16 Nov 2000 16:09 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id QAA24761; Thu, 16 Nov 2000 16:09:51 -0500 (EST) To: Derrick J Brashear Cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells References: <173528928.974407950@skittlebrau.trafford.dementia.org> From: Sam Hartman Date: 16 Nov 2000 16:09:50 -0500 In-Reply-To: Derrick J Brashear's message of "Thu, 16 Nov 2000 15:52:30 -0500" Message-ID: Lines: 26 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Derrick" == Derrick J Brashear writes: Derrick> --On Thursday, November 16, 2000 03:29:19 PM -0500 Sam Derrick> Hartman wrote: >> All you need to avoid ever using noauth is a tool to construct >> a prdb and to modify it if things get broken. MIT has such a >> tool. Sadly, it is designed to build against AFS 3.2. Gary >> Zacheiss has been trying to port to Openafs but is running into >> some trouble. I suspect he will eventually succeed. Derrick> I heard a proposal that ptserver honor UserList when Derrick> system:administrators has no members, which could also Derrick> deal with this. This mostly works although I'd actually want an option to drop it into this mode in case say some cracker gets into the cell and then removes existing s:a members while adding himself. Where should I look if I want to try and implement this? Also, I assume I'd need to remove the logic that prevents database rebuilding if noauth is not true; will there be any harmful effects from this? From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa21856; 16 Nov 2000 16:18 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id QAA25395; Thu, 16 Nov 2000 16:17:45 -0500 (EST) Date: Thu, 16 Nov 2000 16:17:58 -0500 From: Derrick J Brashear To: Sam Hartman cc: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <303168928.974409478@skittlebrau.trafford.dementia.org> In-Reply-To: X-Mailer: Mulberry/2.0.6a8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Thursday, November 16, 2000 04:09:50 PM -0500 Sam Hartman wrote: >>>>>> "Derrick" == Derrick J Brashear writes: > > Derrick> --On Thursday, November 16, 2000 03:29:19 PM -0500 Sam > Derrick> Hartman wrote: > > >> All you need to avoid ever using noauth is a tool to construct > >> a prdb and to modify it if things get broken. MIT has such a > >> tool. Sadly, it is designed to build against AFS 3.2. Gary > >> Zacheiss has been trying to port to Openafs but is running into > >> some trouble. I suspect he will eventually succeed. > > Derrick> I heard a proposal that ptserver honor UserList when > Derrick> system:administrators has no members, which could also > Derrick> deal with this. > > > This mostly works although I'd actually want an option to drop it into > this mode in case say some cracker gets into the cell and then removes > existing s:a members while adding himself. > > > Where should I look if I want to try and implement this? IIRC there are calls like: isadmin = pr_noAuth || IsAMemberOf (mumble, frotz, SYSADMINID); which presumably would need to be made into a function call instead (with at least mumble, frotz, and pr_noAuth as arguments) to add a check against UserList so it can be done in one place. But I'd have to look to be sure. > assume I'd need to remove the logic that prevents database rebuilding > if noauth is not true; will there be any harmful effects from this? Since the header also needs to be zeroed it shouldn't be a problem. -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa21862; 16 Nov 2000 16:19 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id QAA25430 for ; Thu, 16 Nov 2000 16:19:55 -0500 (EST) Date: Thu, 16 Nov 2000 16:20:09 -0500 From: Derrick J Brashear To: openafs-info@central.org Subject: Re: [OpenAFS] Directory layout for new cells Message-ID: <316348928.974409609@skittlebrau.trafford.dementia.org> In-Reply-To: X-Mailer: Mulberry/2.0.6a8 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > Derrick> I heard a proposal that ptserver honor UserList when > Derrick> system:administrators has no members, which could also > Derrick> deal with this. While system:administrators is the topic, I should mention that I did (simple) code for CMU some time ago which creates a magic group system:voyeurs, which as you might guess is a group which can "see" anything implicitly but does not get modification priviledges, the idea being if you have a service which needs to be able to see pts membership, you put it in this group. It's a trivial patch. Any interest? -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mailer.psc.edu by penn.central.org id aa27398; 27 Nov 2000 11:10 EST Received: from minerva.psc.edu (minerva.psc.edu [128.182.62.121]) by mailer.psc.edu (8.9.3/8.9.3/psc) with ESMTP id LAA09957 for ; Mon, 27 Nov 2000 11:10:38 -0500 (EST) Received: from localhost (ecf@localhost) by minerva.psc.edu (8.9.3/8.9.3/psc) with SMTP id LAA29281 for ; Mon, 27 Nov 2000 11:10:37 -0500 (EST) Message-Id: <200011271610.LAA29281@minerva.psc.edu> X-Authentication-Warning: minerva.psc.edu: ecf@localhost didn't use HELO protocol X-Mailer: exmh version 2.0.2 2/24/98 To: openafs-info@central.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 27 Nov 2000 11:10:37 -0500 From: Esther Filderman Subject: [OpenAFS] AFS workshop @ LISA 2000 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: [Please forgive the multiple copies, for those on multiple mailing lists.] Hey, folks, if you're attending LISA 2000 this year, there's still some room left in the AFS workshop, to be held December 5th (Tuesday). The workshop is FREE with conference registration. Application to the workshop is a message to afs-workshop@usenix.org listing some topics you consider most important to be addressed at the workshop. Thanks, Esther Filderman From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from squirt.sigmatel.com by penn.central.org id aa29550; 27 Nov 2000 21:07 EST Received: from spectre.sigmatel.com (spectre [192.168.7.80]) by mcfeely.sigmatel.com (8.10.2/8.10.2) with ESMTP id eAS27Nu27386 for ; Mon, 27 Nov 2000 20:07:31 -0600 Date: Mon, 27 Nov 2000 20:07:25 -0600 (CST) From: "Patrick J. Hennessey" To: openafs-info@WWW-OPENAFS.CENTRAL.ORG Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] SUCCESS: OpenAFS kernel module for Linux 2.4.0-test11 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I was able to compile OpenAFS by commenting out the 'venus' directory and its dependencies, as well as chaning the variable mentioned in src/README from LINUX_VERSION to LINUX_VERS (i.e. make SYS_NAME="i386_linux24" LINUX_VERSION="2.4.0-test11" This compilation was made with the patches by Chas Williams available at ftp://ftp.cmf.nrl.navy.mil/pub/chas/linux-openafs/symlink-fix.diff This compilation produced an SMP kernel module that insmod'd right into my server without a crash. It can be removed from memory as well. I intend to startup a cell soon to truly test this system, but I consider this a succss. (info: gcc 2.95-2 under Mandrake 7.2, Linux kernel 2.4.0-test11) -- Patrick J. Hennessey spectre@sigmatel.com Information Solutions (512) 381-3722 SigmaTel, Inc. www.sigmatel.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa04031; 30 Nov 2000 15:09 EST Received: from JOHNSTOWN.ANDREW.CMU.EDU (JOHNSTOWN.ANDREW.CMU.EDU [128.2.121.156]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id PAA11148 for ; Thu, 30 Nov 2000 15:09:28 -0500 (EST) Date: Thu, 30 Nov 2000 15:09:26 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@johnstown.andrew.cmu.edu cc: openafs-info@OPENAFS.ORG Subject: Re: [OpenAFS] SUCCESS: OpenAFS kernel module for Linux 2.4.0-test11 In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Mon, 27 Nov 2000, Patrick J. Hennessey wrote: > I was able to compile OpenAFS by commenting out the 'venus' directory and > its dependencies, as well as chaning the variable mentioned in src/README > from LINUX_VERSION to LINUX_VERS (i.e. > make SYS_NAME="i386_linux24" LINUX_VERSION="2.4.0-test11" > > This compilation was made with the patches by Chas Williams available at > ftp://ftp.cmf.nrl.navy.mil/pub/chas/linux-openafs/symlink-fix.diff The OpenAFS CVS tree (penn.central.org:/cvs, user anonymous, password anonymous, module openafs) has these fixes applied, and will shortly have a fix for the kdump compilation problem. -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa05990; 5 Dec 2000 21:38 EST Received: from JOHNSTOWN.ANDREW.CMU.EDU (JOHNSTOWN.ANDREW.CMU.EDU [128.2.121.156]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id VAA25595; Tue, 5 Dec 2000 21:38:38 -0500 (EST) Date: Tue, 5 Dec 2000 21:38:38 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@johnstown.andrew.cmu.edu To: openafs-info@openafs.org cc: openafs-announce@openafs.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] OpenAFS 1.0.1 release available Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: An OpenAFS 1.0.1 release is available in source form, and as Debian (Potato) and Redhat (6.2) RPMS. You can always find the latest OpenAFS release at http://www.openafs.org/release/latest.html The code has been built and verified on Debian Potato, Redhat 6.2 and 7.0, and Solaris 7. Solaris 7 binaries and RedHat 7.0 RPMS will be available in a few days. Please send bug reports to openafs-gatekeepers@openafs.org From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from host2.marvell.com by penn.central.org id aa06268; 5 Dec 2000 22:26 EST Received: from flash.marvell.com by marvell.com (SMI-8.6/SMI-SVR4) id TAA24763; Tue, 5 Dec 2000 19:25:31 -0800 Received: from flash by flash.marvell.com (8.8.8+Sun/SMI-SVR4) id TAA03931; Tue, 5 Dec 2000 19:25:36 -0800 (PST) Message-Id: <200012060325.TAA03931@flash.marvell.com> Date: Tue, 5 Dec 2000 19:25:36 -0800 (PST) From: Phuoc Ly Reply-To: Phuoc Ly To: openafs-info@openafs.org MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: H1B9s2VT/tzS1zZ7OyaAjQ== X-Mailer: dtmail 1.2.1 CDE Version 1.2.1 SunOS 5.6 sun4u sparc Subject: [OpenAFS] openafs 1.0.1 Solaris 5.7 compile error Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I got the following error when I try to compile openafs 1.0.1 on Solaris 5.7. I used bison and flex since I don't have lex and yacc. Anyone successfully build openafs on solaris? Please help. Below is the last part of the 'make SYS_NAME="sun4x_57"' command: ... cd obj; cd comerr ; make install SYS_NAME=sun4x_57 DESTDIR=/usr/local/src/openafs-1.0.1/sun4x_57/dest/ SRCDIR=/usr/local/src/openafs-1.0.1/sun4x_57/dest/ make[3]: Entering directory `/usr/local/src/openafs-1.0.1/sun4x_57/obj/comerr' /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic error_msg.c /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic et_name.c /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic com_err.c echo 'char cml_version_number[]="@(#)CML not accessible: No version Information";' >AFS_component_version_number.c; echo 'char* AFSVersion = "afs??"; ' >>AFS_component_version_number.c; /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic AFS_component_version_number.c rm -f libcom_err.a /usr/ccs/bin/ar cruv libcom_err.a error_msg.o et_name.o com_err.o AFS_component_version_number.o a - error_msg.o a - et_name.o a - com_err.o a - AFS_component_version_number.o ar: writing libcom_err.a /bin/true libcom_err.a /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic compile_et.c yacc error_table.y mv -f y.tab.c error_table.c rm -f et_lex.lex.c lex -t et_lex.lex.l >et_lex.lex.c /opt/SUNWspro/bin/cc -c -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic error_table.c case sun4x_57 in \ *_linux* ) \ /opt/SUNWspro/bin/cc -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic -o compile_et compile_et.o error_table.o /usr/local/src/openafs-1.0.1/sun4x_57/dest/lib/afs/util.a -lsocket -lnsl -lintl -ldl;; \ * ) \ /opt/SUNWspro/bin/cc -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy -Bdynamic -o compile_et compile_et.o error_table.o /usr/local/src/openafs-1.0.1/sun4x_57/dest/lib/afs/util.a -lsocket -lnsl -lintl -ldl -ll;; \ esac Undefined first referenced symbol in file yylineno compile_et.o ld: fatal: Symbol referencing errors. No output written to compile_et make[3]: *** [compile_et] Error 1 make[3]: Leaving directory `/usr/local/src/openafs-1.0.1/sun4x_57/obj/comerr' make[2]: *** [comerr] Error 2 make[2]: Leaving directory `/usr/local/src/openafs-1.0.1' make[1]: *** [install] Error 2 make[1]: Leaving directory `/usr/local/src/openafs-1.0.1' make: *** [all] Error 2 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ZALON.REM.CS.CMU.EDU by penn.central.org id aa06630; 5 Dec 2000 23:26 EST Received: from zalon.pc.cs.cmu.edu by zalon.pc.cs.cmu.edu id aa01154; 5 Dec 2000 23:25 EST Date: Tue, 5 Dec 2000 23:25:38 -0500 (EST) From: Jeffrey Hutzelman X-Sender: jhutz@zalon.pc.cs.cmu.edu To: Phuoc Ly cc: openafs-info@openafs.org Subject: Re: [OpenAFS] openafs 1.0.1 Solaris 5.7 compile error In-Reply-To: <200012060325.TAA03931@flash.marvell.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Tue, 5 Dec 2000, Phuoc Ly wrote: > I got the following error when I try to compile openafs 1.0.1 on Solaris 5.7. > I used bison and flex since I don't have lex and yacc. Solaris includes lex and yacc as standard utilities, as do most UNIX systems. On Solaris, they are part of the SUNWbtool package, which includes various tools for building software. They can be found in /usr/ccs/bin. -- Jeffrey T. Hutzelman (N3NHS) Sr. Research Systems Programmer School of Computer Science - Research Computing Facility Carnegie Mellon University - Pittsburgh, PA From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx602-mta.mail.com by penn.central.org id aa18901; 6 Dec 2000 8:18 EST Received: from web394-mc (web394-mc.mail.com [165.251.48.135]) by rmx602-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA11916 for ; Wed, 6 Dec 2000 08:18:12 -0500 (EST) Message-ID: <380602730.976108688156.JavaMail.root@web394-mc> Date: Wed, 6 Dec 2000 08:18:08 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa18943; 6 Dec 2000 8:19 EST Received: from web569-mc (web569-mc.mail.com [165.251.48.92]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA20893 for ; Wed, 6 Dec 2000 08:19:09 -0500 (EST) Message-ID: <385468490.976108698098.JavaMail.root@web569-mc> Date: Wed, 6 Dec 2000 08:18:18 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa18945; 6 Dec 2000 8:19 EST Received: from web569-mc (web569-mc.mail.com [165.251.48.92]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA20956 for ; Wed, 6 Dec 2000 08:19:11 -0500 (EST) Message-ID: <385391283.976108701559.JavaMail.root@web569-mc> Date: Wed, 6 Dec 2000 08:18:21 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa18947; 6 Dec 2000 8:19 EST Received: from web105-mc (web105-mc.mail.com [165.251.48.93]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA20974 for ; Wed, 6 Dec 2000 08:19:12 -0500 (EST) Message-ID: <383425835.976108702080.JavaMail.root@web105-mc> Date: Wed, 6 Dec 2000 08:18:22 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa18949; 6 Dec 2000 8:19 EST Received: from web582-mc (web582-mc.mail.com [165.251.48.95]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA20986 for ; Wed, 6 Dec 2000 08:19:12 -0500 (EST) Message-ID: <384616882.976108702267.JavaMail.root@web582-mc> Date: Wed, 6 Dec 2000 08:18:22 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa19002; 6 Dec 2000 8:20 EST Received: from web577-mc (web577-mc.mail.com [165.251.48.102]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA22289 for ; Wed, 6 Dec 2000 08:20:03 -0500 (EST) Message-ID: <382170596.976108696873.JavaMail.root@web577-mc> Date: Wed, 6 Dec 2000 08:18:16 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx195-mta.mail.com by penn.central.org id aa19006; 6 Dec 2000 8:20 EST Received: from web349-mc (web349-mc.mail.com [165.251.48.91]) by rmx195-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA10185 for ; Wed, 6 Dec 2000 08:20:06 -0500 (EST) Message-ID: <386133299.976108706542.JavaMail.root@web349-mc> Date: Wed, 6 Dec 2000 08:18:26 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx602-mta.mail.com by penn.central.org id aa18951; 6 Dec 2000 8:19 EST Received: from web489-mc (web489-mc.mail.com [165.251.48.99]) by rmx602-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA13610 for ; Wed, 6 Dec 2000 08:19:17 -0500 (EST) Message-ID: <382968445.976108654152.JavaMail.root@web489-mc> Date: Wed, 6 Dec 2000 08:17:34 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa18953; 6 Dec 2000 8:19 EST Received: from web582-mc (web582-mc.mail.com [165.251.48.95]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id IAA21100 for ; Wed, 6 Dec 2000 08:19:16 -0500 (EST) Message-ID: <384494145.976108705092.JavaMail.root@web582-mc> Date: Wed, 6 Dec 2000 08:18:25 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Using OpenAFS for implementing mail servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx195-mta.mail.com by penn.central.org id aa08056; 7 Dec 2000 0:40 EST Received: from web651-mc (web651-mc.mail.com [165.251.48.100]) by rmx195-mta.mail.com (8.9.3/8.9.3) with SMTP id AAA26953 for ; Thu, 7 Dec 2000 00:40:47 -0500 (EST) Message-ID: <384510283.976167646520.JavaMail.root@web651-mc> Date: Thu, 7 Dec 2000 00:40:46 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Subject: RE: [OpenAFS] Using OpenAFS for implementing mail servers Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: people, my apologies for a flood of the following message, as the mail interface had stuck, and it managed to send off so many mail. Jatin ------Original Message------ From: Jatin Nansi To: openafs-info@openafs.org Sent: December 6, 2000 1:17:34 PM GMT Subject: [OpenAFS] Using OpenAFS for implementing mail servers hii we are planning to implement a clustered mail server, having upto 4 physical servers providing mail (smtp, pop3) and web mail (http) to our users. for this we need to keep the filesystems synchronised amongst all the machines, as the user may land up at any server, and his data better be there at that server. the clusetering solution would be something similar to piranha or turbocluster. this is a round robin load balancing solution. all servers are expected to perform high file i/o. since the volume of the mail is very high at our site, typically upto a GB a day, is it possible for me to use openafs for keeping all the servers synced, if not what else can i use for this purpose? please suggest if it would be correct to use openafs here. thank you Jatin Nansi ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com _______________________________________________ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo.cgi/openafs-info ________________________________________ LINUX - Its a way of life ________________________________________ ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa13334; 7 Dec 2000 8:29 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id IAA14878; Thu, 7 Dec 2000 08:29:48 -0500 (EST) Date: Thu, 7 Dec 2000 08:29:48 -0500 (EST) Message-Id: <200012071329.IAA14878@sweet-transvestite.mit.edu> From: Sam Hartman To: openafs-info@openafs.org Subject: [OpenAFS] New Debian packages available Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I realized Tuesday that my Debian packages had a bug in the handling of bos restart -bosserver. They were looking for bosserver in /usr/lib/openafs/bin instead of /usr/sbin. I have fixed this bug and built new versions of Debian 2.2 packages available from the openafs.org website. Also, I have uploaded packages for inclusion in woody to the Debian non-us incoming directory. These packages include the Openafs packages, as well as openafs-krb5, which includes aklog. These packages should be available in woody as soon as the overide file is updated--perhaps a week or so. --Sam From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from windlord.Stanford.EDU by penn.central.org id aa18391; 7 Dec 2000 14:14 EST Received: (qmail 5354 invoked by uid 50); 7 Dec 2000 19:14:51 -0000 To: Jatin Nansi Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers References: <384616882.976108702267.JavaMail.root@web582-mc> In-Reply-To: Jatin Nansi's message of "Wed, 6 Dec 2000 08:18:22 -0500 (EST)" From: Russ Allbery Organization: The Eyrie Date: 07 Dec 2000 11:14:51 -0800 Message-ID: Lines: 32 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Jatin Nansi writes: > we are planning to implement a clustered mail server, having upto 4 > physical servers providing mail (smtp, pop3) and web mail (http) to our > users. for this we need to keep the filesystems synchronised amongst all > the machines, as the user may land up at any server, and his data better > be there at that server. the clusetering solution would be something > similar to piranha or turbocluster. this is a round robin load balancing > solution. all servers are expected to perform high file i/o. > since the volume of the mail is very high at our site, typically upto a > GB a day, is it possible for me to use openafs for keeping all the > servers synced, I've yet to hear from someone who's had good experiences doing high-volume mail into AFS. It's possible to deliver mail into AFS, but it's very slow according to everyone I know who's tried it. You run into a few problems; locking inside AFS isn't something you generally want to use. The write performance is also pretty annoying. > if not what else can i use for this purpose? Probably the most common solution that I've seen for this particular problem is to put the mail on a NetApp machine or something equivalent and then make sure that it and all your servers speak the same variety of NFS locking. We take a different approach and assign each user to one and only one POP server, but use DNS to give them names of the form username.pobox.stanford.edu so that we can move things between POP servers if necessary. -- Russ Allbery (rra@stanford.edu) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa18752; 7 Dec 2000 14:32 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id NAA29060; Thu, 7 Dec 2000 13:31:54 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 7 Dec 2000 13:31:54 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5C85@umr-mail02> From: "Neulinger, Nathan R." To: "'Russ Allbery'" , Jatin Nansi Cc: openafs-info@openafs.org Subject: RE: [OpenAFS] Using OpenAFS for implementing mail servers Date: Thu, 7 Dec 2000 13:31:48 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: If you're going to do it, I recommend that you use Maildir format spools. That eliminates the need for locking entirely. It also lets you deliver into the same spool simultaneously from multiple servers at the same time. Layout would be: [afs servers 1-x ] <-> [mail servers 1-x] Your mail servers could all be 100% identical - that's one significant benefit of using AFS for the mail delivery. Just DNS rotate amongst them (or load balance if you like). Be sure and put really big caches on the mail servers. -- Nathan > -----Original Message----- > From: Russ Allbery [mailto:rra@stanford.edu] > Sent: Thursday, December 07, 2000 1:15 PM > To: Jatin Nansi > Cc: openafs-info@openafs.org > Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers > > > Jatin Nansi writes: > > > we are planning to implement a clustered mail server, having upto 4 > > physical servers providing mail (smtp, pop3) and web mail > (http) to our > > users. for this we need to keep the filesystems > synchronised amongst all > > the machines, as the user may land up at any server, and > his data better > > be there at that server. the clusetering solution would be something > > similar to piranha or turbocluster. this is a round robin > load balancing > > solution. all servers are expected to perform high file i/o. > > > since the volume of the mail is very high at our site, > typically upto a > > GB a day, is it possible for me to use openafs for keeping all the > > servers synced, > > I've yet to hear from someone who's had good experiences > doing high-volume > mail into AFS. It's possible to deliver mail into AFS, but > it's very slow > according to everyone I know who's tried it. You run into a > few problems; > locking inside AFS isn't something you generally want to use. > The write > performance is also pretty annoying. > > > if not what else can i use for this purpose? > > Probably the most common solution that I've seen for this particular > problem is to put the mail on a NetApp machine or something > equivalent and > then make sure that it and all your servers speak the same > variety of NFS > locking. We take a different approach and assign each user to one and > only one POP server, but use DNS to give them names of the form > username.pobox.stanford.edu so that we can move things > between POP servers > if necessary. > > -- > Russ Allbery (rra@stanford.edu) _______________________________________________ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo.cgi/openafs-info From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from windlord.Stanford.EDU by penn.central.org id aa18820; 7 Dec 2000 14:35 EST Received: (qmail 5462 invoked by uid 50); 7 Dec 2000 19:35:06 -0000 To: "Neulinger, Nathan R." Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers References: <9DA8D24B915BD1118911006094516EAF053E5C85@umr-mail02> In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 7 Dec 2000 13:31:48 -0600" From: Russ Allbery Organization: The Eyrie Date: 07 Dec 2000 11:35:06 -0800 Message-ID: Lines: 15 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Neulinger, Nathan R writes: > Be sure and put really big caches on the mail servers. I'd recommend the exact opposite. With mail, you're dealing with a huge number of really small files, particularly with maildir deliveries (which I'd also recommend). That's the worst case scenario for a large cache; you spend all your time searching the cache and your performance goes to hell. I'd instead try a smaller-sized (maybe 100MB to 200MB) memory cache (you definitely want a memory cache, not a disk cache). -- Russ Allbery (rra@stanford.edu) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa19008; 7 Dec 2000 14:45 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id OAA06741; Thu, 7 Dec 2000 14:45:06 -0500 To: Russ Allbery Cc: Jatin Nansi , openafs-info@openafs.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers References: <384616882.976108702267.JavaMail.root@web582-mc> From: Derek Atkins Date: 07 Dec 2000 14:45:05 -0500 In-Reply-To: Russ Allbery's message of "07 Dec 2000 11:14:51 -0800" Message-ID: Lines: 20 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Russ Allbery writes: > locking. We take a different approach and assign each user to one and > only one POP server, but use DNS to give them names of the form > username.pobox.stanford.edu so that we can move things between POP servers > if necessary. MIT does a similar thing (except MIT uses Hesiod instead of raw DNS for the pobox lookup). I would not recommend using AFS for mail storage (except by the end user). Use POP or IMAP for users to connect, and just load-balance the machines by splitting which users use which servers. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from e4.ny.us.ibm.com by penn.central.org id aa23089; 6 Dec 2000 10:38 EST Received: from northrelay02.pok.ibm.com (northrelay02.pok.ibm.com [9.117.200.22]) by e4.ny.us.ibm.com (8.9.3/8.9.3) with ESMTP id KAA407436 for ; Wed, 6 Dec 2000 10:37:25 -0500 Received: from us.ibm.com (stevefx.rchland.ibm.com [9.5.66.42]) by northrelay02.pok.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id KAA149398 for ; Wed, 6 Dec 2000 10:36:20 -0500 Message-ID: <3A2E5E08.DCEF12B0@us.ibm.com> Date: Wed, 06 Dec 2000 09:40:56 -0600 From: Steve Fox X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.16-9mdk i686) X-Accept-Language: en MIME-Version: 1.0 To: openafs-info@openafs.org Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] Troubles with compile on Mandrake Linux 7.1 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Greetings, I'm unable to compile OpenAFS 1.01 on Mandrake Linux 7.1 because it is looking for a UTS_RELEASE string in the version.h header which apparently is not in the Mandrake kernel source (installed via the kernel-source rpm). The exact error is below: ERROR: Cannot build for Linux kernel 2.2.16: No UTS_RELEASE string found in /usr/src/linux-2.2.16/include/linux/version.h. ERROR: Should be able to build at least one of 2.2.15 2.2.16. Valid headers not present for any Linux kernel. make[3]: *** [lkversion] Error 255 Can anyone provide me assistance with figuring out why the UTS_RELEASE string would not be declared? Thanks much. P.S. If you don't mind could you please CC me as I'm not subscribed to the list. Thanks. -- Steve Fox IBM Linux Technology Center http://oss.software.ibm.com/developerworks/opensource/linux/ http://stevefx.rchland.ibm.com (IBM intranet) http://k-lug.com (Rochester Linux Users Group) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from DAEMON-UNPLUGGED.MIT.EDU by penn.central.org id aa05146; 6 Dec 2000 20:21 EST Received: from amu by daemon with local (Exim 3.20 #1 (Debian)) id 143pkL-0002E0-00; Wed, 06 Dec 2000 20:21:17 -0500 To: Phuoc Ly Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] openafs 1.0.1 Solaris 5.7 compile error From: amu@mit.edu (Aaron M. Ucko) Date: 06 Dec 2000 20:21:17 -0500 In-Reply-To: Phuoc Ly's message of "Tue, 5 Dec 2000 19:25:36 -0800 (PST)" Message-ID: <87puj5xbzm.fsf@mit.edu> Lines: 21 X-Mailer: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: (Phuoc Ly) writes: > I got the following error when I try to compile openafs 1.0.1 on Solaris 5.7. > I used bison and flex since I don't have lex and yacc. [...] > /opt/SUNWspro/bin/cc -O -I/usr/local/src/openafs-1.0.1/sun4x_57/dest/include -dy > -Bdynamic -o compile_et compile_et.o error_table.o > /usr/local/src/openafs-1.0.1/sun4x_57/dest/lib/afs/util.a -lsocket -lnsl -lintl > -ldl -ll;; \ > esac > Undefined first referenced > symbol in file > yylineno compile_et.o If you're using flex instead of lex, you need to link against -lfl instead of -ll. -- Aaron M. Ucko, KB1CJC (finger amu@monk.mit.edu) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from e1.ny.us.ibm.com by penn.central.org id aa20798; 7 Dec 2000 16:26 EST Received: from northrelay02.pok.ibm.com (northrelay02.pok.ibm.com [9.117.200.22]) by e1.ny.us.ibm.com (8.9.3/8.9.3) with ESMTP id QAA111852; Thu, 7 Dec 2000 16:25:32 -0500 Received: from RCHASA28.RCHLAND.IBM.COM (d27ml104.rchland.ibm.com [9.5.39.61]) by northrelay02.pok.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id QAA126818; Thu, 7 Dec 2000 16:24:25 -0500 Importance: Normal Subject: Re: [OpenAFS] Troubles with compile on Mandrake Linux 7.1 To: Corey Kovacs Cc: openafs-info@openafs.org X-Mailer: Lotus Notes Release 5.0.4a July 24, 2000 Message-ID: From: "Steve Fox" Date: Thu, 7 Dec 2000 15:27:46 -0600 X-MIMETrack: Serialize by Router on d27ml104/27/M/IBM(Release 5.0.5 |September 22, 2000) at 12/07/2000 03:27:48 PM MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: That fixed it up! Thank you. The problem I'm having now is that OpenAFS is looking for pam_appl.h which is not included in Mandrake's PAM packages. I will go yell at them ;) but I thought I'd document it here in case anyone else is searching the archives. Steve Fox IBM Linux Technology Center http://oss.software.ibm.com/developerworks/opensource/linux http://stevefx.rchland.ibm.com (IBM Intranet) To: Steve Fox/Rochester/IBM@IBMUS cc: Subject: Re: [OpenAFS] Troubles with compile on Mandrake Linux 7.1 Just a thought, did you do a "make dep" in the linux kernel source tree? Steve Fox wrote: > Greetings, > > I'm unable to compile OpenAFS 1.01 on Mandrake Linux 7.1 because it is > looking for a UTS_RELEASE string in the version.h header which > apparently is not in the Mandrake kernel source (installed via the > kernel-source rpm). The exact error is below: > > ERROR: Cannot build for Linux kernel 2.2.16: > No UTS_RELEASE string found in > /usr/src/linux-2.2.16/include/linux/version.h. > ERROR: Should be able to build at least one of 2.2.15 2.2.16. > Valid headers not present for any Linux kernel. > make[3]: *** [lkversion] Error 255 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx325-mta.mail.com by penn.central.org id aa25498; 8 Dec 2000 0:29 EST Received: from web105-mc (web105-mc.mail.com [165.251.48.93]) by rmx325-mta.mail.com (8.9.3/8.9.3) with SMTP id AAA16689; Fri, 8 Dec 2000 00:29:11 -0500 (EST) Message-ID: <382548819.976253301959.JavaMail.root@web105-mc> Date: Fri, 8 Dec 2000 00:28:21 -0500 (EST) From: Jatin Nansi To: Derek Atkins , Russ Allbery Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers Cc: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hi thanks for the information. but unfortunately, our corporate policy here is that the users must get username@domainname.com, thats it, nothing else. and i still need to load balance, and ensure that the mail folders on all servers are in sync at all times. so now how do i do it?? Jatin ------Original Message------ From: Derek Atkins To: Russ Allbery Sent: December 7, 2000 7:45:05 PM GMT Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers Russ Allbery writes: > locking. We take a different approach and assign each user to one and > only one POP server, but use DNS to give them names of the form > username.pobox.stanford.edu so that we can move things between POP servers > if necessary. MIT does a similar thing (except MIT uses Hesiod instead of raw DNS for the pobox lookup). I would not recommend using AFS for mail storage (except by the end user). Use POP or IMAP for users to connect, and just load-balance the machines by splitting which users use which servers. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH warlord@MIT.EDU PGP key available ________________________________________ LINUX - Its a way of life ________________________________________ ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx195-mta.mail.com by penn.central.org id aa25960; 8 Dec 2000 2:43 EST Received: from web582-mc (web582-mc.mail.com [165.251.48.95]) by rmx195-mta.mail.com (8.9.3/8.9.3) with SMTP id CAA09907 for ; Fri, 8 Dec 2000 02:43:39 -0500 (EST) Message-ID: <384366799.976261370823.JavaMail.root@web582-mc> Date: Fri, 8 Dec 2000 02:42:50 -0500 (EST) From: Jatin Nansi To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-WM-FaxTo: X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Subject: [OpenAFS] Problem during implementing AFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hi i have been trying to implement the Open AFS file server on a RedHat Linux 7.0 kernel version 2.2.16-22 on a PIII, with 256 MB RAM as a trial machine. i have downloaded the rpms from the openafs site for the same. i am using the procedure as described in the QkBegin guide that is there in the ibm documentation tarball. the problem is as follows: Altough the kernel module complains, it does load properly. the bosserver also starts properly, on the command: #bosserver -noauth & the 'bos setcellname ' command also works fine, and i am also able to add the pc to the cell. bos listhosts confirms this also. the next step as per the guide should be for starting the database server: the command: bos create node1 kaserver simple \ /usr/afs/bin/kaserver -cell timesgroup.com -noauth this creates the following entry in /usr/afs/local/BosConfig: ------------------ restarttime 11 0 4 0 0 checkbintime 3 0 5 0 0 bnode simple kaserver 1 parm /usr/afs/bin/kaserver end -------------------- however there is no kaserver process running when checked with ps xa similarly, the commands for buserver, ptserver, and vlserver do make entries into BosConfig, but no processes start. then i tried killall bosserver, which killed the boserver process. when i start it again with bosserver -noauth&, i get the following error: /usr/afs/bin/buserver: problems with host name Ubik init failed ........... repeated several times over for all other servers also. the contents of files in /usr/afs/logs: AuthLog-- /usr/afs/bin/kaserver: problems with host name Ubik init failed ll database. Using level crypt for Ubik connections. ubik: primary address 127.0.0.1 does not exist BackupLog-- u: problem with host name Ubik init failed Boslog-- Server directory access is okay BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. ........similarly for all others ----------- Please tell me where am i going wrong in this, as i am following the ibm guide to the letter. thank you Jatin ________________________________________ LINUX - Its a way of life ________________________________________ ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.mirapoint.com by penn.central.org id aa27990; 8 Dec 2000 5:21 EST Received: from tim-bsd.mirapoint.com (tim-bsd.mirapoint.com [192.168.0.117]) by mail.mirapoint.com (Mirapoint) with SMTP id AAV09199; Fri, 8 Dec 2000 02:21:34 -0800 (PST) X-Spook: clones White Water assassination Kennedy spy DES Noriega To: Jatin Nansi Cc: Derek Atkins , Russ Allbery , openafs-info@openafs.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers References: <382548819.976253301959.JavaMail.root@web105-mc> From: Tim Showalter Date: 08 Dec 2000 02:22:50 -0800 In-Reply-To: Jatin Nansi's message of "Fri, 8 Dec 2000 00:28:21 -0500 (EST)" Message-ID: <7dhf4f2ow5.fsf@tim-bsd.mirapoint.com> Lines: 39 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Canyonlands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Jatin Nansi writes: > > > locking. We take a different approach and assign each user to one > > > and only one POP server, but use DNS to give them names of the > > > form username.pobox.stanford.edu so that we can move things > > > between POP servers if necessary. > > > > MIT does a similar thing (except MIT uses Hesiod instead of raw DNS > > for the pobox lookup). I would not recommend using AFS for mail > > storage (except by the end user). Use POP or IMAP for users to > > connect, and just load-balance the machines by splitting which users > > use which servers. > thanks for the information. but unfortunately, our corporate policy > here is that the users must get username@domainname.com, thats it, > nothing else. and i still need to load balance, and ensure that the > mail folders on all servers are in sync at all times. so now how do i > do it?? I don't see how the configuration that has been suggested by Russ and Derek conflicts with the idea of having only a single email address. Have your MX machines route to userid.pop.domain.com based on userid (most large sites seem to be using LDAP now). This technichally means that userid@userid.pop.domain.com is a valid email address as well as userid@domain.com, but I doubt that's an issue, since you probably have a firewall and your POP server is probably on the inside of it. (If this isn't true, don't publish the email addresses and no one will notice -- I know of a few sites using this approach and it seems to have worked well for them). I will refrain from plugging my company's (Mirapoint) fine, high-quality mail server products. I'd also like to not plug IMAP at this point, although I'll note that it's difficult to get most mail clients to read AFS (that idea never really got out of CMU and it's been dying a long time there). Tim From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dsl254-116-177-nyc1.dsl-isp.net by penn.central.org id aa00195; 8 Dec 2000 8:34 EST Received: from o-o.yi.org (IDENT:def@localhost.localdomain [127.0.0.1]) by oo.yi.org (8.9.3/8.8.7) with ESMTP id IAA29148; Fri, 8 Dec 2000 08:33:16 -0500 Message-Id: <200012081333.IAA29148@oo.yi.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Tim Showalter , openafs-info@openafs.org From: aeneous@speakeasy.org Reply-To: aeneous@speakeasy.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers In-Reply-To: Your message of "08 Dec 2000 02:22:50 PST." <7dhf4f2ow5.fsf@tim-bsd.mirapoint.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 08 Dec 2000 08:33:16 -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Tim said: > I will refrain from plugging my company's (Mirapoint) fine, high-quality > mail server products. I'd also like to not plug IMAP at this point, > although I'll note that it's difficult to get most mail clients to read > AFS (that idea never really got out of CMU and it's been dying a long > time there). Mail clients will read AFS just fine - where did you get an idea that was a problem? Perhaps you meant that it's difficult to get clients to read AMDS-style mail directories? The big problem is that mail delivery (and reading) exhibits very poor locality of reference, which is contrary to AFS' design assumptions. Mail storage in AFS would be much more reasonable if you had a technique for writing directly to the server without trying to go through the local cache filesystem, and you used that for final delivery. For an enterprise mail system (though perhaps not the one at hand), the "no matter where you go, there you are" nature of AFS makes handling stored email much more convenient than IMAP does -- especially if one has several hosts that one moves between routinely. This observation may not apply to Jatin Nansi's situation, though. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from firewall1.Lehman.COM by penn.central.org id aa00805; 8 Dec 2000 9:02 EST Received: from relay.lehman.com (relay.Lehman.COM [146.127.39.20]) by firewall1.lehman.com (Postfix) with ESMTP id 82C181177E for ; Fri, 8 Dec 2000 09:01:59 -0500 (EST) Received: from unixeng1.lehman.com by relay.lehman.com (8.9.3/8.8.5) id JAA16307; Fri, 8 Dec 2000 09:01:54 -0500 (EST) Received: from mu.lehman.com by unixeng1.lehman.com (8.8.8+Sun/Lehman Bros. V1.5) id JAA28974; Fri, 8 Dec 2000 09:01:44 -0500 (EST) Received: by mu.lehman.com (8.9.3+Sun/Lehman Bros. V1.5) id JAA21419; Fri, 8 Dec 2000 09:01:18 -0500 (EST) Date: Fri, 8 Dec 2000 09:01:18 -0500 (EST) From: ajf@lehman.COM (Art Freeman) Message-Id: <200012081401.JAA21419@mu.lehman.com> To: openafs-info@openafs.org Subject: [OpenAFS] Using VCS with AFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Has anyone used Veritas' VCS to provide a two node read/write AFS file server cluster? I've been able to get this work in my tests but I'd like to know if anyone has this up and running in production. Also, I'd like to connect my AFS file servers to the EMC symmetrix. Are there any "gotchas" or issues that I need to be aware of? Thanks. --Art Freeman From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from bodhi.src.uchicago.edu by penn.central.org id aa01155; 8 Dec 2000 9:59 EST Received: (from wngdn@localhost) by bodhi.src.uchicago.edu (8.9.3/8.9.3/Debian 8.9.3-21) id IAA08453; Fri, 8 Dec 2000 08:59:48 -0600 From: Kelsang Wangden Date: Fri, 8 Dec 2000 08:59:30 -0600 To: Jatin Nansi Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem during implementing AFS Message-ID: <20001208085930.A8419@src.uchicago.edu> References: <384366799.976261370823.JavaMail.root@web582-mc> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <384366799.976261370823.JavaMail.root@web582-mc>; from linuxguru@india.com on Fri, Dec 08, 2000 at 02:42:50AM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi Jatin, Creating new cells is not my forte, since I have only had to do it once, but I do have one tip. > i have been trying to implement the Open AFS file server on a > RedHat Linux 7.0 kernel version 2.2.16-22 on a PIII, with 256 MB RAM as a > trial machine. [...] > i am using the procedure as described in the QkBegin guide > that is there in the ibm documentation tarball. > > the problem is as follows: > > Altough the kernel module complains, it does load properly. > the bosserver also starts properly, on the command: > > #bosserver -noauth & > the 'bos setcellname ' command also works fine, and i am also able to add > the pc to the cell. bos listhosts confirms this also. > > > the next step as per the guide should be for starting the database server: > > the command: bos create node1 kaserver simple \ > /usr/afs/bin/kaserver -cell timesgroup.com -noauth [...] > however there is no kaserver process running when checked with > ps xa > > similarly, the commands for buserver, ptserver, and vlserver do make entries > into BosConfig, but no processes start. The canonical way of checking for these processes is 'bos status servername'. If you're in the middle of your setup process you may need to add a '-noauth' flag. I wouldn't be surprised if by some wierdness they didn't get allocated as separate processes right away. Regards, Wangden -- Kelsang Wangden (Buddhist monk) Technical Manager Social Science Research Computing, University of Chicago wngdn@src.uchicago.edu (773) 702-3792 PGP key: http://www.src.uchicago.edu/~wngdn/key.txt From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa01262; 8 Dec 2000 10:10 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id KAA07576; Fri, 8 Dec 2000 10:09:50 -0500 To: Jatin Nansi Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem during implementing AFS References: <384366799.976261370823.JavaMail.root@web582-mc> From: Derek Atkins Date: 08 Dec 2000 10:09:50 -0500 In-Reply-To: Jatin Nansi's message of "Fri, 8 Dec 2000 02:42:50 -0500 (EST)" Message-ID: Lines: 100 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Is your machine named node1? Do you have DNS setup? Do you have 'localhost' setup in DNS? -derek Jatin Nansi writes: > hi > > i have been trying to implement the Open AFS file server on a > RedHat Linux 7.0 kernel version 2.2.16-22 on a PIII, with 256 MB RAM as a > trial machine. > > i have downloaded the rpms from the openafs site for the same. > > i am using the procedure as described in the QkBegin guide > that is there in the ibm documentation tarball. > > the problem is as follows: > > Altough the kernel module complains, it does load properly. > the bosserver also starts properly, on the command: > > #bosserver -noauth & > the 'bos setcellname ' command also works fine, and i am also able to add > the pc to the cell. bos listhosts confirms this also. > > > the next step as per the guide should be for starting the database server: > > the command: bos create node1 kaserver simple \ > /usr/afs/bin/kaserver -cell timesgroup.com -noauth > > > this creates the following entry in /usr/afs/local/BosConfig: > ------------------ > restarttime 11 0 4 0 0 > checkbintime 3 0 5 0 0 > bnode simple kaserver 1 > parm /usr/afs/bin/kaserver > end > -------------------- > > however there is no kaserver process running when checked with > ps xa > > similarly, the commands for buserver, ptserver, and vlserver do make entries > into BosConfig, but no processes start. > > then i tried killall bosserver, which killed the boserver process. > when i start it again with bosserver -noauth&, i get the following error: > > /usr/afs/bin/buserver: problems with host name Ubik init failed > ........... repeated several times over for all other servers also. > > the contents of files in /usr/afs/logs: > > AuthLog-- > /usr/afs/bin/kaserver: problems with host name Ubik init failed > ll database. > Using level crypt for Ubik connections. > ubik: primary address 127.0.0.1 does not exist > > BackupLog-- > u: problem with host name > Ubik init failed > > Boslog-- > Server directory access is okay > BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. > ........similarly for all others > ----------- > > > Please tell me where am i going wrong in this, as i am following the ibm > guide to the letter. > > thank you > > Jatin > > ________________________________________ > > LINUX - Its a way of life > ________________________________________ > ------------------------------------ > FREE Personalized Email at India.com > Sign up at http://mail.india.com > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx195-mta.mail.com by penn.central.org id aa01818; 8 Dec 2000 10:50 EST Received: from web105-mc (web105-mc.mail.com [165.251.48.93]) by rmx195-mta.mail.com (8.9.3/8.9.3) with SMTP id KAA19676; Fri, 8 Dec 2000 10:49:36 -0500 (EST) Message-ID: <381405838.976290575620.JavaMail.root@web105-mc> Date: Fri, 8 Dec 2000 10:49:35 -0500 (EST) From: Jatin Nansi To: Derek Atkins Subject: Re: [OpenAFS] Problem during implementing AFS Cc: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.197.16.149 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hi the machines name is node1 DNS is not setup, but every host entry is there in /etc/hosts localhost is there in /etc/hosts. dns is there for the site, but there are no entries for node1. thanks Jatin ------Original Message------ From: Derek Atkins To: Jatin Nansi Sent: December 8, 2000 3:09:50 PM GMT Subject: Re: [OpenAFS] Problem during implementing AFS Is your machine named node1? Do you have DNS setup? Do you have 'localhost' setup in DNS? -derek Jatin Nansi writes: > hi > > i have been trying to implement the Open AFS file server on a > RedHat Linux 7.0 kernel version 2.2.16-22 on a PIII, with 256 MB RAM as a > trial machine. > > i have downloaded the rpms from the openafs site for the same. > > i am using the procedure as described in the QkBegin guide > that is there in the ibm documentation tarball. > > the problem is as follows: > > Altough the kernel module complains, it does load properly. > the bosserver also starts properly, on the command: > > #bosserver -noauth & > the 'bos setcellname ' command also works fine, and i am also able to add > the pc to the cell. bos listhosts confirms this also. > > > the next step as per the guide should be for starting the database server: > > the command: bos create node1 kaserver simple \ > /usr/afs/bin/kaserver -cell timesgroup.com -noauth > > > this creates the following entry in /usr/afs/local/BosConfig: > ------------------ > restarttime 11 0 4 0 0 > checkbintime 3 0 5 0 0 > bnode simple kaserver 1 > parm /usr/afs/bin/kaserver > end > -------------------- > > however there is no kaserver process running when checked with > ps xa > > similarly, the commands for buserver, ptserver, and vlserver do make entries > into BosConfig, but no processes start. > > then i tried killall bosserver, which killed the boserver process. > when i start it again with bosserver -noauth&, i get the following error: > > /usr/afs/bin/buserver: problems with host name Ubik init failed > ........... repeated several times over for all other servers also. > > the contents of files in /usr/afs/logs: > > AuthLog-- > /usr/afs/bin/kaserver: problems with host name Ubik init failed > ll database. > Using level crypt for Ubik connections. > ubik: primary address 127.0.0.1 does not exist > > BackupLog-- > u: problem with host name > Ubik init failed > > Boslog-- > Server directory access is okay > BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. > ........similarly for all others > ----------- > > > Please tell me where am i going wrong in this, as i am following the ibm > guide to the letter. > > thank you > > Jatin > > ________________________________________ > > LINUX - Its a way of life > ________________________________________ > ------------------------------------ > FREE Personalized Email at India.com > Sign up at http://mail.india.com > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH warlord@MIT.EDU PGP key available ________________________________________ LINUX - Its a way of life ________________________________________ ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa02964; 8 Dec 2000 12:07 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id MAA07682; Fri, 8 Dec 2000 12:06:57 -0500 To: Jatin Nansi Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem during implementing AFS References: <381405838.976290575620.JavaMail.root@web105-mc> From: Derek Atkins Date: 08 Dec 2000 12:06:55 -0500 In-Reply-To: Jatin Nansi's message of "Fri, 8 Dec 2000 10:49:35 -0500 (EST)" Message-ID: Lines: 152 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Jatin Nansi writes: > hi > > the machines name is node1 > DNS is not setup, but every host entry is there in /etc/hosts > localhost is there in /etc/hosts. > > dns is there for the site, but there are no entries for node1. > > thanks You might try setting up DNS and making sure 'node1' is in there. This particular error seems to imply that name service is having a problem: > AuthLog-- > /usr/afs/bin/kaserver: problems with host name Ubik init failed > ll database. > Using level crypt for Ubik connections. > ubik: primary address 127.0.0.1 does not exist What happens if you 'host 127.0.0.1', 'host localhost', and 'host node1'? > Jatin -derek > ------Original Message------ > From: Derek Atkins > To: Jatin Nansi > Sent: December 8, 2000 3:09:50 PM GMT > Subject: Re: [OpenAFS] Problem during implementing AFS > > > Is your machine named node1? > Do you have DNS setup? > Do you have 'localhost' setup in DNS? > > -derek > > Jatin Nansi writes: > > > hi > > > > i have been trying to implement the Open AFS file server on a > > RedHat Linux 7.0 kernel version 2.2.16-22 on a PIII, with 256 MB RAM as a > > trial machine. > > > > i have downloaded the rpms from the openafs site for the same. > > > > i am using the procedure as described in the QkBegin guide > > that is there in the ibm documentation tarball. > > > > the problem is as follows: > > > > Altough the kernel module complains, it does load properly. > > the bosserver also starts properly, on the command: > > > > #bosserver -noauth & > > the 'bos setcellname ' command also works fine, and i am also able to add > > the pc to the cell. bos listhosts confirms this also. > > > > > > the next step as per the guide should be for starting the database server: > > > > the command: bos create node1 kaserver simple \ > > /usr/afs/bin/kaserver -cell timesgroup.com -noauth > > > > > > this creates the following entry in /usr/afs/local/BosConfig: > > ------------------ > > restarttime 11 0 4 0 0 > > checkbintime 3 0 5 0 0 > > bnode simple kaserver 1 > > parm /usr/afs/bin/kaserver > > end > > -------------------- > > > > however there is no kaserver process running when checked with > > ps xa > > > > similarly, the commands for buserver, ptserver, and vlserver do make > entries > > into BosConfig, but no processes start. > > > > then i tried killall bosserver, which killed the boserver process. > > when i start it again with bosserver -noauth&, i get the following error: > > > > /usr/afs/bin/buserver: problems with host name Ubik init failed > > ........... repeated several times over for all other servers also. > > > > the contents of files in /usr/afs/logs: > > > > AuthLog-- > > /usr/afs/bin/kaserver: problems with host name Ubik init failed > > ll database. > > Using level crypt for Ubik connections. > > ubik: primary address 127.0.0.1 does not exist > > > > BackupLog-- > > u: problem with host name > > Ubik init failed > > > > Boslog-- > > Server directory access is okay > > BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. > > ........similarly for all others > > ----------- > > > > > > Please tell me where am i going wrong in this, as i am following the ibm > > guide to the letter. > > > > thank you > > > > Jatin > > > > ________________________________________ > > > > LINUX - Its a way of life > > ________________________________________ > > ------------------------------------ > > FREE Personalized Email at India.com > > Sign up at http://mail.india.com > > > > _______________________________________________ > > OpenAFS-info mailing list > > OpenAFS-info@openafs.org > > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info > > -- > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > Member, MIT Student Information Processing Board (SIPB) > URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH > warlord@MIT.EDU PGP key available > > ________________________________________ > > LINUX - Its a way of life > ________________________________________ > ------------------------------------ > FREE Personalized Email at India.com > Sign up at http://mail.india.com > -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.mirapoint.com by penn.central.org id aa04937; 8 Dec 2000 14:32 EST Received: from tim-bsd.mirapoint.com (tim-bsd.mirapoint.com [192.168.0.117]) by mail.mirapoint.com (Mirapoint) with SMTP id AAV11037; Fri, 8 Dec 2000 11:32:44 -0800 (PST) X-Spook: Area 51 radar strategic colonel NWO PLO nuclear To: aeneous@speakeasy.org Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Using OpenAFS for implementing mail servers References: <200012081333.IAA29148@oo.yi.org> From: Tim Showalter Date: 08 Dec 2000 11:33:58 -0800 In-Reply-To: aeneous@speakeasy.org's message of "Fri, 08 Dec 2000 08:33:16 -0500" Message-ID: <7dsnny1zdl.fsf@tim-bsd.mirapoint.com> Lines: 19 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Canyonlands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: aeneous@speakeasy.org writes: > > I will refrain from plugging my company's (Mirapoint) fine, high-quality > > mail server products. I'd also like to not plug IMAP at this point, > > although I'll note that it's difficult to get most mail clients to read > > AFS (that idea never really got out of CMU and it's been dying a long > > time there). > > Mail clients will read AFS just fine - where did you get an idea that > was a problem? Perhaps you meant that it's difficult to get clients to > read AMDS-style mail directories? Most mail clients run on systems that don't support AFS, and most clients don't run on Unix (and therefore need to fetch from POP or IMAP anyway). Sorry that my text implies something else. Tim From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from web906.mail.yahoo.com by penn.central.org id aa07903; 8 Dec 2000 20:42 EST Received: (qmail 19188 invoked by uid 60001); 9 Dec 2000 01:42:48 -0000 Message-ID: <20001209014248.19187.qmail@web906.mail.yahoo.com> Received: from [207.135.67.1] by web906.mail.yahoo.com; Fri, 08 Dec 2000 17:42:48 PST Date: Fri, 8 Dec 2000 17:42:48 -0800 (PST) From: Phuoc Ly To: openafs-info@openafs.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: [OpenAFS] openafs volume size limit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Is there any limit on the size of a volume or directory under AFS? I have been told that the max size is 8G. __________________________________________________ Do You Yahoo!? Yahoo! Shopping - Thousands of Stores. Millions of Products. http://shopping.yahoo.com/ From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from host2.marvell.com by penn.central.org id aa06199; 8 Dec 2000 16:05 EST Received: from flash.marvell.com by marvell.com (SMI-8.6/SMI-SVR4) id NAA17600; Fri, 8 Dec 2000 13:03:49 -0800 Received: from flash by flash.marvell.com (8.8.8+Sun/SMI-SVR4) id NAA14024; Fri, 8 Dec 2000 13:03:56 -0800 (PST) Message-Id: <200012082103.NAA14024@flash.marvell.com> Date: Fri, 8 Dec 2000 13:03:56 -0800 (PST) From: Phuoc Ly Reply-To: Phuoc Ly To: openafs-info@openafs.org MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: 17FdGEXmKK+V8HqWisTrkg== X-Mailer: dtmail 1.2.1 CDE Version 1.2.1 SunOS 5.6 sun4u sparc Subject: [OpenAFS] openafs volume size limit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Is there any limit on the size of a volume or directory under AFS? I have been told that the max size is 8G. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE2FD87.dip.t-dialin.net by penn.central.org id aa27901; 11 Dec 2000 8:25 EST Received: (qmail 30860 invoked from network); 11 Dec 2000 13:24:34 -0000 Received: from cube.ibmpeers.de (192.168.1.2) by p3ee2fd87.dip.t-dialin.net with SMTP; 11 Dec 2000 13:24:34 -0000 Received: from sascha by cube.ibmpeers.de with local (Exim 3.20 #1 (Debian)) id 145SwM-0006B2-00 for ; Mon, 11 Dec 2000 14:24:26 +0100 Date: Mon, 11 Dec 2000 14:24:25 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] New Debian packages available Message-ID: <20001211142425.A23560@cube.ibmpeers.de> Mail-Followup-To: openafs-info@openafs.org References: <200012071329.IAA14878@sweet-transvestite.mit.edu> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="IJpNTDwzlM2Ie8A6" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: <200012071329.IAA14878@sweet-transvestite.mit.edu>; from hartmans@mit.edu on Thu, Dec 07, 2000 at 08:29:48AM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --IJpNTDwzlM2Ie8A6 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Thu, Dec 07, 2000 at 08:29:48AM -0500, Sam Hartman [SH] wrote: SH> I have fixed this bug and built new versions of Debian 2.2 packages SH> available from the openafs.org website. The Packages file now is broken. :( CU/Lnx Sascha --IJpNTDwzlM2Ie8A6 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6NNWJa8hvv2E1w1sRAr4xAJ9qG6XFIIP6uTZh4iJIA+vISaeJngCdEf8h jX//OZ1RqJ8ZvqdAisSxmIg= =bCq9 -----END PGP SIGNATURE----- --IJpNTDwzlM2Ie8A6-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa01142; 11 Dec 2000 14:05 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id OAA06265; Mon, 11 Dec 2000 14:04:52 -0500 (EST) To: Sascha Silbe Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] New Debian packages available References: <200012071329.IAA14878@sweet-transvestite.mit.edu> <20001211142425.A23560@cube.ibmpeers.de> From: Sam Hartman Date: 11 Dec 2000 14:04:52 -0500 In-Reply-To: Sascha Silbe's message of "Mon, 11 Dec 2000 14:24:25 +0100" Message-ID: Lines: 15 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Sascha" == Sascha Silbe writes: Sascha> On Thu, Dec 07, 2000 at 08:29:48AM -0500, Sam Hartman [SH] Sascha> wrote: SH> I have fixed this bug and built new versions of Debian 2.2 SH> packages available from the openafs.org website. Sascha> The Packages file now is broken. :( Sascha> CU/Lnx Sascha Well, it's more like it wasn't a packages file. I have given Jeff a new packages file, and he has installed it. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from acad2.depauw.edu by penn.central.org id aa09700; 12 Dec 2000 11:31 EST Received: from CONVERSION-DAEMON by DEPAUW.EDU (PMDF V5.2-33 #41244) id <01JXM5RHGPDC97QAF7@DEPAUW.EDU> for openafs-info@openafs.org; Tue, 12 Dec 2000 11:28:54 EST Received: from depauw.edu ([163.120.5.98]) by DEPAUW.EDU (PMDF V5.2-33 #41244) with ESMTP id <01JXM5RG2YAM97Q1QE@DEPAUW.EDU> for openafs-info@openafs.org; Tue, 12 Dec 2000 11:28:42 -0500 (EST) Date: Tue, 12 Dec 2000 11:28:39 -0500 From: Corey Kovacs To: openafs-info@openafs.org Message-id: <3A365237.100273F0@depauw.edu> MIME-version: 1.0 X-Mailer: Mozilla 4.75 [en] (Win98; U) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT X-Accept-Language: en Subject: [OpenAFS] afs newbie and startup problem... Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: After compiling the afs code (1.0.1) and going through the quick start guide, I get all the way up to the point of adding the root.afs volume where I get the following error Failed to create the volume root.afs XXXXXXXXXX : Input/output error Error in vos create command. : Input/output error where XXXXXXXXXX is some long number never the same twice... after executing ./vos create root.afs -cell -noauth can someone point me in the right direction as to what this means? -- Corey Kovacs "I know not with what weapons World War III Computer Science Dept. will be fought, but World War IV will be DePauw University. fought with sticks and stones." 765.658.4761 - Albert Einstein From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ietf.207.137.74.236.tx.verio.net by penn.central.org id aa09772; 12 Dec 2000 12:20 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id MAA19229; Tue, 12 Dec 2000 12:20:12 -0500 To: Corey Kovacs Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] afs newbie and startup problem... References: <3A365237.100273F0@depauw.edu> From: Derek Atkins Date: 12 Dec 2000 12:20:12 -0500 In-Reply-To: Corey Kovacs's message of "Tue, 12 Dec 2000 11:28:39 -0500" Message-ID: Lines: 47 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: The XXXXXXXX is the volumeID. It's changing every time because each volume is supposed to have a unique ID within the cell. So, the real question is why it failed to create root.afs. Do you have authentication turned on? If so, do you have valid tokens? -derek Corey Kovacs writes: > After compiling the afs code (1.0.1) and going through the quick start > guide, > I get all the way up to the point of adding the root.afs volume where I > get the > following error > > Failed to create the volume root.afs XXXXXXXXXX > : Input/output error > Error in vos create command. > : Input/output error > > where XXXXXXXXXX is some long number never the same twice... > after executing > > ./vos create root.afs -cell > -noauth > > can someone point me in the right direction as to what this means? > > > > -- > Corey Kovacs "I know not with what weapons World War III > Computer Science Dept. will be fought, but World War IV will be > DePauw University. fought with sticks and stones." > 765.658.4761 - Albert Einstein > > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from e23.nc.us.ibm.com by penn.central.org id aa09852; 12 Dec 2000 12:35 EST Received: from southrelay02.raleigh.ibm.com (southrelay02.raleigh.ibm.com [9.37.3.209]) by e23.nc.us.ibm.com (8.9.3/8.9.3) with ESMTP id MAA28516; Tue, 12 Dec 2000 12:31:50 -0600 Received: from d04nm201.raleigh.ibm.com (d04nm201.raleigh.ibm.com [9.67.228.38]) by southrelay02.raleigh.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id MAA47554; Tue, 12 Dec 2000 12:34:50 -0500 Importance: Normal Subject: Re: [OpenAFS] afs newbie and startup problem... To: Corey Kovacs Cc: openafs-info@openafs.org X-Mailer: Lotus Notes Release 5.0.5 September 22, 2000 Message-ID: From: "Jared Spencer" Date: Tue, 12 Dec 2000 12:34:48 -0500 X-MIMETrack: Serialize by Router on D04NM201/04/M/IBM(Release 5.0.3 (Intl)|21 March 2000) at 12/12/2000 12:34:50 PM MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >The XXXXXXXX is the volumeID. It's changing every time because each >volume is supposed to have a unique ID within the cell. So, the real >question is why it failed to create root.afs. Do you have authentication >turned on? If so, do you have valid tokens? ... >> >> Failed to create the volume root.afs XXXXXXXXXX >> : Input/output error >> Error in vos create command. >> : Input/output error >> Make sure your vice partition(s) is valid and you've restarted the fileserver since creating it. Thanks, Jared From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from sunny.pacific.net.sg by penn.central.org id aa13693; 12 Dec 2000 21:53 EST Received: from pop1.pacific.net.sg (pop1.pacific.net.sg [203.120.90.85]) by sunny.pacific.net.sg with ESMTP id eBD2rQo10421 for ; Wed, 13 Dec 2000 10:53:27 +0800 (SGT) Received: from earth.elinux.com.sg (earth.elinux.com.sg [203.120.99.101]) by pop1.pacific.net.sg with ESMTP id KAA12897 for ; Wed, 13 Dec 2000 10:53:26 +0800 (SGT) Received: (from root@localhost) by earth.elinux.com.sg (8.9.3/8.9.3) id KAA22413 for openafs-info@openafs.org.AVP; Wed, 13 Dec 2000 10:50:52 +0800 Received: from elinux.com.sg (michael@dhcp112.int.elinux.com.sg [192.168.0.112]) by earth.elinux.com.sg (8.9.3/8.9.3) with ESMTP id KAA22407 for ; Wed, 13 Dec 2000 10:50:52 +0800 Message-ID: <3A36E553.ACFC9668@elinux.com.sg> Date: Wed, 13 Dec 2000 10:56:19 +0800 From: Michael Boman Organization: eLinux Pte Ltd X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.16-storm.reiserfs.lvm-storm i686) X-Accept-Language: en MIME-Version: 1.0 To: openafs-info Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] Status of OpenAFS code? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Is OpenAFS ready for the prime time? Is anyone running it in a production enviroment? Best regards Michael Boman -- "eLINUX --- Enabling the Net Economy on Linux" ---------------------------------------------------------- Michael Boman eLinux Pte Ltd Technical Consultant http://www.elinux.com.sg michael@elinux.com.sg Tel: (65) 227 6180 Fax: (65) 227 5808 ---------------------------------------------------------- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from bent.cait.org by penn.central.org id aa18854; 13 Dec 2000 10:45 EST Received: (from awettste@localhost) by bent.cait.org (8.8.8+Sun/8.8.8) id JAA09941 for openafs-info@openafs.org; Wed, 13 Dec 2000 09:37:11 -0600 (CST) Date: Wed, 13 Dec 2000 09:37:11 -0600 From: Andrew Wettstein To: openafs-info@openafs.org Message-ID: <20001213093711.A9874@bent.cait.org> Mail-Followup-To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2i Subject: [OpenAFS] afs and powerpc Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi I'm trying to compile afs for powerpc and there seems to be a few problems. First there is no Makefile.ppc_linux22, but there is a Makefile.ppc_linux24. Does this mean that linux 2.2 doesn't work on ppc? Second, I tried building with ppc_linux24 (i think i had to change a few things in the makefile) and that worked until it got the des directory. It's looking for a conf-ppc-linux.h but there is none, only a conf-i386-linux.h, so I guess i need to know what that file should contain. By the way, I'm also using debian potato, and the entries in sources.gz are pointing to the wrong directory. It should be debian instead of releases. thanks andy From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ginger.cmf.nrl.navy.mil by penn.central.org id aa20131; 13 Dec 2000 12:33 EST Received: from cmf.nrl.navy.mil (borg.cmf.nrl.navy.mil [134.207.10.148]) (authenticated) by ginger.cmf.nrl.navy.mil (8.10.1/8.10.1) with ESMTP id eBDHX8R10718; Wed, 13 Dec 2000 12:33:08 -0500 (EST) Message-Id: <200012131733.eBDHX8R10718@ginger.cmf.nrl.navy.mil> To: Andrew Wettstein cc: openafs-info@openafs.org Subject: Re: [OpenAFS] afs and powerpc In-reply-to: Your message of "Wed, 13 Dec 2000 09:37:11 CST." <20001213093711.A9874@bent.cait.org> X-url: http://www.nrl.navy.mil/CCS/people/chas/index.html X-mailer: nmh 1.0 Date: Wed, 13 Dec 2000 12:33:08 -0500 From: Chas Williams Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >Does this mean that linux 2.2 doesn't work on ppc? >Second, I tried building with ppc_linux24 (i think i had >to change a few things in the makefile) and that worked until >it got the des directory. >It's looking for a conf-ppc-linux.h but there is none, only a >conf-i386-linux.h, so I guess i need to know what that file >should contain. sounds like the 1.0.1 release doesnt have the files i submitted to provide ppc support. i will check this out myself but the file you probably need are in ftp://ftp.cmf.nrl.navy.mil/pub/chas/openafs/ppc From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa20990; 13 Dec 2000 14:21 EST Received: from JOHNSTOWN.ANDREW.CMU.EDU (JOHNSTOWN.ANDREW.CMU.EDU [128.2.121.156]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id OAA17671 for ; Wed, 13 Dec 2000 14:21:08 -0500 (EST) Date: Wed, 13 Dec 2000 14:21:06 -0500 (EST) From: Derrick J Brashear X-Sender: shadow@johnstown.andrew.cmu.edu cc: openafs-info@openafs.org Subject: Re: [OpenAFS] afs and powerpc In-Reply-To: <200012131733.eBDHX8R10718@ginger.cmf.nrl.navy.mil> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Wed, 13 Dec 2000, Chas Williams wrote: > >Does this mean that linux 2.2 doesn't work on ppc? > >Second, I tried building with ppc_linux24 (i think i had > >to change a few things in the makefile) and that worked until > >it got the des directory. > >It's looking for a conf-ppc-linux.h but there is none, only a > >conf-i386-linux.h, so I guess i need to know what that file > >should contain. > > sounds like the 1.0.1 release doesnt have the files i submitted > to provide ppc support. i will check this out myself but the > file you probably need are in ftp://ftp.cmf.nrl.navy.mil/pub/chas/openafs/ppc It's possible I missed this file. If so, it will be included in the 1.0.2 release which should be available in a few days. -D From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from egghead.curl.com by penn.central.org id aa11805; 12 Dec 2000 16:04 EST Received: (qmail 20231 invoked by uid 10171); 12 Dec 2000 16:03:55 -0500 To: openafs-info@openafs.org From: patl@curl.com (Patrick J. LoPresti) Date: 12 Dec 2000 16:03:55 -0500 Message-ID: Lines: 43 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: [OpenAFS] Many questions Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: (If any of these are answered elsewhere, please feel free to flame me and give a pointer.) What is the real status of Open AFS as far as IBM is concerned? (I have read the FAQ, but I am curious about the reality.) Will IBM be working to make fixes and improvements flow from IBM AFS to Open AFS and vice-versa, or is this a one-time code fork into two trees which will diverge from now on? IBM says that Open AFS is "unsupported", which is reasonable. But what if, hypothetically, I wanted to license Windows clients but use OpenAFS servers? Would I still be able to get support for the clients? The Quick Beginnings guide makes no mention of a special "fsck" program for Linux. Does this mean that the standard e2fsck is appropriate for checking Linux /viceXX partitions? (It sounds like it, but I want to make sure.) Are the Linux OpenAFS modules compiled with symbol versioning? On a related note, is there any effort underway to get the Open AFS kernel modules integrated into the mainline Linux kernel sources? Would the license on Open AFS even allow it? The RPMs do not seem to include modified versions of the rsh suite (or any similar tools). The source for rsh.c (OpenAFS 1.0.1 distribution) defines a local "pass_tokens" variable but never uses it for anything; in other words, the source code itself seems to make no sense. I understand that I can use PAM to allow various services to authenticate against the AFS Authentication Server, but what I want is to be able to pass authentication tokens *without* typing my password again... Is that possible, even in principle? Thanks in advance for any answers. - Pat From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from windlord.Stanford.EDU by penn.central.org id aa26096; 13 Dec 2000 20:30 EST Received: (qmail 5673 invoked by uid 50); 14 Dec 2000 01:30:37 -0000 To: openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions References: In-Reply-To: "Patrick J. LoPresti"'s message of "12 Dec 2000 16:03:55 -0500" From: Russ Allbery Organization: The Eyrie Date: 13 Dec 2000 17:30:37 -0800 Message-ID: Lines: 34 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I'm going to bail on most of these questions... some of them are ones that the people here probably can't ask and that you're going to need to ask directly of Transarc. Patrick J LoPresti writes: > The RPMs do not seem to include modified versions of the rsh suite > (or any similar tools). The source for rsh.c (OpenAFS 1.0.1 > distribution) defines a local "pass_tokens" variable but never uses > it for anything; in other words, the source code itself seems to > make no sense. I understand that I can use PAM to allow various > services to authenticate against the AFS Authentication Server, but > what I want is to be able to pass authentication tokens *without* > typing my password again... Is that possible, even in principle? I think it's generally considered to be the Right Thing to just use Kerberos versions of rsh, rlogin, etc. if you want to pass authentication rather than trying to use the token-passing versions. The Kerberos versions can, with Kerberos v5, pass tickets, which is even better. This however requires a Kerberos v5 infrastructure with the migration stuff so that you can generate Kerberos v4 tickets for AFS's uses. If you have just a Kerberos v4 infrastructure, what we did here at Stanford was write a fairly ugly hack to forward Kerberos v4 tickets that we've used for some years now. It works, but we'll be happy to get rid of it in favor of Kerberos v5 ticket passing, as it has various annoyances and security concerns. I'm hoping that in the long run, all of the authentication passing mechanisms will disappear in favor of standardizing on Kerberos v5 ticket passing (either via the native clients or via ssh). -- Russ Allbery (rra@stanford.edu) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dsl254-116-177-nyc1.dsl-isp.net by penn.central.org id aa26938; 13 Dec 2000 22:40 EST Received: from o-o.yi.org (IDENT:def@localhost.localdomain [127.0.0.1]) by oo.yi.org (8.9.3/8.8.7) with ESMTP id WAA08684; Wed, 13 Dec 2000 22:39:50 -0500 Message-Id: <200012140339.WAA08684@oo.yi.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: Corey Kovacs , openafs-info@openafs.org From: aeneous@speakeasy.org Reply-To: aeneous@speakeasy.org Subject: Re: [OpenAFS] afs newbie and startup problem... In-Reply-To: Your message of "Tue, 12 Dec 2000 11:28:39 EST." <3A365237.100273F0@depauw.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 13 Dec 2000 22:39:50 -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > following error > > Failed to create the volume root.afs XXXXXXXXXX > : Input/output error > Error in vos create command. > : Input/output error > > where XXXXXXXXXX is some long number never the same twice... > after executing It's probably a volume id. I'd check the vos code to be sure. I don't know what it means off the top of my head -- is there anything in the volserver log? From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from post.rzg.mpg.de by penn.central.org id aa01369; 14 Dec 2000 5:50 EST Received: from rzg.mpg.de (hwr@reuter.rzg.mpg.de [130.183.2.117]) by post.rzg.mpg.de (8.9.2/8.9.2) with ESMTP id LAA83940; Thu, 14 Dec 2000 11:50:01 +0100 Message-ID: <3A38A5D6.92B792B8@rzg.mpg.de> Date: Thu, 14 Dec 2000 11:49:58 +0100 From: Hartmut Reuter Organization: Max-Planck-Inst. f. Plasmaphysik, RZG X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.14 i686) X-Accept-Language: en, de-DE MIME-Version: 1.0 To: "Patrick J. LoPresti" CC: openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: "Patrick J. LoPresti" wrote: > > (If any of these are answered elsewhere, please feel free to flame me > and give a pointer.) > > What is the real status of Open AFS as far as IBM is concerned? (I > have read the FAQ, but I am curious about the reality.) Will IBM be > working to make fixes and improvements flow from IBM AFS to Open AFS > and vice-versa, or is this a one-time code fork into two trees which > will diverge from now on? > The OpenAFS advisary board has promized not to break compatibility to IBM's AFS. > IBM says that Open AFS is "unsupported", which is reasonable. But > what if, hypothetically, I wanted to license Windows clients but use > OpenAFS servers? Would I still be able to get support for the > clients? You may build your Windows client in OpenAFS as well. > > The Quick Beginnings guide makes no mention of a special "fsck" > program for Linux. Does this mean that the standard e2fsck is > appropriate for checking Linux /viceXX partitions? (It sounds like > it, but I want to make sure.) The AFS-server in Linux uses the so called NAMEI-Interface: i.e. all AFS files and directories are regular files in the filesystem which, however, have synthetic names. Therefor no special fsck is required. > > Are the Linux OpenAFS modules compiled with symbol versioning? > > On a related note, is there any effort underway to get the Open AFS > kernel modules integrated into the mainline Linux kernel sources? > Would the license on Open AFS even allow it? There were already some discussions about this point in the mailing list. Since the AFS kernel extensions are huge compared with the Linux kernel itself and since the kernel versions are develloping rather fast it seams for the moment better to keep things separate. Also, the Linux port should not fork from the general AFS client code. > > The RPMs do not seem to include modified versions of the rsh suite > (or any similar tools). The source for rsh.c (OpenAFS 1.0.1 > distribution) defines a local "pass_tokens" variable but never uses > it for anything; in other words, the source code itself seems to > make no sense. I understand that I can use PAM to allow various > services to authenticate against the AFS Authentication Server, but > what I want is to be able to pass authentication tokens *without* > typing my password again... Is that possible, even in principle? This is old stuff. You better use ssh instead of rsh. We have local changes to ssh which do a token transfer. -Hartmut Reuter > > Thanks in advance for any answers. > > - Pat > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- ----------------------------------------------------------------- Hartmut Reuter e-mail reuter@rzg.mpg.de phone +49-89-3299-1328 RZG (Rechenzentrum Garching) fax +49-89-3299-1301 Computing Center of the Max-Planck-Gesellschaft (MPG) and the Institut fuer Plasmaphysik (IPP) ----------------------------------------------------------------- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa02894; 14 Dec 2000 9:56 EST Received: from oraadv.cc.umr.edu (oraadv.cc.umr.edu [131.151.1.94]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id IAA19739; Thu, 14 Dec 2000 08:56:04 -0600 Received: from (root@localhost) by oraadv.cc.umr.edu (8.9.3/M.4.00) id IAA29372; Thu, 14 Dec 2000 08:56:04 -0600 Date: Thu, 14 Dec 2000 08:56:04 -0600 From: System Administrator To: openafs-info@openafs.org, info-afs@transarc.com Message-ID: <20001214085604.A29234@umr.edu> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="X1bOJ3K7DJ5YkBrT" Content-Disposition: inline User-Agent: Mutt/1.3.11i Subject: [OpenAFS] Scripts for cleaning tokens/pags Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --X1bOJ3K7DJ5YkBrT Content-Type: text/plain; charset=us-ascii Content-Disposition: inline These are two scripts we use on machines with the following criteria: a. Lots of authentications that involve tokens - this in our case does _NOT_ include POP and IMAP servers, those are krb5 only, and do not get tokens, however, they _DO_ include telnet logins, netatalk-afpd, and samba. b. Setup such that tokens don't go away in general. In the case of telnet sessions, people often leave stuff running in background - having the tokens go away would cause a problem. c. (HP-UX) Tokens are not owned by userids that don't match their afsid. (This is a limitation of HP-UX, I have no way of determining the pags that are in use by a process.) On linux, /proc can be used to determine all active pags from processes that are running. ---- For reference, if you run this script on a machine that is overly bogged down by pags currently - it will likely appear to lock up the machine for a few seconds as it collapses a huge in-kernel hash into a tiny one after you've cleared out all the old tokens. I'm sure someone could improve this immensely by triggering the unlog system call from perl directly instead of system("unlog"); --- The way the scripts work is, using kdump, they retrieve a list of all the pags in the kernel hash, they they attempt to determine which of those pags contain tokens that need to be kept. (In the case of the hpux10 script, that means 'the userid associated with this afsid for this token has processes running on the machine.) (In the case of linux, that means 'a process exists in this pag'.) it then loops through all those pags, putting the script into that pag temporarily (setgroups) and issuing unlog. --- Note - this is necessary even on the most current afs for linux, as it still does not do garbage collection of tokens/pags. -- Nathan --X1bOJ3K7DJ5YkBrT Content-Type: application/x-perl Content-Disposition: attachment; filename="clean-tokens-hpux10.pl" #!/umr/bin/perl $| = 1; # # Input sources # $in_kdump = "/usr/afsws/etc/kdump -users|"; $in_ps = "/usr/bin/ps -ef|"; # # Read in the kdump data # open(IN, $in_kdump); print "Processing KDump: "; $count = 0; %pags = (); %pags_per_uid = (); while ( $line = ) { next if ( $line !~ /:/o ); if ( $line =~ /uid=x(41[0-9a-fA-F]+),.* vid=([0-9]+)/o ) { $pag = $1; $uid = $2; $pagnum = hex($pag) - hex("41000000"); $pag2 = $pagnum % 16384; $pag1 = ($pagnum-$pag2)/16384; $pag1 += 33536; $pag2 += 32512; if ( $uid > 1 ) { $pags{"$pag1:$pag2"} = $uid; } # if ( $count % 250 == 0 ) { print "." }; $count++; } } print "$count active tokens.\n"; close(IN); # # Read in the ps data # open(IN, $in_ps); print "Processing Users: "; %uids = (); $count = 0; while ( chomp($line = ) ) { $line =~ s/^ +//go; $line =~ s/ +.*//go; $uid = getpwnam($line); next if ($uid eq ""); $uids{$uid} = 1; # if ( $count % 50 == 0 ) { print "." }; $count++; } close(IN); print "$count active user processes.\n"; # # Clean pags # $cleaned = 0; $skipped = 0; $count = 0; while ( ($pag,$uid) = each(%pags) ) { if ( $count % 100 == 0 ) { print "Processed Tokens: $count\n"; } $count++; if ( $uids{$uid} ) { # print "-"; $skipped++; } else { ($pag1, $pag2) = split(/:/, $pag); $) = "0 $pag1 $pag2"; system("/usr/afsws/bin/unlog"); $cleaned++; # print "+"; } } print "Cleaning Tokens: $skipped tokens skipped.\n"; print "Cleaning Tokens: $cleaned tokens unlogged.\n"; --X1bOJ3K7DJ5YkBrT Content-Type: application/x-perl Content-Disposition: attachment; filename="clean-tokens-redhat.pl" #!/umr/bin/perl $| = 1; # # Input sources # $in_kdump = "/usr/afsws/etc/kdump -users|"; $in_ps = "/bin/ps -ef|"; # # Read in the kdump data # open(IN, $in_kdump); print "Processing KDump: "; $count = 0; %pags = (); %pags_per_uid = (); while ( $line = ) { next if ( $line !~ /:/o ); if ( $line =~ /uid=x(41[0-9a-fA-F]+),.*vid=([0-9]+)/o ) { $pag = $1; $uid = $2; $pagnum = hex($pag) - hex("41000000"); $pag2 = $pagnum % 16384; $pag1 = ($pagnum-$pag2)/16384; $pag1 += 33536; $pag2 += 32512; if ( $uid > 0 ) { $pags{"$pag1:$pag2"} = $uid; print "\t$pag1,$pag2\n"; } # if ( $count % 250 == 0 ) { print "." }; $count++; } } print "$count active tokens.\n"; close(IN); # # Read in the info about active processes, delete from # %pags for each active # print "Scanning processes:\n"; opendir(DIR, "/proc"); while ( $file = readdir(DIR) ) { next if ( $file !~ /^[0-9]+$/o ); print "\t$file: "; $groups = ""; open(IN, "/proc/$file/status"); while ( chomp($line = ) ) { if ( $line =~ /^Groups:(.*)$/o ) { $groups = $1; last; } } close(IN); $groups =~ s/[\r\n\t ]+/ /gio; $groups =~ s/^ +//gio; $groups =~ s/ +$//gio; @groups = split(' ', $groups); $pag1 = $groups[0]; $pag2 = $groups[1]; if ( $pag1 >= 32512 && $pag2 >= 33536 ) { print "$pag1,$pag2 "; if ( $pags{"$pag1:$pag2"} ) { print "skipping pag\n"; delete $pags{"$pag1:$pag2"}; } else { print "no active token\n"; } } else { print "not active pag\n"; } } closedir(DIR); # # Clean pags # $cleaned = 0; $skipped = 0; $count = 0; while ( ($pag,$uid) = each(%pags) ) { if ( $count % 100 == 0 ) { print "Processed Tokens: $count\n"; } $count++; ($pag1, $pag2) = split(/:/, $pag); $) = "0 $pag1 $pag2"; system("/usr/afsws/bin/unlog"); $cleaned++; print "\t$pag unlogged\n"; } print "Cleaning Tokens: $cleaned tokens unlogged.\n"; --X1bOJ3K7DJ5YkBrT-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from egghead.curl.com by penn.central.org id aa03126; 14 Dec 2000 10:20 EST Received: (qmail 19645 invoked by uid 10171); 14 Dec 2000 10:20:20 -0500 Date: 14 Dec 2000 10:20:20 -0500 Message-ID: <20001214152020.19644.qmail@egghead.curl.com> From: patl@curl.com (Patrick J. LoPresti) To: Hartmut Reuter Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions References: <3A38A5D6.92B792B8@rzg.mpg.de> Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Thank you for your answers. I have a couple of follow-up questions: Hartmut Reuter writes: > You may build your Windows client in OpenAFS as well. Interesting. Does that include Win2k? > This is old stuff. You better use ssh instead of rsh. We have local > changes to ssh which do a token transfer. Are these changes available somewhere? I already use ssh; I only asked about rsh because that is what the AFS docs talk about. Thanks again! - Pat From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa06780; 14 Dec 2000 16:22 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id PAA20959; Thu, 14 Dec 2000 15:22:25 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Thu, 14 Dec 2000 15:22:25 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5CD2@umr-mail02> From: "Neulinger, Nathan R." To: openafs-info@openafs.org, info-afs@transarc.com Subject: RE: [OpenAFS] Scripts for cleaning tokens/pags Date: Thu, 14 Dec 2000 15:22:22 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: FYI - I accidentally sent these from the wrong window... If you have any questions about them, send them to nneul@umr.edu, as I won't see the replies otherwise. -- Nathan > -----Original Message----- > From: System Administrator [mailto:root@umr.edu] > Sent: Thursday, December 14, 2000 8:56 AM > To: openafs-info@openafs.org; info-afs@transarc.com > Subject: [OpenAFS] Scripts for cleaning tokens/pags > > > These are two scripts we use on machines with the following criteria: > > a. Lots of authentications that involve tokens - this in our > case does > _NOT_ include POP and IMAP servers, those are krb5 only, and > do not get > tokens, however, they _DO_ include telnet logins, netatalk-afpd, and > samba. > > b. Setup such that tokens don't go away in general. In the > case of telnet > sessions, people often leave stuff running in background - having the > tokens go away would cause a problem. > > c. (HP-UX) Tokens are not owned by userids that don't match > their afsid. > (This is a limitation of HP-UX, I have no way of determining the pags > that are in use by a process.) On linux, /proc can be used to > determine all > active pags from processes that are running. > > ---- > > For reference, if you run this script on a machine that is > overly bogged > down by pags currently - it will likely appear to lock up the > machine for > a few seconds as it collapses a huge in-kernel hash into a > tiny one after > you've cleared out all the old tokens. > > I'm sure someone could improve this immensely by triggering the unlog > system call from perl directly instead of system("unlog"); > > --- > > The way the scripts work is, using kdump, they retrieve a > list of all the > pags in the kernel hash, they they attempt to determine which > of those > pags contain tokens that need to be kept. (In the case of the hpux10 > script, that means 'the userid associated with this afsid for > this token > has processes running on the machine.) (In the case of linux, > that means > 'a process exists in this pag'.) it then loops through all > those pags, > putting the script into that pag temporarily (setgroups) and > issuing unlog. > > --- > > Note - this is necessary even on the most current afs for > linux, as it > still does not do garbage collection of tokens/pags. > > -- Nathan > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa06942; 14 Dec 2000 16:33 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id QAA03181 for ; Thu, 14 Dec 2000 16:33:46 -0500 (EST) Date: Thu, 14 Dec 2000 16:34:09 -0500 From: Derrick J Brashear To: openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions Message-ID: <1583972704.976829649@skittlebrau.trafford.dementia.org> In-Reply-To: <20001214152020.19644.qmail@egghead.curl.com> X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Thursday, December 14, 2000 10:20:20 AM -0500 "Patrick J. LoPresti" wrote: > Thank you for your answers. I have a couple of follow-up questions: > > > Hartmut Reuter writes: > >> You may build your Windows client in OpenAFS as well. > > Interesting. Does that include Win2k? Changes that support Win2k (thought I have no tested them) have been contributed and are included in the 1.0.1 release. >> This is old stuff. You better use ssh instead of rsh. We have local >> changes to ssh which do a token transfer. > > Are these changes available somewhere? I already use ssh; I only > asked about rsh because that is what the AFS docs talk about. openssh has them included. Otherwise, something like http://www.monkey.org/~dugsong/ssh-afs/ (a web search for dug song ssh afs will certainly turn them up) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dingo.sta.cathedral.org by penn.central.org id aa07331; 14 Dec 2000 17:02 EST Received: by dingo.sta.cathedral.org (STA/mail, from userid 500) id 809462061EC; Thu, 14 Dec 2000 17:02:47 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by dingo.sta.cathedral.org (STA/mail) with ESMTP id 7B4112061EB; Thu, 14 Dec 2000 17:02:47 -0500 (EST) Date: Thu, 14 Dec 2000 17:02:47 -0500 (EST) From: Michael Nelson X-Sender: To: "Patrick J. LoPresti" Cc: Hartmut Reuter , Subject: Re: [OpenAFS] Many questions In-Reply-To: <20001214152020.19644.qmail@egghead.curl.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Return-Path: Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On 14 Dec 2000, Patrick J. LoPresti wrote: > > You may build your Windows client in OpenAFS as well. > > Interesting. Does that include Win2k? I have submitted changes to make things work pretty well. You have to jump through some hoops though. NetBEUI has to be installed and you have to set your LANA ID in the AFS configuration GUI to 6 (at least for me and one or two other folks, it could be different on your system). -mike From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa07550; 14 Dec 2000 17:24 EST Received: (qmail 22710 invoked from network); 14 Dec 2000 22:24:55 -0000 Received: from nneul.rollanet.org (HELO umr.edu) (208.18.13.137) by mx.rollanet.org with SMTP; 14 Dec 2000 22:24:55 -0000 Message-ID: <3A3948B7.30A884D7@umr.edu> Date: Thu, 14 Dec 2000 16:24:55 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.17 i686) X-Accept-Language: en MIME-Version: 1.0 To: Michael Nelson CC: "Patrick J. LoPresti" , Hartmut Reuter , openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Michael Nelson wrote: > > On 14 Dec 2000, Patrick J. LoPresti wrote: > > > You may build your Windows client in OpenAFS as well. > > > > Interesting. Does that include Win2k? > > I have submitted changes to make things work pretty well. You have to jump > through some hoops though. NetBEUI has to be installed and you have to set > your LANA ID in the AFS configuration GUI to 6 (at least for me and one or > two other folks, it could be different on your system). Hopefully transarc/ibm will contribute their 3.6 fixes for LANA stuff in the 2.5 patches. -- Nathan From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from acad1.depauw.edu by penn.central.org id aa15112; 15 Dec 2000 16:07 EST Received: from CONVERSION-DAEMON by DEPAUW.EDU (PMDF V5.2-33 #41244) id <01JXQMD7592896TH1Q@DEPAUW.EDU> for openafs-info@openafs.org; Fri, 15 Dec 2000 16:06:47 EST Received: from depauw.edu ([163.120.5.3]) by DEPAUW.EDU (PMDF V5.2-33 #41244) with ESMTP id <01JXQMD6YB7S97SEY5@DEPAUW.EDU> for openafs-info@openafs.org; Fri, 15 Dec 2000 16:06:45 -0500 (EST) Date: Fri, 15 Dec 2000 16:06:45 -0500 From: Corey Kovacs To: "openafs-info@openafs.org" Message-id: <3A3A87E5.D8B03128@depauw.edu> MIME-version: 1.0 X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.17 i686) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT X-Accept-Language: en Subject: [OpenAFS] login information Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: First, thanx again to everyone who responded to my previous question regarding the volume creation error. The solution was of course that I had skipped a step (restarting the server process) all is well with the volumes now :) New Question: I want to use the AFS server to handle the authentication for the clients as well as supply their home directories etc. so that the users have only to issue one password response. I have read that one must either have local login information (home dir etc) predefined by some other means, either by using a bunk passwd file or NIS/NIS+ etc. In the documentation for AFS it seems to point out under creating users etc that you can set this information through AFS. Is this true or must I use one of the other mechanisms? -- Corey Kovacs "I know not with what weapons World War III Computer Science Dept. will be fought, but World War IV will be DePauw University. fought with sticks and stones." 765.658.6538 - Albert Einstein From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa15300; 15 Dec 2000 16:35 EST Received: (qmail 28476 invoked from network); 15 Dec 2000 21:34:39 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 15 Dec 2000 21:34:39 -0000 Message-ID: <3A3A8E6F.CE6EF822@umr.edu> Date: Fri, 15 Dec 2000 15:34:39 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18 i686) X-Accept-Language: en MIME-Version: 1.0 To: Corey Kovacs CC: "openafs-info@openafs.org" Subject: Re: [OpenAFS] login information References: <3A3A87E5.D8B03128@depauw.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Correct, afs and kerberos only do authentication, they do not do authorization (except in very limited cases.) The usual approach is to define the password file in NIS/NIS+/LDAP/HESIOD/etc. but put a null/invalid password in the password field, then to put pam'ified or replacement /bin/login commands on the unix stations. -- Nathan Corey Kovacs wrote: > > First, thanx again to everyone who responded to my previous question > regarding the volume creation error. The solution was of course that I > had skipped a step (restarting the server process) all is well with the > volumes now :) > > New Question: > > I want to use the AFS server to handle the authentication for the > clients as > well as supply their home directories etc. so that the users have only > to issue > one password response. I have read that one must either have local login > > information (home dir etc) predefined by some other means, either by > using a bunk passwd file or NIS/NIS+ etc. In the documentation for > AFS it seems to point out under creating users etc that you can set this > > information through AFS. Is this true or must I use one of the other > mechanisms? > > -- > Corey Kovacs "I know not with what weapons World War III > Computer Science Dept. will be fought, but World War IV will be > DePauw University. fought with sticks and stones." > 765.658.6538 - Albert Einstein > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from beniaminus.red.cert.org by penn.central.org id aa07160; 14 Dec 2000 16:52 EST Received: from smtp.indigo.cert.org (smtp.indigo.cert.org [192.88.209.150]) by beniaminus.red.cert.org (8.9.3/8.9.3/1.9) with ESMTP id QAA12693; Thu, 14 Dec 2000 16:52:49 -0500 (EST) Received: from a3.blue.cert.org (a3.blue.cert.org [192.88.210.35]) by smtp.indigo.cert.org (8.9.3/8.9.3/2.4) with ESMTP id QAA27783; Thu, 14 Dec 2000 16:52:50 -0500 (EST) Date: Thu, 14 Dec 2000 16:52:50 -0500 From: Rudolph T Maceyko To: Derrick J Brashear cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Many questions Message-ID: <233300000.976830770@a3.blue.cert.org> In-Reply-To: <1583972704.976829649@skittlebrau.trafford.dementia.org> X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Dug made those patches inaccessible and is currently recommending the use of openssh... -Rudy -- Rudy Maceyko Voice +1 412 268 6818 CERT Coordination Center Fax +1 412 268 6989 Software Engineering Institute Carnegie Mellon University --On Thursday, December 14, 2000 16:34:09 -0500 Derrick J Brashear wrote: >> Are these changes available somewhere? I already use ssh; I only >> asked about rsh because that is what the AFS docs talk about. > > openssh has them included. Otherwise, something like > http://www.monkey.org/~dugsong/ssh-afs/ > (a web search for dug song ssh afs will certainly turn them up) From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from euclid.jpl.nasa.gov by penn.central.org id aa08213; 14 Dec 2000 21:52 EST Received: from pepe.jpl.nasa.gov (vpn-232-101.jpl.nasa.gov [128.149.232.101]) by euclid.JPL.NASA.GOV (8.9.3/8.9.3) with ESMTP id SAA07241; Thu, 14 Dec 2000 18:52:40 -0800 (PST) Message-Id: <4.3.2.7.2.20001214184412.00b22d40@psdt.com> X-Sender: pjscott@mail2a.jpl.nasa.gov X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 14 Dec 2000 18:54:02 -0800 To: openafs-info@openafs.org, info-afs@transarc.com From: Peter Scott Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Subject: [OpenAFS] Web server over AFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hello. I am looking for input from anyone who has successfully run an institutional web server using AFS and provided users the ability to create web interfaces to persistent data. The hard part is ensuring that customer CGI programs are the only processes allowed to modify that data and that no-one else with an AFS account can get at it through CGI programs in their own directory. Non-AFS systems solve this with setuid, which is not available on AFS. I am talking about a centralized web server which is administered by AFS admins, no user access allowed to local disk space. We have a couple of theoretical solutions already, so I really want to constrain the answers to solutions that have been proven in practice. Bonus points if yours works with load balancing or round-robin type web server multiplexing. Thanks in advance. -- Peter Scott Peter.J.Scott@jpl.nasa.gov From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from quince.ifs.umich.edu by penn.central.org id aa14294; 15 Dec 2000 14:02 EST Received: from pepper-pot (pepper-pot.ifs.umich.edu [141.211.168.91]) by quince.ifs.umich.edu (8.6.13/8.6.12) with ESMTP id OAA21002; Fri, 15 Dec 2000 14:02:18 -0500 Message-Id: <200012151902.OAA21002@quince.ifs.umich.edu> To: openafs-info@openafs.org, info-afs@transarc.com In-reply-to: Your message of "Thu, 14 Dec 2000 18:54:02 PST." <4.3.2.7.2.20001214184412.00b22d40@psdt.com> Date: Fri, 15 Dec 2000 14:02:18 -0500 From: Marcus Watts Subject: [OpenAFS] Re: Web server over AFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Peter Scott writes: > Message-Id: <4.3.2.7.2.20001214184412.00b22d40@psdt.com> > Date: Thu, 14 Dec 2000 18:54:02 -0800 > To: openafs-info@openafs.org, info-afs@transarc.com > From: Peter Scott > Subject: Web server over AFS > > Hello. I am looking for input from anyone who has successfully run an > institutional web server using AFS and provided users the ability to create > web interfaces to persistent data. The hard part is ensuring that customer > CGI programs are the only processes allowed to modify that data and that > no-one else with an AFS account can get at it through CGI programs in their > own directory. Non-AFS systems solve this with setuid, which is not > available on AFS. I am talking about a centralized web server which is > administered by AFS admins, no user access allowed to local disk space. > > We have a couple of theoretical solutions already, so I really want to > constrain the answers to solutions that have been proven in > practice. Bonus points if yours works with load balancing or round-robin > type web server multiplexing. Thanks in advance. > -- > Peter Scott > Peter.J.Scott@jpl.nasa.gov > > SUID doesn't exist in AFS, because the trust model doesn't exist. SUID works under Unix because the OS can be trusted. With AFS, a user could boot a rogue OS that tweaks things, or halt the OS at any time and munge things, so the fileserver can't trust that the program running on the client machine is really working the way intended by the owner of the file from which the program was loaded. Even in straight Unix, SUID gets turned off by ptrace for similar reasons. If you are storing data in AFS, then whatever accesses that data has to have a token (ticket for AFS) that permits that access. There are a couple of ways to arrange for this. The straight-forward way would be to arrange for the CGI script to do an AFS setpag, klog, and unlog. In order to do the klog, it will need to have a srvtab. The srvtab will function in much the same way that a password would work for a human. The AFS klog program won't work for "klog" here - you need a special program to do this. We (umich.edu) use gettoken, a copy of which can be found in /afs/umich.edu/group/itd/build/mdw/gettoken for this purpose. (This isn't packaged up nicely; use at own risk,&etc&etc). If your cgi script were a C program, you'd probably want to hard-wire the equivalent logic into your cgi... There are two main issues with above cgi logic: (1) the cgi script has to have access to the srvtab. With the typical apache httpd setup, this could be done one of two ways: (a) permit the srvtab read-everybody; *any* cgi script or other program has potential access to the srvtab. This works best when the web server has tightly controlled access, dedicated functionality, so the functionality of all cgi scripts can be trusted. (b) the cgi script is SUID, to an owner that has access to the srvtab. (2) each run of the cgi script has to authenticate all over again. For a server with a low to medium volume of traffic, this may be fine. For a server with a large volume of traffic, there may be various issues, including the "only 1 new pag per second" logic in the cache manager. Another way to package things up is that the cgi script runs, and talks to a separate server daemon that does the actual work. This is a more complicated approach to setup, but has some advantages: since the daemon can be persistant, it can remember things between calls (careful how you use this.) More importantly; the daemon can maintain a persistant AFS token shared among multiple calls. There are lots of ways the daemon could work; it could run on the same machine and use Unix domain sockets (if all users on that machine can be trusted), or it could run on a different machine and require that a kerberos ticket be used to access it. If run on a different machine, it can enforce its own trust rules without concern that the web server might be compromised. On the other hand, the web server still needs to do some form of "proxy" authentication; it needs to say "Hi, I'm bill-the-web-server@X, working on behalf of andrew-the-user@X" to the daemon. Since there are no pags or AFS file server connections, using kerberos to authenticate to the daemon is still nicer than getting an AFS token straight off. There are lots of other ways to set things up, but I think this covers a good range of possibilities. One other possibility, for instance, is to teach apache about pags, and using the same pag across multiple cgi script runs, doing an "unlog" between each cgi script. It's possible to do this with an apache loadable module, so this isn't as hard as it might sound. I even have source code, somewhere, that does this, which I last built with apache 1.3.12. -Marcus Watts UM ITCS Umich Systems Group From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from portal.pnl.gov by penn.central.org id aa14684; 15 Dec 2000 15:21 EST Received: from CONVERSION-DAEMON by pnl.gov (PMDF V5.2-33 #42505) id <01JXQEHA37KY8Y5CJT@pnl.gov> for openafs-info@openafs.org; Fri, 15 Dec 2000 12:21:01 PST Received: from vscan2.pnl.gov ([130.20.64.141]) by pnl.gov (PMDF V5.2-33 #42505) with SMTP id <01JXQEGWHJ1A8WWDY5@pnl.gov>; Fri, 15 Dec 2000 12:20:41 -0800 (PST) Received: from 130.20.128.21 by vscan2.pnl.gov (InterScan E-Mail VirusWall NT) ; Fri, 15 Dec 2000 12:20:40 -0800 (Pacific Standard Time) Received: by PNLMSE1.pnl.gov with Internet Mail Service (5.5.2448.0) id ; Fri, 15 Dec 2000 12:18:19 -0800 Date: Fri, 15 Dec 2000 12:18:14 -0800 From: "Mathieu, Thomas J" To: 'Peter Scott' , openafs-info@openafs.org, info-afs@transarc.com Message-id: <6DBB005AF9DDD31193030008C7A49DC82D402A@pnlmse8.pnl.gov> MIME-version: 1.0 X-Mailer: Internet Mail Service (5.5.2448.0) Content-type: text/plain; charset="iso-8859-1" Subject: [OpenAFS] RE: Web server over AFS Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hello Peter -- We run a webserver (Sun Ultra 80 class, 4 processor, currently Solaris 2.6). Only a handful of admins have access and it has a highly secure configuration -- no telnet or ftp access, etc., other than the webserver. It runs Apache web services and serves pages out of our AFS space. Several logical servers are supported -- and external server (which anyone can view at www.emsl.pnl.gov), an internal server limited to on site access, a secure server which allows authenticated external access to the internal only area, and a few project specific areas. The external web pages are maintained by a few web developers who know rules about what can be released and what can't. The internal pages are maintained by project web developers. CGI scripts are peer-reviewed for potential problems or policy violations. This is strictly enforced on the external site, more loosely enforced internally. Most file access is to databases which have secondary controls. Scripts aren't allowed write to the web server or AFS space, but we haven't extensively tested the capabilities here. We use AFS WebSecure product from IBM/Transarc to manage the secure web access controls. It works in conjunction with AFS groups to control access. After a users authenticate, they can access web pages as allowed by groups in which they are members. No bonus points. only one moderately loaded server taking about 1.5M hits per month. Tom Mathieu (tom.mathieu@pnl.gov) EMSL Computer Support Battelle, Pacific Northwest National Laboratory -----Original Message----- From: Peter Scott [mailto:Peter.J.Scott@jpl.nasa.gov] Sent: Thursday, December 14, 2000 6:54 PM To: openafs-info@openafs.org; info-afs@transarc.com Subject: Web server over AFS Hello. I am looking for input from anyone who has successfully run an institutional web server using AFS and provided users the ability to create web interfaces to persistent data. The hard part is ensuring that customer CGI programs are the only processes allowed to modify that data and that no-one else with an AFS account can get at it through CGI programs in their own directory. Non-AFS systems solve this with setuid, which is not available on AFS. I am talking about a centralized web server which is administered by AFS admins, no user access allowed to local disk space. We have a couple of theoretical solutions already, so I really want to constrain the answers to solutions that have been proven in practice. Bonus points if yours works with load balancing or round-robin type web server multiplexing. Thanks in advance. -- Peter Scott Peter.J.Scott@jpl.nasa.gov From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mindwarp.smart.net by penn.central.org id aa20730; 16 Dec 2000 10:04 EST Received: from mindwarp.smart.net (rlhamil@mindwarp.smart.net [192.168.1.4]) by mindwarp.smart.net (8.9.3+Sun/8.9.3/SpamBlocker) with ESMTP id KAA12086 for ; Sat, 16 Dec 2000 10:04:50 -0500 (EST) Message-ID: <3A3B8488.B3A043F3@mindwarp.smart.net> Date: Sat, 16 Dec 2000 10:04:40 -0500 From: "Richard L. Hamilton" Organization: Timetravellers Anonymous X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.8 sun4m) X-Accept-Language: en MIME-Version: 1.0 To: openafs-info@openafs.org Content-Type: multipart/mixed; boundary="------------E926203D754B3BCEA85A98CD" Subject: [OpenAFS] SPARC Solaris 8 binary? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: This is a multi-part message in MIME format. --------------E926203D754B3BCEA85A98CD Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Anyone have one packaged up (preferably _not_ for residence in /usr/local)? With only gcc, this isn't exactly self-service... -- ftp> get |fortune 377 I/O error: smart remark generator failed Bogonics: the primary language inside the Beltway mailto:rlhamil@mindwarp.smart.net http://www.smart.net/~rlhamil --------------E926203D754B3BCEA85A98CD Content-Type: text/x-vcard; charset=us-ascii; name="rlhamil.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Richard L. Hamilton Content-Disposition: attachment; filename="rlhamil.vcf" begin:vcard n:Hamilton;Richard x-mozilla-html:FALSE org:Timetravellers Anonymous;Improbability adr:;;;;;; version:2.1 email;internet:rlhamil@mindwarp.smart.net title:Mr. x-mozilla-cpt:;0 fn:Richard Hamilton end:vcard --------------E926203D754B3BCEA85A98CD-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from d06lmsgate-2.uk.ibm.com by penn.central.org id aa07418; 18 Dec 2000 15:45 EST Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22]) by d06lmsgate-2.uk.ibm.com (1.0.0) with ESMTP id UAA167868 for ; Mon, 18 Dec 2000 20:30:33 GMT Received: from de.ibm.com (lig32-239-94-137.emea.lig-dial.ibm.com [32.239.94.137]) by d12relay01.de.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id VAA79506 for ; Mon, 18 Dec 2000 21:43:02 +0100 Message-ID: <3A3E76D2.C359@de.ibm.com> Date: Mon, 18 Dec 2000 21:42:58 +0100 From: Markus Alt Organization: IBM Deutschland Entwicklung GmbH X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.16 i686) X-Accept-Language: en-US,de-DE MIME-Version: 1.0 To: openafs-info Mailing List Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] OpenAFS on SuSE 7.0 anyone? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi all, there has been some talk about other distributions apart from Red Hat lately (Mandrake, Debian, etc.). Now I'd like to know if someome has ever successfully compiled OpenAFS on the latest version of SuSE Linux, i.e. SuSE 7.0. From the experience I've made so far, I know that this isn't straightforward. There are a few patches that SuSE added to the pristine kernel sources and these patches also prevent the "regular" AFS client from starting successfully on SuSE 7.0. (Up to 6.4 this was no problem.) It looks like memory management - which has changed in 7.0 to support memory > 2 GB - is the culprit, at least as I understand it so far. I've already contacted Transarc's product support, but as there's only official support for Red Hat, they couldn't help me further. Now, when trying to compile OpenAFS on SuSE 7.0, the process exits at this stage: gcc -O2 -fomit-frame-pointer -fno-strength-reduce -pipe -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -D__KERNEL__ -DCPU=586 -DKERNEL -D_KERNEL -DMODULE -DAFS_SMP -I. -I../ -c ../afs/osi_vnodeops.c; ../afs/osi_vnodeops.c: In function `afs_linux_readpage': ../afs/osi_vnodeops.c:1135: structure has no member named `offset' ../afs/osi_vnodeops.c: In function `afs_linux_updatepage': ../afs/osi_vnodeops.c:1265: structure has no member named `offset' make[4]: *** [osi_vnodeops.o] Error 1 make[4]: Leaving directory `/build/openafs-1.0.1/i386_linux22/obj/libafs/MODLOAD-2.2.16-MP' make[3]: *** [linux_compdirs] Error 2 make[3]: Leaving directory `/build/openafs-1.0.1/i386_linux22/obj/libafs' make[2]: *** [libafs] Error 2 make[2]: Leaving directory `/build/openafs-1.0.1' make[1]: *** [install] Error 2 make[1]: Leaving directory `/build/openafs-1.0.1' make: *** [all] Error 2 The message "structure has no member named `offset'" leads to the same conclusion as above: The memory management in 7.0 has no more "offset" field in the struct "page" within /usr/src/linux/include/linux/mm.h - in contrast to 6.4. So finally: Has anyone already created some patches to either the OpenAFS sources or the SuSE 7.0 kernel sources in order to compile OpenAFS successfully? I will as well contact SuSE on this issue to see what they say. Of course, I could also use the vanilla kernel sources to build my kernel. But a) there are some really useful patches in the SuSE kernel sources and b) IMHO the other distributions will sooner or later also change their memory management to support bigger RAM size. At the latest then the OpenAFS sources will have to be adopted to support this. If someone has information he's willing to share or is also interested in getting OpenAFS to run on SuSE 7.0, then I'd like to hear from it. TIA, Markus -- Markus Alt IBM Lab Boeblingen, Germany altmark@de.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ECE.CMU.EDU by penn.central.org id aa07959; 18 Dec 2000 17:14 EST Received: from pyanfar (PYANFAR.REM.CMU.EDU [128.2.87.225]) by ece.cmu.edu (8.9.2/8.8.8) with ESMTP id RAA20461; Mon, 18 Dec 2000 17:10:26 -0500 (EST) Date: Mon, 18 Dec 2000 17:10:25 -0500 From: "Brandon S. Allbery KF8NH" To: Markus Alt , openafs-info Mailing List Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 anyone? Message-ID: <3880000.977177425@pyanfar> In-Reply-To: <3A3E76D2.C359@de.ibm.com> X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Monday, December 18, 2000 09:42:58 PM +0100, Markus Alt wrote: +----- | Hi all, | | there has been some talk about other distributions apart from Red Hat | lately (Mandrake, Debian, etc.). Now I'd like to know if someome has | ever successfully compiled OpenAFS on the latest version of SuSE Linux, | i.e. SuSE 7.0. +--->8 Not yet, but I had already gotten the local AFS gods to create a SuSE 7 AFS kernel module for me and I have requested the patches in an OpenAFS-compatible form; I'll send them on in when I get them. -- brandon s. allbery [os/2][linux][solaris][japh] allbery@kf8nh.apk.net system administrator [WAY too many hats] allbery@ece.cmu.edu electrical and computer engineering KF8NH carnegie mellon university ["better check the oblivious first" -ke6sls] From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SMTP2.ANDREW.CMU.EDU by penn.central.org id aa08138; 18 Dec 2000 18:00 EST Received: from TRIGATI.ANDREW.CMU.EDU (TRIGATI.ANDREW.CMU.EDU [128.2.122.7]) by smtp2.andrew.cmu.edu (8.9.3/8.9.3) with ESMTP id SAA29018 for ; Mon, 18 Dec 2000 18:00:26 -0500 (EST) Date: Mon, 18 Dec 2000 18:00:28 -0500 (EST) From: Chaskiel M Grundman To: openafs-info@openafs.org Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 anyone? In-Reply-To: <3A3E76D2.C359@de.ibm.com> Message-ID: MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="41944954-890487850-977180428=:26137" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. Send mail to mime@docserver.cac.washington.edu for more info. --41944954-890487850-977180428=:26137 Content-Type: TEXT/PLAIN; charset=US-ASCII > So finally: Has anyone already created some patches to either the > OpenAFS sources or the SuSE 7.0 kernel sources in order to compile > OpenAFS successfully? I've attached a patch which I believe will allow compilation on SuSE systems. I don't actually have any, and my attempts at using kernel headers extracted from SuSE rpms was not very successful, so I don't know that it will work.... The kernel patches in question do not have to do with large memory machine as much as supporting large files in the filesystem. (needless to say, afs does not support large files, so my patch did not actually add 64bit interfaces to afs, it merely extracts the 32 bit values it has always used.) In addition, I also adjusted an AFS structure that needs to be kept in sync with a linux struct inode. I would appreciate finding out if this works for you so I can consider submitting it to the openafs-gatekeeper (well, once I figure out how to identify LFS-patched kernels in a way other than #ifdef _REISER_FS_I) --41944954-890487850-977180428=:26137 Content-Type: TEXT/PLAIN; charset=US-ASCII; name="openafs-suse.diff" Content-Transfer-Encoding: BASE64 Content-ID: Content-Description: Content-Disposition: attachment; filename="openafs-suse.diff" ZGlmZiAtcnUgMDI4L2xpYmFmcy1PMS4wL2Fmcy9vc2lfdmZzLmggMDI5L2xp YmFmcy1PMS4wL2Fmcy9vc2lfdmZzLmgNCi0tLSAwMjgvbGliYWZzLU8xLjAv YWZzL29zaV92ZnMuaAlXZWQgTm92ICAxIDEzOjM4OjMzIDIwMDANCisrKyAw MjkvbGliYWZzLU8xLjAvYWZzL29zaV92ZnMuaAlXZWQgRGVjIDEzIDE0OjQz OjIzIDIwMDANCkBAIC0yOCw3ICsyOCwxMSBAQA0KIAl1aWRfdAkJCWlfdWlk Ow0KIAlnaWRfdAkJCWlfZ2lkOw0KIAlrZGV2X3QJCQlpX3JkZXY7DQorI2lm ZGVmIF9SRUlTRVJfRlNfSSAvKiBHYWNrLiBob3cgc2hvdWxkIEkgaWRlbnRp ZnkgTEZTIHBhdGNoPyAqLw0KKwlsb2ZmX3QJCQlpX3NpemU7DQorI2Vsc2UN CiAJb2ZmX3QJCQlpX3NpemU7DQorI2VuZGlmDQogCXRpbWVfdAkJCWlfYXRp bWU7DQogCXRpbWVfdAkJCWlfbXRpbWU7DQogCXRpbWVfdAkJCWlfY3RpbWU7 DQpkaWZmIC1ydSAwMjgvbGliYWZzLU8xLjAvYWZzL29zaV92bm9kZW9wcy5j IDAyOS9saWJhZnMtTzEuMC9hZnMvb3NpX3Zub2Rlb3BzLmMNCi0tLSAwMjgv bGliYWZzLU8xLjAvYWZzL29zaV92bm9kZW9wcy5jCVdlZCBOb3YgIDEgMTM6 Mzg6MzMgMjAwMA0KKysrIDAyOS9saWJhZnMtTzEuMC9hZnMvb3NpX3Zub2Rl b3BzLmMJV2VkIERlYyAxMyAxNDozODo0NyAyMDAwDQpAQCAtMjQsNiArMjQs MTIgQEANCiAjaW5jbHVkZSAiLi4vaC9tbS5oIg0KICNpbmNsdWRlICIuLi9o L3BhZ2VtYXAuaCINCiANCisjaWZkZWYgcGdvZmYydWxvbmcNCisjZGVmaW5l IHBhZ2VvZmYocHApIHBnb2ZmMnVsb25nKChwcCktPmluZGV4KQ0KKyNlbHNl DQorI2RlZmluZSBwYWdlb2ZmKHBwKSBwcC0+b2Zmc2V0DQorI2VuZGlmDQor DQogZXh0ZXJuIHN0cnVjdCB2Y2FjaGUgKmFmc19nbG9iYWxWcDsNCiANCiBl eHRlcm4gc3RydWN0IGRlbnRyeV9vcGVyYXRpb25zICphZnNfZG9wczsNCkBA IC05NzAsNyArOTc2LDcgQEANCiAgICAgc2V0X2JpdChQR19sb2NrZWQsICZw cC0+ZmxhZ3MpOyAvKiBvdGhlciBiaXRzPyBTZWUgbW0uaCAqLw0KICAgICBj bGVhcl9iaXQoUEdfZXJyb3IsICZwcC0+ZmxhZ3MpOw0KIA0KLSAgICBzZXR1 cF91aW8oJnR1aW8sICZpb3ZlYywgKGNoYXIqKWFkZHJlc3MsIHBwLT5vZmZz ZXQsIFBBR0VTSVpFLA0KKyAgICBzZXR1cF91aW8oJnR1aW8sICZpb3ZlYywg KGNoYXIqKWFkZHJlc3MsIHBhZ2VvZmYocHApLCBQQUdFU0laRSwNCiAJICAg ICAgVUlPX1JFQUQsIEFGU19VSU9TWVMpOw0KICAgICBjb2RlID0gYWZzX3Jk d3IoKHN0cnVjdCB2Y2FjaGUqKWlwLCAmdHVpbywgVUlPX1JFQUQsIDAsIGNy ZWRwKTsNCiANCkBAIC0xMDYxLDcgKzEwNjcsNyBAQA0KIAkgICAgICAgSUNM X1RZUEVfUE9JTlRFUiwgcHAsDQogCSAgICAgICBJQ0xfVFlQRV9JTlQzMiwg YXRvbWljX3JlYWQoJnBwLT5jb3VudCksDQogCSAgICAgICBJQ0xfVFlQRV9J TlQzMiwgOTk5OTkpOw0KLSAgICBzZXR1cF91aW8oJnR1aW8sICZpb3ZlYywg cGFnZV9hZGRyICsgb2Zmc2V0LCBwcC0+b2Zmc2V0ICsgb2Zmc2V0LCBjb3Vu dCwNCisgICAgc2V0dXBfdWlvKCZ0dWlvLCAmaW92ZWMsIHBhZ2VfYWRkciAr IG9mZnNldCwgcGFnZW9mZihwcCkgKyBvZmZzZXQsIGNvdW50LA0KIAkgICAg ICBVSU9fV1JJVEUsIEFGU19VSU9TWVMpOw0KIA0KICAgICBjb2RlID0gYWZz X3dyaXRlKHZjcCwgJnR1aW8sIGZwLT5mX2ZsYWdzLCBjcmVkcCwgMCk7DQo= --41944954-890487850-977180428=:26137-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from delta.rollanet.org by penn.central.org id aa08383; 18 Dec 2000 19:26 EST Received: (qmail 22639 invoked from network); 19 Dec 2000 00:26:28 -0000 Received: from cessna.rollanet.org (HELO umr.edu) (208.18.13.139) by mx.rollanet.org with SMTP; 19 Dec 2000 00:26:28 -0000 Message-ID: <3A3EAB33.EAFB1E9D@umr.edu> Date: Mon, 18 Dec 2000 18:26:27 -0600 From: Nathan Neulinger X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.18 i686) X-Accept-Language: en MIME-Version: 1.0 To: openafs-info@openafs.org, info-afs@transarc.com Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] Transarc AFS 3.6 2.5 server release AAARRRGGGHHH!!!! Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Today we did a major AFS server upgrade that has been delayed for far too long. Unfortunately thanks to an (apparently, I might have missed it, but it sucks regardless) undocumented new feature in the 2.5 patch release - we wound up having problems for about 8 hours with NT clients unable to talk to the server. Apparently the 2.5 code has a nwe delta in it that causes servers to stop talking to clients completely when it thinks that they are flooding the server with requests. There is no way to tell that this is happening, no way to shut it off, and no way to get a list of affected stations. In our case, almost 1500 NT stations with the AFS client had extremely sporadic and unstable access to afs. What would happen is - the fs checks output woult include all of the servers running 2.5, or some selection of them. So - if you're thinking of upgrading to 3.6 2.5, think twice, or at least be very cautious about it. Shutting down all your clients ahead of time, and SLOWLY bringing them back up might help, but that's hardly an option with 1500 stations. The end result - we wound up backing down to 3.6 2.3 (a huge upgrade from the 3.4a 5.53 we were running), which fixed the problem. (To openafs gatekeepers - if you get a delta from transarc/ibm [yeah right!] that includes this, I _STRONGLY_ suggest that you say 'no thanks!' or at the very least, make the feature optional.) BTW - this doesn't seem to affect unix clients. It only affected the NT clients, runnig 3.5 or 3.6, didn't seem to matter which, although the particular behavior differed with 3.5 and 3.6. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from d12lmsgate.de.ibm.com by penn.central.org id aa11504; 19 Dec 2000 5:17 EST Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22]) by d12lmsgate.de.ibm.com (1.0.0) with ESMTP id LAA29994 for ; Tue, 19 Dec 2000 11:17:10 +0100 Received: from de.ibm.com (altmark.boeblingen.de.ibm.com [9.164.181.140]) by d12relay01.de.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id LAA84392 for ; Tue, 19 Dec 2000 11:17:10 +0100 Message-ID: <3A3F35A4.63391C37@de.ibm.com> Date: Tue, 19 Dec 2000 11:17:08 +0100 From: Markus Alt Organization: IBM Deutschland Entwicklung GmbH X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.16 i686) X-Accept-Language: en-US,de-DE MIME-Version: 1.0 To: openafs-info Mailing List Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 anyone? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Chaskiel M Grundman wrote: > [snip] > > In addition, I also adjusted an AFS structure that needs to be kept in > sync with a linux struct inode. I would appreciate finding out if this > works for you so I can consider submitting it to the openafs-gatekeeper > (well, once I figure out how to identify LFS-patched kernels in a way > other than #ifdef _REISER_FS_I) Thanks for the patch - of course I will try if it works and let you know how far I got. Just one question before I start: My system has a kernel which supports ReiserFS and in fact I also use ReiserFS for my filesystems. Do you see any difficulties in that? Regards, Markus -- Markus Alt IBM Lab Boeblingen, Germany altmark@de.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from FORT-POINT-STATION.MIT.EDU by penn.central.org id aa13180; 19 Dec 2000 10:03 EST Received: from grand-central-station.MIT.EDU (GRAND-CENTRAL-STATION.MIT.EDU [18.69.0.34]) by fort-point-station.mit.edu (8.9.2/8.9.2) with ESMTP id KAA27653; Tue, 19 Dec 2000 10:03:13 -0500 (EST) Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id KAA12808; Tue, 19 Dec 2000 10:03:13 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id KAA01533; Tue, 19 Dec 2000 10:03:12 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id KAA17169; Tue, 19 Dec 2000 10:03:12 -0500 (EST) To: Markus Alt Cc: openafs-info Mailing List Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 anyone? References: <3A3F35A4.63391C37@de.ibm.com> From: Derek Atkins Date: 19 Dec 2000 10:03:12 -0500 In-Reply-To: Markus Alt's message of "Tue, 19 Dec 2000 11:17:08 +0100" Message-ID: Lines: 19 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Markus Alt writes: > Thanks for the patch - of course I will try if it works and let you know > how far I got. Just one question before I start: My system has a kernel > which supports ReiserFS and in fact I also use ReiserFS for my > filesystems. Do you see any difficulties in that? There is a known problem with a slow-start on Reiserfs if your AFS Cache is on e.g. your root partition. The problem is that iget() is slow on large partitions. The only workaround is to have AFS on its own partition. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE3262F.dip.t-dialin.net by penn.central.org id aa09747; 18 Dec 2000 22:33 EST Received: (qmail 13988 invoked from network); 19 Dec 2000 03:32:54 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 19 Dec 2000 03:32:54 -0000 Received: (qmail 2711 invoked by uid 1000); 19 Dec 2000 03:32:53 -0000 Date: Tue, 19 Dec 2000 04:32:53 +0100 From: Sascha Silbe To: openafs-info@openafs.org Message-ID: <20001219043252.A2595@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="RnlQjJ0d97Da+TV1" Content-Disposition: inline User-Agent: Mutt/1.3.9i Subject: [OpenAFS] AFS-Client behind masquerading firewall Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --RnlQjJ0d97Da+TV1 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hello! I'm trying to get the OpenAFS client to work behind a masquerading firewall (PPP dial-up with dynamic IP address that changes every 24 hours). For the first few minutes it works perfectly, but after a while it just complains about the network being down. :( Are there any options that could cause this problem? I saw that only UDP is used. Does AFS also work over TCP (don't know if the protocol supports it and couldn't find any info on this topic)? If the answer is yes, how do I g= et=20 the OpenAFS client to use TCP? CU/Lnx Sascha --RnlQjJ0d97Da+TV1 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6Ptbka8hvv2E1w1sRArSJAKCELnsYyggfB6sEIEPB+9GQzT68WACdHQYF WYSBoqSlKeUqWXJHWnsVcB4= =2UN/ -----END PGP SIGNATURE----- --RnlQjJ0d97Da+TV1-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from egghead.curl.com by penn.central.org id aa13767; 19 Dec 2000 11:35 EST Received: (qmail 13716 invoked by uid 10171); 19 Dec 2000 11:35:50 -0500 To: Info-AFS@transarc.com, openafs-info@openafs.org From: patl@curl.com (Patrick J. LoPresti) Date: 19 Dec 2000 11:35:50 -0500 Message-ID: Lines: 43 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: [OpenAFS] AFS and Macintosh Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: (I apologize if you get this message twice; it did not appear to go through to info-afs the first time.) We are considering deploying AFS at our organization. In addition to Linux, Solaris, and Windows clients, we have to worry about Macs; specifically, MacOS 8.5, 9, and (soon) X. We can probably wait for the native OpenAFS client for OS X to stabilize. But we need to support MacOS 8.5 and 9 right away. I have searched the info-afs and netatalk-admins archives. I have browsed the UMich RSUG pages. I have stumbled onto AuthMan and the AuthMan UAM: http://www.umich.edu/~rsug/authman/ ...but some messages from the netatalk-admins archive suggest that AuthMan may not work with OS 8.5 and higher. I see a link to a slightly different version of the same tools at this URL: http://rsug.itd.umich.edu/software/mackerb.html However, I am not even sure AuthMan/AuthMan UAM are what I need. So my question is this: What is the best way to provide access to AFS for Macs? What software do I need, and where should I obtain it? Is the 1.5 prerelease of netatalk at http://netatalk.sourceforge.net/ suitable? What do most organizations with AFS and Macs do? (Note: In our environment, we do not particularly mind sending cleartext passwords over the network. And we do not care about any services except AFS. So Kerberos on Mac is not a requirement; a version of netatalk which simply used the cleartext password to get AFS tokens would be fine.) Thanks. - Pat From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SOUTH-STATION-ANNEX.MIT.EDU by penn.central.org id aa14928; 19 Dec 2000 13:29 EST Received: from GRAND-CENTRAL-STATION.MIT.EDU by MIT.EDU with SMTP id AA18529; Tue, 19 Dec 00 13:28:46 EST Received: from melbourne-city-street.MIT.EDU (MELBOURNE-CITY-STREET.MIT.EDU [18.69.0.45]) by grand-central-station.MIT.EDU (8.9.2/8.9.2) with ESMTP id NAA24380; Tue, 19 Dec 2000 13:29:42 -0500 (EST) Received: from indiana.mit.edu (INDIANA.MIT.EDU [18.18.1.138]) by melbourne-city-street.MIT.EDU (8.9.3/8.9.2) with ESMTP id NAA18122; Tue, 19 Dec 2000 13:29:41 -0500 (EST) Received: (from warlord@localhost) by indiana.mit.edu (8.9.3) id NAA17205; Tue, 19 Dec 2000 13:29:41 -0500 (EST) To: Sascha Silbe Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall References: <20001219043252.A2595@cube.sascha.silbe.org> From: Derek Atkins Date: 19 Dec 2000 13:29:41 -0500 In-Reply-To: Sascha Silbe's message of "Tue, 19 Dec 2000 04:32:53 +0100" Message-Id: Lines: 31 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sascha Silbe writes: > Hello! > > I'm trying to get the OpenAFS client to work behind a masquerading firewall > (PPP dial-up with dynamic IP address that changes every 24 hours). For the > first few minutes it works perfectly, but after a while it just complains > about the network being down. :( > Are there any options that could cause this problem? I saw that only UDP is > used. Does AFS also work over TCP (don't know if the protocol supports it > and couldn't find any info on this topic)? If the answer is yes, how do I g= > et=20 > the OpenAFS client to use TCP? > > CU/Lnx Sascha AFS does not currently support TCP operations. AFS _does_ work behind a masquerading firewall provided that you set the UDP timeouts high enough to allow callbacks to occur. I would recommend UDP timeouts in the range of 10-15 minutes. Note that this requires changes the firewall, not your client. You should also note that AFS might "hiccup" when the IP address changes, at least until new callbacks can be set with the new IP. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from d12lmsgate-3.de.ibm.com by penn.central.org id aa15655; 19 Dec 2000 15:40 EST Received: from d12relay02.de.ibm.com (d12relay02.de.ibm.com [9.165.215.23]) by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id VAA132404 for ; Tue, 19 Dec 2000 21:40:53 +0100 Received: from de.ibm.com (lig32-239-94-57.emea.lig-dial.ibm.com [32.239.94.57]) by d12relay02.de.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id VAA181666; Tue, 19 Dec 2000 21:40:51 +0100 Message-ID: <3A3FC7D2.96270963@de.ibm.com> Date: Tue, 19 Dec 2000 21:40:50 +0100 From: Markus Alt Organization: IBM Deutschland Entwicklung GmbH X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.16 i686) X-Accept-Language: en-US,de-DE MIME-Version: 1.0 To: openafs-info Mailing List Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 anyone? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Chaskiel M Grundman wrote: > > > So finally: Has anyone already created some patches to either the > > OpenAFS sources or the SuSE 7.0 kernel sources in order to compile > > OpenAFS successfully? > I've attached a patch which I believe will allow compilation on SuSE > systems. I don't actually have any, and my attempts at using kernel > headers extracted from SuSE rpms was not very successful, so I don't know > that it will work.... The kernel patches in question do not have to do Great, it *does* work! Thanks a lot! Even on my system, where all partitions (except for /boot) have ReiserFS filesystems [1], there seems to be no problem so far. Of course I will some further testing and post any problems I may encounter to this list. Thanks again and best regards, Markus [1] For the AFS cache's partition, I tried with an Ext2 filesystem first, but after switching to ReiserFS again, it also works flawlessly. -- Markus Alt IBM Lab Boeblingen, Germany altmark@de.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa15686; 19 Dec 2000 15:47 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.37.121]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id OAA29848; Tue, 19 Dec 2000 14:47:32 -0600 Received: by umr-mail01 with Internet Mail Service (5.5.2650.21) id ; Tue, 19 Dec 2000 14:47:32 -0600 Message-ID: <9DA8D24B915BD1118911006094516EAF053E5CFD@umr-mail02> From: "Neulinger, Nathan R." To: "'Markus Alt'" , openafs-info Mailing List Subject: RE: [OpenAFS] OpenAFS on SuSE 7.0 anyone? Date: Tue, 19 Dec 2000 14:47:30 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > [1] For the AFS cache's partition, I tried with an Ext2 filesystem > first, but after switching to ReiserFS again, it also works > flawlessly. FYI - we've been running quite a number of machines pure-reiserfs with afs without any trouble at all. One definite benefit - being able to build a 500 MB/50,000 file afs cache in a 30 seconds sure is nice. -- Nathan From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from d06lmsgate-3.uk.ibm.com by penn.central.org id aa16192; 19 Dec 2000 17:22 EST Received: from d06relay02.portsmouth.uk.ibm.com (d06relay02.portsmouth.uk.ibm.com [9.166.84.148]) by d06lmsgate-3.uk.ibm.com (1.0.0) with ESMTP id WAA09628 for ; Tue, 19 Dec 2000 22:14:22 GMT Received: from de.ibm.com (lig32-239-94-98.emea.lig-dial.ibm.com [32.239.94.98]) by d06relay02.portsmouth.uk.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id WAA249100 for ; Tue, 19 Dec 2000 22:22:46 GMT Message-ID: <3A3FDFB4.4F3B0C98@de.ibm.com> Date: Tue, 19 Dec 2000 23:22:44 +0100 From: Markus Alt Organization: IBM Deutschland Entwicklung GmbH X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.16 i686) X-Accept-Language: en-US,de-DE MIME-Version: 1.0 To: openafs-info Mailing List Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] OpenAFS on SuSE 7.0 - not yet! Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hmm, looks like I've been a little too happy about the success I apparently had ... After some further testing I found out that there are severe problems with all files > 4096 bytes: After 4096 bytes, the file's content starts over again with its beginning, then 4096 bytes later again and so on. I first noticed this with files like .alias or .Xdefaults as they looked like crap. Meanwhile I tried ReiserFS and Ext2 for the filesystem on the AFS cache's partition, as well as a memory cache - no success. ("fs flush" and "fs checkv" don't help either, of course.) All of the commands I've tested so far work (some of the bos, vos and fs suites' subcommands), but without a functioning cache manager, this is rather useless ... Any words of wisdom? Regards, Markus -- Markus Alt IBM Lab Boeblingen, Germany altmark@de.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from e21.nc.us.ibm.com by penn.central.org id aa04469; 22 Dec 2000 9:28 EST Received: from southrelay02.raleigh.ibm.com (southrelay02.raleigh.ibm.com [9.37.3.209]) by e21.nc.us.ibm.com (8.9.3/8.9.3) with ESMTP id JAA26528 for ; Fri, 22 Dec 2000 09:25:01 -0600 Received: from d04nm201.raleigh.ibm.com (d04nm201.raleigh.ibm.com [9.67.228.38]) by southrelay02.raleigh.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id JAA67746; Fri, 22 Dec 2000 09:28:07 -0500 Importance: Normal Subject: Re: [OpenAFS] Transarc AFS 3.6 2.5 server release AAARRRGGGHHH!!!! To: openafs-info@openafs.org, info-afs@transarc.com X-Mailer: Lotus Notes Release 5.0.5 September 22, 2000 Message-ID: From: "Jared Spencer" Date: Fri, 22 Dec 2000 09:28:08 -0500 X-MIMETrack: Serialize by Router on D04NM201/04/M/IBM(Release 5.0.3 (Intl)|21 March 2000) at 12/22/2000 09:28:09 AM MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Here is some more information on this problem: First, it appears that the scenario of NT clients being unable to access volumes on a fileserver is specific to a 3.6 Patch 1 fileserver and a *GA release* NT client. Other NT clients (and UNIX clients for that matter) could become internally marked down by a fileserver, but this would only be under specific circumstances and is recoverable from. Read on.... AFS 3.6 patch 1 contains a delta to address Denial of Service types of attacks from rogue clients. The release notes generically describe this as: This fix resolves a problem that occurred when an AFS File Server received requests from AFS Client machines to which it could not respond. The requests locked up threads in the File Server and rendered the server unusable. Basically, if the fileserver is unable to respond to a client, it exponentially backs off the responses until it eventually marks the client as down. Once a client is marked down by the fileserver, that entry can be cleared by restarting the fileserver. If you suspect that a fileserver has marked a client down, one of the initial symptoms is that fs checks returns quickly showing the fileserver down when, in fact, it is up. You can verify further by running tcpdump and looking for the following: [jared] tcpdump -l -N -s 256 port 7000 and host cube and host afstest58 tcpdump: listening on hme0 10:15:41.531302 afstest58.cm.e9052e9c.0460 > cube.fs: 1.046e data C.L. 4 GetTime 10:15:41.532957 afstest58.cm.e9052e9c.0460 < cube.fs: 0.0407 abort .... 4 -1 (DF) Here, the fileserver has immediately returned an abort to the GetTime request from the client. This is a good indication the fileserver has this client marked down. Under normal circumstances, there are very few scenarios where a "good" client will get marked down. And, if this happens, the fileserver can be restarted to clear the entry for that host. But enter the 3.5 GA release NT client. In the 3.5 GA release NT client, there was a bug that would cause it to fail to return an appropriate code to a fileserver's WhoAreYou requests. At the time, this was basically a benign bug. The client still functioned normally, and the only indication that the call was failing was from the WhoAreYou failed messages in the FileLog. However, with the new 3.6 Patch 1 fileserver, we believe it's now seeing these failed RPCs as failure to respond, and is marking the client as down. Although restarting the fileserver clears the host entry, in this case it will just quickly get marked down again. This NT client bug was fixed in 3.5 Patch 1. This is a preliminary analysis, and we're continuing to investigate this here. If you have any further questions, please contact AFS support. Also, there was a response to this thread identifying the -dontdelay switch that has been added. This is actually unrelated to the DoS delta. This switch was to address a situation where clients would spend a long time trying to stat a directory in which they didn't have appropriate permissions, and this fix is not yet in the released code for either 3.5 or 3.6. Thanks, Jared ~~~~~~~~~~~~~~~~~~~~~~~~~~~ Jared Spencer AFS Technical Lead Staff Software Engineer IBM Transarc Labs jareds@us.ibm.com ~~~~~~~~~~~~~~~~~~~~~~~~~~~ Nathan Neulinger @openafs.org on 12/18/2000 07:26:27 PM Sent by: openafs-info-admin@openafs.org To: openafs-info@openafs.org, info-afs@transarc.com cc: Subject: [OpenAFS] Transarc AFS 3.6 2.5 server release AAARRRGGGHHH!!!! Today we did a major AFS server upgrade that has been delayed for far too long. Unfortunately thanks to an (apparently, I might have missed it, but it sucks regardless) undocumented new feature in the 2.5 patch release - we wound up having problems for about 8 hours with NT clients unable to talk to the server. Apparently the 2.5 code has a nwe delta in it that causes servers to stop talking to clients completely when it thinks that they are flooding the server with requests. There is no way to tell that this is happening, no way to shut it off, and no way to get a list of affected stations. In our case, almost 1500 NT stations with the AFS client had extremely sporadic and unstable access to afs. What would happen is - the fs checks output woult include all of the servers running 2.5, or some selection of them. So - if you're thinking of upgrading to 3.6 2.5, think twice, or at least be very cautious about it. Shutting down all your clients ahead of time, and SLOWLY bringing them back up might help, but that's hardly an option with 1500 stations. The end result - we wound up backing down to 3.6 2.3 (a huge upgrade from the 3.4a 5.53 we were running), which fixed the problem. (To openafs gatekeepers - if you get a delta from transarc/ibm [yeah right!] that includes this, I _STRONGLY_ suggest that you say 'no thanks!' or at the very least, make the feature optional.) BTW - this doesn't seem to affect unix clients. It only affected the NT clients, runnig 3.5 or 3.6, didn't seem to matter which, although the particular behavior differed with 3.5 and 3.6. -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 CIS - Systems Programming Fax: (573) 341-4216 _______________________________________________ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo.cgi/openafs-info From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from e24.nc.us.ibm.com by penn.central.org id aa05540; 22 Dec 2000 12:05 EST Received: from southrelay02.raleigh.ibm.com (southrelay02.raleigh.ibm.com [9.37.3.209]) by e24.nc.us.ibm.com (8.9.3/8.9.3) with ESMTP id MAA14826 for ; Fri, 22 Dec 2000 12:06:15 -0600 From: kkassab@us.ibm.com Received: from d54mta01.raleigh.ibm.com (d54mta01.raleigh.ibm.com [9.67.228.33]) by southrelay02.raleigh.ibm.com (8.8.8m3/NCO v4.95) with SMTP id MAA78278 for ; Fri, 22 Dec 2000 12:04:55 -0500 Received: by d54mta01.raleigh.ibm.com(Lotus SMTP MTA v4.6.5 (863.2 5-20-1999)) id 852569BD.005DD0EC ; Fri, 22 Dec 2000 12:04:43 -0500 X-Lotus-FromDomain: IBMUS To: openafs-info@openafs.org Message-ID: <852569BD.005DCFE5.00@d54mta01.raleigh.ibm.com> Date: Fri, 22 Dec 2000 12:03:55 -0500 Mime-Version: 1.0 Content-type: text/plain; charset=us-ascii Content-Disposition: inline Subject: [OpenAFS] AFS initial setup Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi all, I have been having a hard time with this and hope anyone can help. I'm still in the initial stages with configuring an initial AFS cell on RH 6.2 and I'm currently on the step to create Authentication Database entries. Within kas, I get "Creating user afs : [____] server or network not responding" when I issue "create afs". And after I look at the BosLog I see this: Fri Dec 22 12:38:11 2000: Server directory access is okay Fri Dec 22 12:38:11 2000: BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'vlserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'ptserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'buserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'ptserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'vlserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'buserver' repeatedly failed to start, perhaps missing executable. I have started the bosserver with the -noauth flag along with all my other commands. I have followed the QkBegin.pdf exactly and everything appears to be set up exactly as they presented it. I have looked in the archives a bit and saw that Jatin Nansi had a similar problem but it looked like it was related to having no DNS entry. I do and I don't see any bind errors anywhere. Is there something I'm missing, thanks in advance for any help I receive -Kevin ****************** Kevin Kassab Systems Engineer - Adv. Internet Technology Phone: 203 486 5608 E-mail: kkassab@us.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from pc124121.stofanet.dk by penn.central.org id aa20641; 20 Dec 2000 8:09 EST Received: (qmail 9815 invoked by uid 1000); 20 Dec 2000 13:09:36 -0000 Date: 20 Dec 2000 13:09:36 -0000 Message-ID: <20001220130936.9814.qmail@karthy.net> From: Karsten Thygesen To: openafs-info@openafs.org Subject: [OpenAFS] NT access to servers Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi I'm in the process of setting up a new cell, and I'm using afs on Solaris as file/database server. The server is behind a PIX firewall, which does NAT. The cell is accessible by sun, linux and windows-2K clients from the local network (behind the NAT), and everything works great. From outside the firewall, everything works great for Linux and Sun clients, but the NT/2K clients barf. They can acquire a token, but not access any files. The firewall is open for: UDP 7000-7008 (from the AFS FAQ) UDP 750 <-- kerberos 4 - required to make NT get a token!!! UDP 123 (ntp - is that required?) UDP xlate timeout is 15 minuttes. The NT box just claims, that the server is paused or not running. Any ideas???? According to the firewall log, it does not seem, that it is rejecting anything... Best regards, Karsten From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mercury.ccmr.cornell.edu by penn.central.org id aa07777; 22 Dec 2000 19:19 EST Received: from ruby.ccmr.cornell.edu (IDENT:0@ruby.ccmr.cornell.edu [128.84.231.115]) by mercury.ccmr.cornell.edu (8.9.3/8.9.3) with ESMTP id TAA23508; Fri, 22 Dec 2000 19:19:15 -0500 Received: from localhost (mitch@localhost) by ruby.ccmr.cornell.edu (8.9.3/8.9.3) with ESMTP id TAA18662; Fri, 22 Dec 2000 19:19:13 -0500 X-Authentication-Warning: ruby.ccmr.cornell.edu: mitch owned process doing -bs Date: Fri, 22 Dec 2000 19:19:13 -0500 (EST) From: Mitch Collinsworth To: Karsten Thygesen cc: openafs-info@openafs.org Subject: Re: [OpenAFS] NT access to servers In-Reply-To: <20001220130936.9814.qmail@karthy.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On 20 Dec 2000, Karsten Thygesen wrote: > The NT box just claims, that the server is paused or not running. This sounds suspiciously like the recently discussed NT client problem. What version of AFS are you running on your servers and on the NT clients that think the server is paused? -Mitch From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from PO10.ANDREW.CMU.EDU by penn.central.org id aa14613; 24 Dec 2000 22:21 EST Received: (from postman@localhost) by po10.andrew.cmu.edu (8.9.3/8.9.3) id WAA00931 for openafs-info@openafs.org; Sun, 24 Dec 2000 22:21:29 -0500 (EST) Received: via switchmail; Sun, 24 Dec 2000 22:21:28 -0500 (EST) Received: from trigati.andrew.cmu.edu via qmail ID ; Sun, 24 Dec 2000 22:20:24 -0500 (EST) Received: from trigati.andrew.cmu.edu via qmail ID ; Sun, 24 Dec 2000 22:20:22 -0500 (EST) Received: from mms.4.60.Apr..3.2000.21.32.31.i386.Linux.22.EzMail.2.0.CUILIB.3.45.SNAP.NOT.LINKED.trigati.andrew.cmu.edu.i386.linux22 via MS.5.6.trigati.andrew.cmu.edu.i386_Linux_22; Sun, 24 Dec 2000 22:20:22 -0500 (EST) Message-ID: Date: Sun, 24 Dec 2000 22:20:22 -0500 (EST) From: Chaskiel M Grundman To: openafs-info@openafs.org Subject: Re: [OpenAFS] OpenAFS on SuSE 7.0 - not yet! Cc: In-Reply-To: <3A3FDFB4.4F3B0C98@de.ibm.com> References: <3A3FDFB4.4F3B0C98@de.ibm.com> Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Excerpts from internet.computing.openafs.info: 19-Dec-100 [OpenAFS] OpenAFS on SuSE 7.. by Markus Alt@de.ibm.com > Hmm, looks like I've been a little too happy about the success I > apparently had ... Upon further reflection of the suse kernel sources, it looks like my patch references the wrong function. in the post-patch (src/afs/LINUX/)osi_vnodeops.c, you want to change #define pageoff(pp) pgoff2ulong((pp)->index) to #define pageoff(pp) pgoff2loff((pp)->index) I've tested this on a freshly installed suse 7.0 system, and it seems to work. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from rmx470-mta.mail.com by penn.central.org id aa14917; 25 Dec 2000 0:29 EST Received: from web170-mc (web170-mc.mail.com [165.251.48.146]) by rmx470-mta.mail.com (8.9.3/8.9.3) with SMTP id AAA03780; Mon, 25 Dec 2000 00:29:31 -0500 (EST) Message-ID: <384024738.977722171186.JavaMail.root@web170-mc> Date: Mon, 25 Dec 2000 00:29:31 -0500 (EST) From: Jatin Nansi To: kkassab@us.ibm.com, openafs-info@openafs.org Subject: RE: [OpenAFS] AFS initial setup Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: mail.com X-Originating-IP: 203.199.42.195 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: most likely u have a bad hosts file. just make sure that ur machine points to the correct IP and not to lo0. i could never understand tho why afs refuses to run on loopback and will run nicely on a real ip. any comments afs masters?? Jatin Have a very merry christmas ------Original Message------ From: kkassab@us.ibm.com To: openafs-info@openafs.org Sent: December 22, 2000 5:03:55 PM GMT Subject: [OpenAFS] AFS initial setup Hi all, I have been having a hard time with this and hope anyone can help. I'm still in the initial stages with configuring an initial AFS cell on RH 6.2 and I'm currently on the step to create Authentication Database entries. Within kas, I get "Creating user afs : [____] server or network not responding" when I issue "create afs". And after I look at the BosLog I see this: Fri Dec 22 12:38:11 2000: Server directory access is okay Fri Dec 22 12:38:11 2000: BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'vlserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'ptserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'buserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'ptserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'vlserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'kaserver' repeatedly failed to start, perhaps missing executable. Fri Dec 22 12:38:11 2000: BNODE 'buserver' repeatedly failed to start, perhaps missing executable. I have started the bosserver with the -noauth flag along with all my other commands. I have followed the QkBegin.pdf exactly and everything appears to be set up exactly as they presented it. I have looked in the archives a bit and saw that Jatin Nansi had a similar problem but it looked like it was related to having no DNS entry. I do and I don't see any bind errors anywhere. Is there something I'm missing, thanks in advance for any help I receive -Kevin ****************** Kevin Kassab Systems Engineer - Adv. Internet Technology Phone: 203 486 5608 E-mail: kkassab@us.ibm.com _______________________________________________ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo.cgi/openafs-info ________________________________________ LINUX - Its a way of life ________________________________________ ------------------------------------ FREE Personalized Email at India.com Sign up at http://mail.india.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from pc124121.stofanet.dk by penn.central.org id aa23993; 27 Dec 2000 6:15 EST Received: (qmail 14713 invoked by uid 1000); 27 Dec 2000 11:15:44 -0000 To: Mitch Collinsworth Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] NT access to servers References: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit From: Karsten Thygesen Date: 27 Dec 2000 12:15:44 +0100 In-Reply-To: Mitch Collinsworth's message of "Fri, 22 Dec 2000 19:19:13 -0500 (EST)" Message-ID: <87puie2jz3.fsf@karthy.net> Lines: 25 User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7 MIME-Version: 1.0 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >>>>> "Mitch" == Mitch Collinsworth writes: Mitch> On 20 Dec 2000, Karsten Thygesen wrote: >> The NT box just claims, that the server is paused or not running. Mitch> This sounds suspiciously like the recently discussed NT client Mitch> problem. What version of AFS are you running on your servers Mitch> and on the NT clients that think the server is paused? I think I found the problem. My server is behind a NAT-firewall, and it have several internal interfaces. _One_ of them is mapped externally. The VOS server seems to report all of the interfaces as available for fileserver access, and unix clients is intelligent enough to only contact the external mapped interfaces when communicating from outside the firewall. NT clients try to access one of the internal adresses instead, which, of cause, is not available. The solution is to use the NetRestrict file and disable the other external-non-accessible interfaces - then it works with both unix and nt clients. However - I still wonder why there is a differnce in semantics... Karsten From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from d06lmsgate.uk.ibm.com by penn.central.org id aa06263; 29 Dec 2000 7:16 EST Received: from d06relay02.portsmouth.uk.ibm.com (d06relay02.portsmouth.uk.ibm.com [9.166.84.148]) by d06lmsgate.uk.ibm.COM (1.0.0) with ESMTP id LAA62656; Fri, 29 Dec 2000 11:56:03 GMT Received: from de.ibm.com (lig32-239-94-245.emea.lig-dial.ibm.com [32.239.94.245]) by d06relay02.portsmouth.uk.ibm.com (8.8.8m3/NCO v4.95) with ESMTP id MAA260712; Fri, 29 Dec 2000 12:10:47 GMT Message-ID: <3A4C7F43.F1D4736E@de.ibm.com> Date: Fri, 29 Dec 2000 13:10:43 +0100 From: Markus Alt Organization: IBM Deutschland Entwicklung GmbH X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.17 i686) X-Accept-Language: en-US,de-DE MIME-Version: 1.0 To: openafs-info Mailing List , cg2v@andrew.cmu.edu Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [OpenAFS] OpenAFS on SuSE 7.0 - finally! Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Although I had switched to stock 2.2.17 with ReiserFS patches applied meanwhile (where "standard" IBM AFS works fine), I've tried Chaskiel's modified patch now on a SuSE 7.0 system (2.2.16 built from SuSE kernel sources). It works great (so far ;-) ), the oddity with files' repeating contents after 4 kB has gone! Thanks much for the patch, Chaskiel. I think you should consider submitting it to openafs-gatekeeper, even with that #ifdef _REISER_FS_I LFS hack. People running kernel compiled from SuSE modified kernel sources will definetely appreciate it. Needless to say that I will post problems/oddities/etc. here in case there are any. I'd prefer if they stay away ... Regards, Markus -- Markus Alt IBM Lab Boeblingen, Germany altmark@de.ibm.com From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.tpi.pl by penn.central.org id aa06833; 29 Dec 2000 8:50 EST Received: from mail.tpi.pl (mail.tpi.pl [213.25.79.65]) by tpi.pl (Postfix) with ESMTP id 3A5881AF for ; Fri, 29 Dec 2000 14:50:16 +0100 (CET) Date: Fri, 29 Dec 2000 14:50:16 +0100 (CET) From: Anna Szczechura To: openafs-info@openafs.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] Problem with klog Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I'm trying to setup AFS on RH6.2 using Quick Start for Beginners docs, getting some problems when trying to authorize as admin(or any other created user): [root@czatnius local]# klog -principal admin -password xxxxxxx Unable to authenticate to AFS because no such entry. I've been looking for any notices about this a problem without succes. I can get those information when running bosserver -noauth: [root@czatnius local]# pts membership admin Groups admin (id: 2) is a member of: system:administrators [root@czatnius local]# bos listkeys -server czatnius key 0 has cksum 1358133947 Keys last changed on Fri Dec 29 12:47:13 2000. All done. [root@czatnius local]# kas examin afs -noauth User data for afs key (2) cksum is 1358133947, last cpw: Fri Dec 29 12:46:48 2000 password will never expire. An unlimited number of unsuccessful authentications is permitted. entry never expires. Max ticket lifetime 100.00 hours. last mod on Fri Dec 29 09:08:26 2000 by permit password reuse [root@czatnius local]# kas examin admin -noauth User data for admin (ADMIN) key (2) cksum is 1895119345, last cpw: Fri Dec 29 13:35:41 2000 password will never expire. An unlimited number of unsuccessful authentications is permitted. entry never expires. Max ticket lifetime 25.00 hours. last mod on Fri Dec 29 12:44:22 2000 by permit password reuse end.. everything seems to be ok? Beside this odd message explained above.. (I've already tried to use klog -setpag and pagsh..) best regards Anna Szczechura From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa10804; 30 Dec 2000 16:39 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id JAA27942; Sat, 30 Dec 2000 09:33:20 -0500 (EST) To: Anna Szczechura Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem with klog References: From: Sam Hartman Date: 30 Dec 2000 09:33:20 -0500 In-Reply-To: Anna Szczechura's message of "Fri, 29 Dec 2000 14:50:16 +0100 (CET)" Message-ID: Lines: 3 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: You need to follow the directions about creating the afs user before klog will work. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dsl254-116-177-nyc1.dsl-isp.net by penn.central.org id aa11245; 30 Dec 2000 21:08 EST Received: from o-o.yi.org (IDENT:def@localhost.localdomain [127.0.0.1]) by oo.yi.org (8.9.3/8.8.7) with ESMTP id TAA01616 for ; Sat, 30 Dec 2000 19:32:18 -0500 Message-Id: <200012310032.TAA01616@oo.yi.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: openafs-info@openafs.org From: aeneous@speakeasy.org Reply-To: aeneous@speakeasy.org Subject: Re: [OpenAFS] Problem with klog In-Reply-To: Your message of "Fri, 29 Dec 2000 14:50:16 +0100." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 30 Dec 2000 19:32:18 -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: It looks like your server's keyfile has only keyid 0, while the kdc has keyid 2. Those need to be in sync. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from SWEET-TRANSVESTITE.MIT.EDU by penn.central.org id aa11566; 31 Dec 2000 0:05 EST Received: (from hartmans@localhost) by sweet-transvestite.mit.edu (8.9.3) id AAA29749; Sun, 31 Dec 2000 00:05:38 -0500 (EST) Date: Sun, 31 Dec 2000 00:05:38 -0500 (EST) Message-Id: <200012310505.AAA29749@sweet-transvestite.mit.edu> To: openafs-info@openafs.org From: Sam Hartman Subject: [OpenAFS] Boxed Penguin Prototype demonstrates easy AFS cell setup Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Boxed Penguin (http://www.boxedpenguin.com) is working to develop infrastructure software. We'd like to draw your attention to our initial prototype, based on Debian GNU/Linux, available at our website. The prototype demonstrates how a site could combine various components of Debian together to set up an infrastructure. AFS is used as a secure, distributed file system for the site's data. Kerberos acts as a central authentication service. SASL and PAM are used to provide security to applications like LDAP and IMAP. LDAP is used to provide a means for distributing account information. While designing the prototype, we worked to make the AFS cell setup painless and easy. As questions on this list indicate, setting up an AFS cell for the first time can be fairly confusing. While documentation is provided, the documentation makes the unfortunate assumption that the cell will be using kaserver. We do not want to encourage new sites to deploy Kerberos version 4. Boxed Penguin provides a better solution to AFS cell creation. In addition to integrating the cell creation with setting up a Kerberos version 5 realm, Boxed Penguin scripts the install so that the user does not have to manually create the initial volumes. Hopefully, our work will help focus effort on making AFS cell creation easier in all environments as well as focusing on Kerberos5 integration. Thanks, --Sam From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.tpi.pl by penn.central.org id aa20127; 2 Jan 2001 1:49 EST Received: from mail.tpi.pl (mail.tpi.pl [213.25.79.65]) by tpi.pl (Postfix) with ESMTP id 8CE301A6 for ; Tue, 2 Jan 2001 07:49:13 +0100 (CET) Date: Tue, 2 Jan 2001 07:49:13 +0100 (CET) From: Anna Szczechura To: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem with klog In-Reply-To: <200012310032.TAA01616@oo.yi.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > It looks like your server's keyfile has only keyid 0, > while the kdc has keyid 2. Those need to be in sync. Ehhh I think thats not a problem: [root@pluskwa anias]# kas examine afs -noauth User data for afs key (1) cksum is 2351075426, last cpw: Tue Jan 2 07:37:02 2001 password will never expire. An unlimited number of unsuccessful authentications is permitted. entry never expires. Max ticket lifetime 100.00 hours. last mod on Fri Dec 29 15:41:46 2000 by permit password reuse [root@pluskwa anias]# bos listkeys -server pluskwa -noauth key 1 has cksum 2351075426 Keys last changed on Tue Jan 2 07:40:26 2001. All done. Thay're now the same(I've just changed them) but after restarting .. I got the same message..(user admin which I use for testing purposes in authorization process is also available and doesn't have other uid in my /etc/passwd file - i don't have system admin user at all). regards, Anna Szczechura From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail.tpi.pl by penn.central.org id aa20842; 2 Jan 2001 4:09 EST Received: from mail.tpi.pl (mail.tpi.pl [213.25.79.65]) by tpi.pl (Postfix) with ESMTP id D47F3873 for ; Tue, 2 Jan 2001 10:09:41 +0100 (CET) Date: Tue, 2 Jan 2001 10:09:41 +0100 (CET) From: Anna Szczechura To: openafs-info@openafs.org Subject: Re: [OpenAFS] Problem with klog In-Reply-To: <200012310032.TAA01616@oo.yi.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Ok thanks for help.. I've solved my problem.. I was just too creative when choosing my cell name ;) The smallest problem the hardest to find.. regards anka From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ironside.nohup.se by penn.central.org id aa21999; 2 Jan 2001 8:03 EST Received: from shaft.int.nohup.se (IDENT:root@shaft.int.nohup.se [192.168.2.23]) by columbo.int.nohup.se (8.9.3/8.9.3) with ESMTP id OAA01425 for ; Tue, 2 Jan 2001 14:02:57 +0100 From: Kalle Svensson Received: (from kalle@localhost) by shaft.int.nohup.se (8.9.3/8.9.3) id OAA02763 for openafs-info@openafs.org; Tue, 2 Jan 2001 14:02:55 +0100 Date: Tue, 2 Jan 2001 14:02:55 +0100 To: openafs-info@openafs.org Message-ID: <20010102140255.A2746@shaft.int.nohup.se> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="0F1p//8PRICkK4MW" Content-Disposition: inline User-Agent: Mutt/1.2.5i Organization: Nohup AB Subject: [OpenAFS] backup addvolentry problem Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --0F1p//8PRICkK4MW Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi! We have a strange problem with our OpenAFS backup, backup addvolentry fails mysteriously with the following error messages: $ backup addvolentry blah blah ... backup: Unknown code bpi 132 (134914948) ; Volume entry not added $ backup backup> addvolentry blah blah ... backup: Unknown code bpe 196 (134913988) ; Volume entry not added Any ideas? I'm sorry about the short problem description, I'd appreciate if you told me what info you consider relevant. I've checked the logs, nothing that seems interesting. OS is GNU/Linux (RedHat 6.2) on i386. Regards, Kalle --=20 Kalle Svensson, Tech, Nohup AB | You can tune a filesystem, but you kalle@nohup.se, +46 8 4587812 | can't tune a fish. -- man tunefs(8) PGP fingerprint: 4A28 371A CCCA ED32 85E2 2335 D805 F251 8C79 A2F1 --0F1p//8PRICkK4MW Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.1 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6UdF/2AXyUYx5ovERArU5AKDSI7S+rCo+G/rB3m6y0SA/hA+HZgCfSglM u+EAJOoEb5X8WjFDLi6JqeU= =IsDO -----END PGP SIGNATURE----- --0F1p//8PRICkK4MW-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa00812; 3 Jan 2001 11:49 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id LAA15963 for ; Wed, 3 Jan 2001 11:49:19 -0500 (EST) Date: Wed, 03 Jan 2001 11:50:53 -0500 From: Derrick J Brashear To: openafs-info@openafs.org Message-ID: <681640000.978540653@skittlebrau.trafford.dementia.org> In-Reply-To: X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Subject: [OpenAFS] Re: Contrib Area Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Wednesday, January 03, 2001 04:43:59 PM +0000 Andrei Maslennikov wrote: > > Hi Mrs./Mr. Webmaster, > > I feel like I miss a contrib area on the www.openafs.org. Since the very > first days of the openAFS source availability, we are deploying the > extended AFS RPMs of our build and are quite eager to share them with > the other people. > > Please tell me if you could consider creating one such area on your site. > If this is not possible: could you at least foresee a "Contrib" section in > your "Links" area? (So that you could point to us and/or anyone else who > would like to share their AFS-related stuff). In the near future a "Contrib" area will be present on www.central.org; "Contrib" items other than those which can be integrated into AFS proper will be provided there. The AFS cell is already operational, but not completely configured. More details will be available soon. Derrick J. Brashear From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from kiwi.caspur.it by penn.central.org id aa00877; 3 Jan 2001 11:57 EST Received: from localhost (andrei@localhost) by kiwi.caspur.it (8.11.0/8.11.0) with SMTP id f03GvdX21607; Wed, 3 Jan 2001 17:57:39 +0100 X-Authentication-Warning: kiwi.caspur.it: andrei owned process doing -bs Date: Wed, 3 Jan 2001 16:57:39 +0000 ( ) From: Andrei Maslennikov To: webmaster@openafs.org cc: openafs-info@openafs.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] Contrib Area Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi Mrs./Mr. Webmaster, I feel like I miss a contrib area on the www.openafs.org. Since the very first days of the openAFS source availability, we are deploying the extended AFS RPMs of our build and are quite eager to share them with the other people. Please tell me if you could consider creating one such area on your site. If this is not possible: could you at least foresee a "Contrib" section in your "Links" area? (So that you could point to us and/or anyone else who would like to share their AFS-related stuff). Thanks - Andrei. ____________________________________________________________________________ Andrei Maslennikov phone : +39 06 4463354 CASPUR Inter-Univ. Computing Consortium and INFN cell : +39 335 6214776 c/o Universita' "La Sapienza", 00185 Rome, Italy fax : +39 06 4957083 ___________________________________________________ mail : andrei@caspur.it From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from cssvr17.cs.ust.hk by penn.central.org id aa03751; 3 Jan 2001 20:39 EST Received: from guitar.cs.ust.hk (IDENT:root@guitar.cs.ust.hk [143.89.43.97]) by cssvr17.cs.ust.hk (8.11.0/8.10.1) with ESMTP id f041dgA17439 for ; Thu, 4 Jan 2001 09:39:42 +0800 (HKT) Full-Name: Dominic Lai Received: (from dominic@localhost) by guitar.cs.ust.hk (8.11.1/8.9.2) id f041dgM06387 for openafs-info@openafs.org; Thu, 4 Jan 2001 09:39:42 +0800 Date: Thu, 4 Jan 2001 09:39:41 +0800 From: Dominic Lai To: openafs-info@openafs.org Message-ID: <20010104093941.A6256@guitar.cs.ust.hk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i X-Webpage: http://www.cs.ust.hk/~dominic X-Affiliation: Department of Computer Science, HKUST Subject: [OpenAFS] openafs problem on Solaris 8 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: In the past, I used to have openafs (1.0) working on Solaris 8, after applying Chas Williams' patch. Recently after applying the kernel patch #108528-04, I find that openafs (even I tried 1.0.1) no longer works. Basically, it could still mount the Cell volume. But when accessing a file in the volume, it gives a kernel crash: panic[cpu0]/thread=f6594ce0: BAD TRAP: type=9 (Data fault) rp=fbfe5644 addr=c mmu_fsr=326 rw=1 occurred in module "unix" due to a NULL pointer dereference less: Data fault kernel read fault at addr=0xc, pme=0x0 MMU sfsr=326: Invalid Address on supv data fetch at level 3 pte addr = 0xf59d3a00, level = 3 pid=285, pc=0xf005e6c8, sp=0xfbfe5690, psr=0x404010c2, context=29 g1-g7: 1, ff, f114e, 1, f6598e00, f0041000, f6594ce0 fbfe54c0 unix:die+b8 (9, f024c9f0, f024c84c, 326, 1, 326) %l0-7: fbfe5644 0000000c fbfe5644 00000004 f5ae6a38 00000001 00000000 00000004fbfe5540 unix:trap+790 (0, 1, f0000000, f659a080, f6250240, 0) %l0-7: 00000000 00000000 fbfe5644 00000009 00000326 00000001 000001d7 0000039bfbfe55e8 unix:fault+84 (f6594ce0, 0, ffb2ca67, 1, f6594ce4, 0) %l0-7: 404010c2 f005e6c8 f005e6cc 00000040 00000009 00000004 f024fc98 fbfe55e8fbfe5690 afs:afs_GetOnePage+1fc (0, 0, f0248b24, f024e428, f114e000, 1000) %l0-7: 409010c3 f62b6628 f6337a44 00000020 00000004 00000000 00000001 00000404fbfe56f0 unix:page_create_va+308 (3, 0, 1, 0, f02508d8, 4eac) %l0-7: 000013ab 0000000b f114e000 f0250a10 00000000 f644eacc f02582ac 00000000fbfe5778 afs:afs_GetOnePage+8a0 (f644eacc, 0, f6512b60, f644eacc, 0, 0) %l0-7: f02582ac f644eacc 00000000 00000000 f114e000 00001000 00000003 ef7ed762fbfe5840 afs:afs_getpage+100 (f5a14fa0, f114e000, 1, 1000, fbfe594c, 0) %l0-7: f644eacc 00001000 00000000 00000000 f5a13e48 fbfe594c fbfe5940 00002000fbfe58d0 genunix:segmap_fault+100 (0, 0, 0, 1000, 2, 1) %l0-7: f6337408 00000000 f59303c0 f644eacc f114e000 f5a2e000 f6512b60 00000000fbfe5950 afs:afs_nfsrdwr+1858 (f644eacc, 0, 0, 40, f114e000, 0) %l0-7: f5a14fa0 f5a2e000 f114e000 00001000 00000001 00000002 00000000 0000339bfbfe5a20 afs:afs_vmread+80 (f644eacc, fbfe5af0, 0, f5a13e48, 6, f663b000) %l0-7: fbfe5af0 00000000 00000000 f5a13e48 0003f678 f6594ce0 f024fc98 fbfe5b80fbfe5a88 genunix:read+270 (6, 1, 40, 0, f61f67b8, f644eacc) %l0-7: f6337268 00000040 00000040 00000000 00000000 f6594ce0 f024fc98 fbfe5b80 It is on a SS20 platform, while the package was compiled by Workshop 5.0. Have anyone experienced this problem? Pointers are much appreciated. Thanks. Dominic -- Dominic Lai, Assistant Computer Officer, CS System Group Computer Science Department, The Hong Kong University of Science & Technology Home Page : http://www.cs.ust.hk/~dominic PGP Public Key: ftp://ftp.cs.ust.hk/pub/cssystem/dominic/pgp.pubkey.asc Mojo: Rock N' Roll Forever Any unsolicited mails will be reported to to have the corresponding mail domains added to the blackhole list. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mailgateway1.uni-freiburg.de by penn.central.org id aa06501; 4 Jan 2001 6:00 EST Received: from mawa.ruf.uni-freiburg.de [132.230.1.30] by mailgateway1.uni-freiburg.de with esmtp (Exim 1.82 #3) id 14E883-0007Q0-00; Thu, 4 Jan 2001 12:00:19 +0100 Date: Thu, 4 Jan 2001 12:00:20 +0100 (MET) From: Martin Walter X-Sender: To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] data encryption Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: hi, is there any full data encryption mode in afs? if not, is there any plan to extend afs in this direction? regards, martin. -- Martin Walter Office: OG 115 University Computing Center Mail: mawa@uni-freiburg.de Hermann-Herder-Str.10 Phone: +49 761 203 4651 D-79104 Freiburg i.Br. Fax: +49 761 203 4643 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dsl254-116-177-nyc1.dsl-isp.net by penn.central.org id aa07198; 4 Jan 2001 8:17 EST Received: from o-o.yi.org (IDENT:def@localhost.localdomain [127.0.0.1]) by oo.yi.org (8.9.3/8.8.7) with ESMTP id IAA07957 for ; Thu, 4 Jan 2001 08:17:16 -0500 Message-Id: <200101041317.IAA07957@oo.yi.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: openafs-info@openafs.org From: aeneous@speakeasy.org Reply-To: aeneous@speakeasy.org Subject: Re: [OpenAFS] data encryption In-Reply-To: Your message of "Thu, 04 Jan 2001 12:00:20 +0100." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 04 Jan 2001 08:17:16 -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > is there any full data encryption mode in afs? > if not, is there any plan to extend afs in this direction? There's half a one, but it uses fcrypt, which isn't really a great algorithm. It would be useful, even so, if it were finished. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from locutus.cmf.nrl.navy.mil by penn.central.org id aa08067; 4 Jan 2001 10:33 EST Received: from locutus.cmf.nrl.navy.mil (chas@localhost) by locutus.cmf.nrl.navy.mil (8.9.3/8.9.3) with ESMTP id KAA03586; Thu, 4 Jan 2001 10:33:26 -0500 Message-Id: <200101041533.KAA03586@locutus.cmf.nrl.navy.mil> To: Dominic Lai cc: openafs-info@openafs.org Subject: Re: [OpenAFS] openafs problem on Solaris 8 In-reply-to: Your message of "Thu, 04 Jan 2001 09:39:41 +0800." <20010104093941.A6256@guitar.cs.ust.hk> X-url: http://www.nrl.navy.mil/CCS/people/chas/index.html X-mailer: nmh 1.0 Date: Thu, 04 Jan 2001 10:33:25 -0500 From: chas@cmf.nrl.navy.mil Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: >In the past, I used to have openafs (1.0) working on Solaris 8, after >applying Chas Williams' patch. Recently after applying the kernel >patch #108528-04, I find that openafs (even I tried 1.0.1) no longer >works. Basically, it could still mount the Cell volume. But when accessing >a file in the volume, it gives a kernel crash: i will apply this patch and give it a try. if you could get a stack trace using adb that would be pretty handy From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mail-out2.apple.com by penn.central.org id aa08322; 4 Jan 2001 10:58 EST Received: from mailgate1.apple.com (A17-128-100-225.apple.com [17.128.100.225]) by mail-out2.apple.com (8.9.3/8.9.3) with ESMTP id HAA17616 for ; Thu, 4 Jan 2001 07:58:03 -0800 (PST) Received: from scv2.apple.com (scv2.apple.com) by mailgate1.apple.com (Content Technologies SMTPRS 4.1.2) with ESMTP id ; Thu, 4 Jan 2001 07:58:03 -0800 Received: from apple.com (vpn-gh-1038.apple.com [17.254.140.13]) by scv2.apple.com (8.9.3/8.9.3) with ESMTP id HAA28941; Thu, 4 Jan 2001 07:58:02 -0800 (PST) Message-ID: <3A549D8A.2E1A2A7E@apple.com> Date: Thu, 04 Jan 2001 07:58:35 -0800 From: Thomas Vincent X-Mailer: Mozilla 4.7C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; I; PPC) X-Accept-Language: en MIME-Version: 1.0 To: aeneous@speakeasy.org CC: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption References: <200101041317.IAA07957@oo.yi.org> Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; x-mac-creator="4D4F5353" Content-Transfer-Encoding: 7bit Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: My only concern with encrypting all the day would be CPU. What kind of performance hit would the server take? I believe the strategy that Microsoft is pushing with AD enabled apps is ipsec for data encryption, and keberos for authentication. Sounds good on paper. Has a road map been put out for OpenAFS development? Something like teir 1 priorities, teir 2, etc. I have looked at the changelogs, but have aeneous@speakeasy.org wrote: > > is there any full data encryption mode in afs? > > if not, is there any plan to extend afs in this direction? > > There's half a one, but it uses fcrypt, which isn't really a great > algorithm. It would be useful, even so, if it were finished. > > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa08341; 4 Jan 2001 10:59 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.1.108]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id JAA31747; Thu, 4 Jan 2001 09:59:12 -0600 Received: by umr-mail01.cc.umr.edu with Internet Mail Service (5.5.2650.21) id ; Thu, 4 Jan 2001 09:59:12 -0600 Message-ID: From: "Neulinger, Nathan R." To: "'openafs-info@openafs.org'" Date: Thu, 4 Jan 2001 09:59:05 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Subject: [OpenAFS] will transarc put openafs cells in their CellServDB? Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I know they don't want their CellServDB to be used by default with OpenAFS, I'm wondering if they will include OpenAFS cells in it though if people provide the info? -- Nathan ------------------------------------------------------------ Nathan Neulinger EMail: nneul@umr.edu University of Missouri - Rolla Phone: (573) 341-4841 Computing Services Fax: (573) 341-4216 From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa08497; 4 Jan 2001 11:22 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id LAA25734; Thu, 4 Jan 2001 11:22:01 -0500 To: Thomas Vincent Cc: aeneous@speakeasy.org, openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption References: <200101041317.IAA07957@oo.yi.org> <3A549D8A.2E1A2A7E@apple.com> From: Derek Atkins Date: 04 Jan 2001 11:22:01 -0500 In-Reply-To: Thomas Vincent's message of "Thu, 04 Jan 2001 07:58:35 -0800" Message-ID: Lines: 21 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Thomas Vincent writes: > My only concern with encrypting all the day would be CPU. What kind of > performance hit would the server take? > I believe the strategy that Microsoft is pushing with AD enabled apps is > ipsec for data encryption, and keberos for authentication. Sounds good > on paper. This would require at least as much processing power as just encrypting AFS itself ;) The other problem with "just use IPSec" is that application protocols don't benefit from the security infrastructure. I.e. an application cannot query the IPSec SA to find out who sent a particular packet. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa08503; 4 Jan 2001 11:23 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id LAA25737; Thu, 4 Jan 2001 11:23:18 -0500 To: "Neulinger, Nathan R." Cc: "'openafs-info@openafs.org'" Subject: Re: [OpenAFS] will transarc put openafs cells in their CellServDB? References: From: Derek Atkins Date: 04 Jan 2001 11:23:18 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 4 Jan 2001 09:59:05 -0600" Message-ID: Lines: 30 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I was under the impression that OpenAFS was planning to host a list of OpenAFS cells, so users could grab the list from both places. Hopefully the "AFSDB RR" support will reduce the need for CellServDB files. -derek "Neulinger, Nathan R." writes: > I know they don't want their CellServDB to be used by default with OpenAFS, > I'm wondering if they will include OpenAFS cells in it though if people > provide the info? > > -- Nathan > > ------------------------------------------------------------ > Nathan Neulinger EMail: nneul@umr.edu > University of Missouri - Rolla Phone: (573) 341-4841 > Computing Services Fax: (573) 341-4216 > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mrelay.cc.umr.edu by penn.central.org id aa08529; 4 Jan 2001 11:26 EST Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu [131.151.1.108]) via ESMTP by mrelay.cc.umr.edu (8.9.3/R.4.20) id KAA07931; Thu, 4 Jan 2001 10:26:23 -0600 Received: by umr-mail01.cc.umr.edu with Internet Mail Service (5.5.2650.21) id ; Thu, 4 Jan 2001 10:26:23 -0600 Message-ID: From: "Neulinger, Nathan R." To: "'Derek Atkins'" Cc: "'openafs-info@openafs.org'" Subject: RE: [OpenAFS] will transarc put openafs cells in their CellServDB ? Date: Thu, 4 Jan 2001 10:26:20 -0600 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="ISO-8859-1" Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I was thinking more from the perspective of sites not using OpenAFS - they'd likely never think to retrieve the OpenAFS CellServDB. Particularly if those sites don't do their own CellServDB maintenance and simply sync with transarcs export file. -- Nathan > -----Original Message----- > From: Derek Atkins [mailto:warlord@MIT.EDU] > Sent: Thursday, January 04, 2001 10:23 AM > To: Neulinger, Nathan R. > Cc: 'openafs-info@openafs.org' > Subject: Re: [OpenAFS] will transarc put openafs cells in their > CellServDB? > > > I was under the impression that OpenAFS was planning to host a list > of OpenAFS cells, so users could grab the list from both places. > > Hopefully the "AFSDB RR" support will reduce the need for CellServDB > files. > > -derek > > "Neulinger, Nathan R." writes: > > > I know they don't want their CellServDB to be used by > default with OpenAFS, > > I'm wondering if they will include OpenAFS cells in it > though if people > > provide the info? > > > > -- Nathan > > > > ------------------------------------------------------------ > > Nathan Neulinger EMail: nneul@umr.edu > > University of Missouri - Rolla Phone: (573) 341-4841 > > Computing Services Fax: (573) 341-4216 > > _______________________________________________ > > OpenAFS-info mailing list > > OpenAFS-info@openafs.org > > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info > > -- > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > Member, MIT Student Information Processing Board (SIPB) > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH > warlord@MIT.EDU PGP key available > From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa08594; 4 Jan 2001 11:30 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id LAA25749; Thu, 4 Jan 2001 11:30:43 -0500 To: "Neulinger, Nathan R." Cc: "'openafs-info@openafs.org'" Subject: Re: [OpenAFS] will transarc put openafs cells in their CellServDB ? References: From: Derek Atkins Date: 04 Jan 2001 11:30:39 -0500 In-Reply-To: "Neulinger, Nathan R."'s message of "Thu, 4 Jan 2001 10:26:20 -0600" Message-ID: Lines: 61 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: I suppose we'll have to wait for an answer from Transarc/IBM.... -derek "Neulinger, Nathan R." writes: > I was thinking more from the perspective of sites not using OpenAFS - they'd > likely never think to retrieve the OpenAFS CellServDB. Particularly if those > sites don't do their own CellServDB maintenance and simply sync with > transarcs export file. > > -- Nathan > > > -----Original Message----- > > From: Derek Atkins [mailto:warlord@MIT.EDU] > > Sent: Thursday, January 04, 2001 10:23 AM > > To: Neulinger, Nathan R. > > Cc: 'openafs-info@openafs.org' > > Subject: Re: [OpenAFS] will transarc put openafs cells in their > > CellServDB? > > > > > > I was under the impression that OpenAFS was planning to host a list > > of OpenAFS cells, so users could grab the list from both places. > > > > Hopefully the "AFSDB RR" support will reduce the need for CellServDB > > files. > > > > -derek > > > > "Neulinger, Nathan R." writes: > > > > > I know they don't want their CellServDB to be used by > > default with OpenAFS, > > > I'm wondering if they will include OpenAFS cells in it > > though if people > > > provide the info? > > > > > > -- Nathan > > > > > > ------------------------------------------------------------ > > > Nathan Neulinger EMail: nneul@umr.edu > > > University of Missouri - Rolla Phone: (573) 341-4841 > > > Computing Services Fax: (573) 341-4216 > > > _______________________________________________ > > > OpenAFS-info mailing list > > > OpenAFS-info@openafs.org > > > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info > > > > -- > > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory > > Member, MIT Student Information Processing Board (SIPB) > > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH > > warlord@MIT.EDU PGP key available > > -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from citi.umich.edu by penn.central.org id aa08813; 4 Jan 2001 12:12 EST Received: from citi.umich.edu (ooty.citi.umich.edu [141.211.169.121]) by citi.umich.edu (Postfix) with ESMTP id E45AE207C5 for ; Thu, 4 Jan 2001 12:12:57 -0500 (EST) Subject: Re: [OpenAFS] data encryption To: openafs-info@openafs.org From: Jim Rees In-Reply-To: aeneous@speakeasy.org, Thu, 04 Jan 2001 08:17:16 EST Date: Thu, 04 Jan 2001 12:12:57 -0500 Message-Id: <20010104171257.E45AE207C5@citi.umich.edu> Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: We have been using full data encryption in our Transarc afs cell for about a year now. Before that, cpu load on the server was a problem, but upgrading from a Sparc IPC to a Sparcstation (something) solved that. The Transarc code requires a couple of small patches to make this work. The code is all there but "fs" needs to be able to turn encryption on and off, and the encryption flag needs to get carried down to the proper place. I haven't looked at that part of the OpenAFS code but it shouldn't be hard. The fcrypt altgorithm is better than nothing but less than ideal. NFS v4 uses gss for authentication and encryption. I suspect that's overkill for afs, but it has some nice features. For example, the OpenBSD client will eventually use the OpenBSD kernel crypt library, and that means a choice of algorithms, as well as hardware acceleration where it's available. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa08844; 4 Jan 2001 12:28 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id MAA27055 for ; Thu, 4 Jan 2001 12:28:42 -0500 (EST) Date: Thu, 04 Jan 2001 12:30:27 -0500 From: Derrick J Brashear To: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption Message-ID: <1760300000.978629427@skittlebrau.trafford.dementia.org> In-Reply-To: <200101041317.IAA07957@oo.yi.org> X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Thursday, January 04, 2001 08:17:16 AM -0500 aeneous@speakeasy.org wrote: >> is there any full data encryption mode in afs? >> if not, is there any plan to extend afs in this direction? > > There's half a one, but it uses fcrypt, which isn't really a great > algorithm. It would be useful, even so, if it were finished. I have patches to use it, ready to commit. Derrick J Brashear From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa08971; 4 Jan 2001 12:39 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id MAA25781; Thu, 4 Jan 2001 12:39:30 -0500 To: Jim Rees Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption References: <20010104171257.E45AE207C5@citi.umich.edu> From: Derek Atkins Date: 04 Jan 2001 12:39:30 -0500 In-Reply-To: Jim Rees's message of "Thu, 04 Jan 2001 12:12:57 -0500" Message-ID: Lines: 36 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Jim, Did it require changes to the server or just the client? I'd certainly be interested in this patch (as would a bunch of others, I'm sure) if you could find it. -derek Jim Rees writes: > We have been using full data encryption in our Transarc afs cell for about a > year now. Before that, cpu load on the server was a problem, but upgrading > from a Sparc IPC to a Sparcstation (something) solved that. > > The Transarc code requires a couple of small patches to make this work. The > code is all there but "fs" needs to be able to turn encryption on and off, > and the encryption flag needs to get carried down to the proper place. I > haven't looked at that part of the OpenAFS code but it shouldn't be hard. > > The fcrypt altgorithm is better than nothing but less than ideal. > > NFS v4 uses gss for authentication and encryption. I suspect that's > overkill for afs, but it has some nice features. For example, the OpenBSD > client will eventually use the OpenBSD kernel crypt library, and that means > a choice of algorithms, as well as hardware acceleration where it's > available. > _______________________________________________ > OpenAFS-info mailing list > OpenAFS-info@openafs.org > https://lists.openafs.org/mailman/listinfo.cgi/openafs-info -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa09021; 4 Jan 2001 12:42 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id MAA25787; Thu, 4 Jan 2001 12:42:30 -0500 To: Derrick J Brashear Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption References: <1760300000.978629427@skittlebrau.trafford.dementia.org> From: Derek Atkins Date: 04 Jan 2001 12:42:30 -0500 In-Reply-To: Derrick J Brashear's message of "Thu, 04 Jan 2001 12:30:27 -0500" Message-ID: Lines: 15 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Derrick J Brashear writes: > I have patches to use it, ready to commit. WONDERFUL! :) > Derrick J Brashear -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ALYCIA.DEMENTIA.ORG by penn.central.org id aa09199; 4 Jan 2001 12:52 EST Received: from skittlebrau.trafford.dementia.org (VPN-100-227.CC.CMU.EDU [128.2.100.227]) by alycia.dementia.org (8.9.1a/8.9.0) with ESMTP id MAA27260 for ; Thu, 4 Jan 2001 12:52:03 -0500 (EST) Date: Thu, 04 Jan 2001 12:53:47 -0500 From: Derrick J Brashear To: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption Message-ID: <1785700000.978630827@skittlebrau.trafford.dementia.org> In-Reply-To: X-Mailer: Mulberry/2.0.6b2 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --On Thursday, January 04, 2001 12:39:30 PM -0500 Derek Atkins wrote: > Jim, > > Did it require changes to the server or just the client? I'd > certainly be interested in this patch (as would a bunch of others, I'm > sure) if you could find it. The relevant patch makes 2 changes in the cache manager, one (simple) to make encryption possible and one to add a pioctl to enable and disable it, and one change to fs, to allow that command to enable and disable it. Derrick J Brashear From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from MINBAR.FAC.CS.CMU.EDU by penn.central.org id aa10022; 4 Jan 2001 14:22 EST Received: from minbar.fac.cs.cmu.edu by minbar.fac.cs.cmu.edu id aa05235; 4 Jan 2001 14:22 EST Date: Thu, 4 Jan 2001 14:22:34 -0500 (EST) From: Jeffrey Hutzelman X-Sender: jhutz@minbar.fac.cs.cmu.edu To: Jim Rees cc: openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption In-Reply-To: <20010104171257.E45AE207C5@citi.umich.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On Thu, 4 Jan 2001, Jim Rees wrote: > NFS v4 uses gss for authentication and encryption. I suspect that's > overkill for afs, but it has some nice features. For example, the OpenBSD > client will eventually use the OpenBSD kernel crypt library, and that means > a choice of algorithms, as well as hardware acceleration where it's > available. Actually, I don't think it's overkill at all. In fact, my current plan for rxkad v3 is to use GSSAPI, for some of the same reasons. However, we'll only use GSSAPI for authentication and key exchange, and handle encryption of AFS traffic independently. Anything else would require significant chunks of GSSAPI code, including krb5 and whatever other mechanisms you want, to be in the kernel. -- Jeffrey T. Hutzelman (N3NHS) Sr. Research Systems Programmer School of Computer Science - Research Computing Facility Carnegie Mellon University - Pittsburgh, PA From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa10179; 4 Jan 2001 14:40 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id OAA25908; Thu, 4 Jan 2001 14:40:03 -0500 To: sommerfeld@orchard.arlington.ma.us Cc: Thomas Vincent , aeneous@speakeasy.org, openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption References: <20010104192100.3F11A2A3E@orchard.arlington.ma.us> From: Derek Atkins Date: 04 Jan 2001 14:40:03 -0500 In-Reply-To: Bill Sommerfeld's message of "Thu, 04 Jan 2001 14:20:54 -0500" Message-ID: Lines: 41 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Bill, Unfortunately there is no standard API to obtain IPSec credentials. Sure, different implementation _may_ decide to supply an API, but then you'd have to figure out what IPSec implementation you are using and code specifically to that. However the fact remains that most implementations DO NOT supply application-level access to SA creds. Sure, we could standardize an API. But first we'd have to come up with one, and then we'd have to go through the process of standardizing it. Oh, yea, then we'd have to wait for it to get implemented. So I suppose in 2005 we might have this available... But we're talking about the here and now, making changes to subsystems that we control. I'm not going to wait until Micro$oft decides they want to support a (not-yet-created) standard API to access IPSec SA credentials. I'd rather get AFS encrypting now, thank you very much. -derek Bill Sommerfeld writes: > > This would require at least as much processing power as just > > encrypting AFS itself ;) The other problem with "just use IPSec" is > > that application protocols don't benefit from the security > > infrastructure. I.e. an application cannot query the IPSec SA to find > > out who sent a particular packet. > > This is an implementation limitation of current ipsec implementations. > Do not assume this will always be the case. > > it's worth noting that ipsec is much more amenable to hardware > accelleration than many other security protocols. > > - Bill -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from kiwi.caspur.it by penn.central.org id aa00758; 3 Jan 2001 11:43 EST Received: from localhost (andrei@localhost) by kiwi.caspur.it (8.11.0/8.11.0) with SMTP id f03Ghxr21468; Wed, 3 Jan 2001 17:43:59 +0100 X-Authentication-Warning: kiwi.caspur.it: andrei owned process doing -bs Date: Wed, 3 Jan 2001 16:43:59 +0000 ( ) From: Andrei Maslennikov Reply-To: Andrei Maslennikov To: webmaster@openafs.org cc: openafs-info@openafs.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: [OpenAFS] Contrib Area Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Hi Mrs./Mr. Webmaster, I feel like I miss a contrib area on the www.openafs.org. Since the very first days of the openAFS source availability, we are deploying the extended AFS RPMs of our build and are quite eager to share them with the other people. Please tell me if you could consider creating one such area on your site. If this is not possible: could you at least foresee a "Contrib" section in your "Links" area? (So that you could point to us and/or anyone else who would like to share their AFS-related stuff). Thanks - Andrei. ____________________________________________________________________________ Andrei Maslennikov phone : +39 06 4463354 CASPUR Inter-Univ. Computing Consortium and INFN cell : +39 335 6214776 c/o Universita' "La Sapienza", 00185 Rome, Italy fax : +39 06 4957083 ___________________________________________________ mail : andrei@caspur.it From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from orchard.hamachi.org by penn.central.org id aa10000; 4 Jan 2001 14:21 EST Received: by orchard.arlington.ma.us (Postfix, from userid 587) id 3F11A2A3E; Thu, 4 Jan 2001 14:21:00 -0500 (EST) Received: from orchard.arlington.ma.us (localhost [127.0.0.1]) by orchard.arlington.ma.us (Postfix) with ESMTP id 1D0091FD8; Thu, 4 Jan 2001 14:21:00 -0500 (EST) To: Derek Atkins Cc: Thomas Vincent , aeneous@speakeasy.org, openafs-info@openafs.org Subject: Re: [OpenAFS] data encryption In-Reply-To: Message from Derek Atkins of "04 Jan 2001 11:22:01 EST." Reply-To: sommerfeld@orchard.arlington.ma.us Date: Thu, 04 Jan 2001 14:20:54 -0500 From: Bill Sommerfeld Message-Id: <20010104192100.3F11A2A3E@orchard.arlington.ma.us> Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > This would require at least as much processing power as just > encrypting AFS itself ;) The other problem with "just use IPSec" is > that application protocols don't benefit from the security > infrastructure. I.e. an application cannot query the IPSec SA to find > out who sent a particular packet. This is an implementation limitation of current ipsec implementations. Do not assume this will always be the case. it's worth noting that ipsec is much more amenable to hardware accelleration than many other security protocols. - Bill From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE327D1.dip.t-dialin.net by penn.central.org id aa10944; 4 Jan 2001 16:29 EST Received: (qmail 9420 invoked from network); 4 Jan 2001 21:29:16 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 4 Jan 2001 21:29:16 -0000 Received: (qmail 18544 invoked by uid 1000); 4 Jan 2001 21:29:15 -0000 Date: Thu, 4 Jan 2001 22:29:15 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall Message-ID: <20010104222914.B606@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org References: <20001219043252.A2595@cube.sascha.silbe.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="KFztAG8eRSV9hGtP" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: ; from warlord@MIT.EDU on Tue, Dec 19, 2000 at 01:29:41PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --KFztAG8eRSV9hGtP Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Dec 19, 2000 at 01:29:41PM -0500, Derek Atkins [DA] wrote: Sorry for the long delay, I was away for christmas and new year's eve. DA> AFS does not currently support TCP operations. That could be a real problem for masqueraded machines, depending on how the protocol works. Unfortunately, I could not find any docs at all regarding t= he protocol itself. DA> AFS _does_ work behind a masquerading firewall provided that you set t= he=20 DA> UDP timeouts high enough to allow callbacks to occur. When do callbacks occur? Only within some time after a client request or anytime? On the same ports as the request or on a separate one? DA> I would recommend UDP timeouts in the range of 10-15 minutes. Thanks! I'll see if it helps. Strangely the OpenAFS client does not work at= =20 all currently. Perhaps I have to reboot the machine after stopping arla and before starting OpenAFS. DA> You should also note that AFS might "hiccup" when the IP address DA> changes, at least until new callbacks can be set with the new IP. That's currently a big problem here, both with arla and OpenAFS. Even=20 restarting the client does not help. :( CU/Lnx Sascha --KFztAG8eRSV9hGtP Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6VOsqa8hvv2E1w1sRAjs3AJ9kUwNjfNSIQZLXz2UkzjbsBCJG8QCfcSYg VYO9xgpvVgRi71uAK2aErSU= =Pc8G -----END PGP SIGNATURE----- --KFztAG8eRSV9hGtP-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa11166; 4 Jan 2001 17:21 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id RAA26071; Thu, 4 Jan 2001 17:20:46 -0500 To: Sascha Silbe Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> From: Derek Atkins Date: 04 Jan 2001 17:20:46 -0500 In-Reply-To: Sascha Silbe's message of "Thu, 4 Jan 2001 22:29:15 +0100" Message-ID: Lines: 58 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sascha Silbe writes: > Sorry for the long delay, I was away for christmas and new year's eve. Well, happy new year. :) > DA> AFS _does_ work behind a masquerading firewall provided that you set t= > he=20 > DA> UDP timeouts high enough to allow callbacks to occur. > When do callbacks occur? Only within some time after a client request or > anytime? On the same ports as the request or on a separate one? Basically the cache manager opens port 7001 on the client and uses that port to contact ports 7000-7008 on the AFS servers. All output from the client cache manager exits via UDP port 7001, and all responses to the cache manager (including callbacks) return via UDP port 7001. Callbacks can pretty much occur at any time. HOWEVER, the client cache manager 'pings' each server periodically to make sure it is still there. This ping happens, IIRC, every 10 minutes. So, if you set a 15-minute UDP timeout, that should keep everything alive. > DA> I would recommend UDP timeouts in the range of 10-15 minutes. > Thanks! I'll see if it helps. Strangely the OpenAFS client does not > work at all currently. Perhaps I have to reboot the machine after > stopping arla and before starting OpenAFS. Maybe -- I've not tried arla, and I've never tried changing over from one to the other. Rebooting is always a good idea in such strange situations ;) > DA> You should also note that AFS might "hiccup" when the IP address > DA> changes, at least until new callbacks can be set with the new IP. > That's currently a big problem here, both with arla and OpenAFS. Even=20 > restarting the client does not help. :( Perhaps you should find a better ISP -- one that gives you a more stable IP Address? Yea, if your IP Address changes, you probably need to: 1) stop AFS 2) remove the AFS module 3) restart AFS It definitely wont work without step #2. Just keep in mind that AFS was not designed with NAT in mind (NAT is an abomination). The fact that it works across NAT is just luck. AFS barely copes when your own IP address changes, but at least the AFS client KNOWS when this change happens. The client cannot know when the NAT-box changes IP Address. > CU/Lnx Sascha -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE327D1.dip.t-dialin.net by penn.central.org id aa11490; 4 Jan 2001 18:08 EST Received: (qmail 17938 invoked from network); 4 Jan 2001 23:08:33 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 4 Jan 2001 23:08:33 -0000 Received: (qmail 19223 invoked by uid 1000); 4 Jan 2001 23:08:33 -0000 Date: Fri, 5 Jan 2001 00:08:33 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall Message-ID: <20010105000833.F606@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="ULyIDA2m8JTe+TiX" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: ; from warlord@MIT.EDU on Thu, Jan 04, 2001 at 05:20:46PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --ULyIDA2m8JTe+TiX Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Jan 04, 2001 at 05:20:46PM -0500, Derek Atkins [DA] wrote: DA> Well, happy new year. :) Thanks, same to you and everyone else on this list. DA> All output from the client cache manager exits via UDP port 7001, and= =20 DA> all responses to the cache manager (including callbacks) return via UDP DA> port 7001. Fine. How about the source port of the callback packets? Is it the same as the destination port of the original request? DA> Callbacks can pretty much occur at any time. HOWEVER, the client DA> cache manager 'pings' each server periodically to make sure it is DA> still there. Wonderful. Does it use the same destination ports for both requests and pin= gs? DA> Rebooting is always a good idea in such strange situations ;) Though it should be rare on Linux machines. :) DA> Perhaps you should find a better ISP It's not a technical or stability issue, but a policy one: The connection is cut (and reopened by my firewall) after 24 hours. This is the only differen= ce between a so-called "Flatrate" and a real leased line, besides the prices (Flatrate: around 40$, leased line: around 150$). There is no ISP in Germany that provides a static IP for less than 150$ per month which is more than I am able to afford. DA> 1) stop AFS Did that. DA> 2) remove the AFS module /etc/init.d/afs seems to already do that for me. DA> 3) restart AFS Did that, but no effect. :( DA> Just keep in mind that AFS was not designed with NAT in mind=20 DA> (NAT is an abomination). I'm of the same opinion, but I have to live with it. IPv6 will take at least several months. CU/Lnx Sascha --ULyIDA2m8JTe+TiX Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6VQJxa8hvv2E1w1sRAgFXAJ9UKcLGnItUIFSvzxoZ/oj22lCSIgCfet7X GdzItmeBL1uDsGkSaP8ZDVk= =w03A -----END PGP SIGNATURE----- --ULyIDA2m8JTe+TiX-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from mercury.ccmr.cornell.edu by penn.central.org id aa11776; 4 Jan 2001 19:05 EST Received: from ruby.ccmr.cornell.edu (IDENT:0@ruby.ccmr.cornell.edu [128.84.231.115]) by mercury.ccmr.cornell.edu (8.9.3/8.9.3) with ESMTP id TAA30416; Thu, 4 Jan 2001 19:05:39 -0500 Received: from localhost (mitch@localhost) by ruby.ccmr.cornell.edu (8.9.3/8.9.3) with ESMTP id TAA22704; Thu, 4 Jan 2001 19:05:38 -0500 X-Authentication-Warning: ruby.ccmr.cornell.edu: mitch owned process doing -bs Date: Thu, 4 Jan 2001 19:05:38 -0500 (EST) From: Mitch Collinsworth To: Derek Atkins cc: Sascha Silbe , openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: On 4 Jan 2001, Derek Atkins wrote: > Just keep in mind that AFS > was not designed with NAT in mind (NAT is an abomination). The fact > that it works across NAT is just luck. AFS barely copes when your own > IP address changes, but at least the AFS client KNOWS when this change > happens. The client cannot know when the NAT-box changes IP Address. I agree that NAT is not the greatest thing to come down the road, but it's already here. The battle to prevent it was lost a long time ago. It's going to be with us for a while and there are lots of people for whom it's going to be a fact of life. Some of them are even going to turn out to be important to your or my or someone else's livelihood here and it would be better for most of us if we figured out a way to make it work. I'd like to see us take this sort of thing as a challenge to improve the software rather than write off the whole idea. Ok, the client can't know when the NAT-box changes IP. But certainly the server can notice a new IP is talking to it. And if we add a method for client authentication that's not based on IP then surely we can weather the NAT-box IP change without losing our marbles. What would it take to implement something like this? Didn't Transarc recently add support for multi-homed clients? Isn't there something from that work that can be leveraged here? -Mitch From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from dsl254-116-177-nyc1.dsl-isp.net by penn.central.org id aa12124; 4 Jan 2001 21:45 EST Received: from o-o.yi.org (IDENT:def@localhost.localdomain [127.0.0.1]) by oo.yi.org (8.9.3/8.8.7) with ESMTP id VAA09391 for ; Thu, 4 Jan 2001 21:45:22 -0500 Message-Id: <200101050245.VAA09391@oo.yi.org> X-Mailer: exmh version 2.1.1 10/15/1999 To: openafs-info@openafs.org From: aeneous@speakeasy.org Reply-To: aeneous@speakeasy.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall In-Reply-To: Your message of "Thu, 04 Jan 2001 19:05:38 EST." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 04 Jan 2001 21:45:22 -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: > Ok, the client can't know when the NAT-box changes IP. But certainly > the server can notice a new IP is talking to it. And if we add a > method for client authentication that's not based on IP then surely we > can weather the NAT-box IP change without losing our marbles. What > would it take to implement something like this? Didn't Transarc > recently add support for multi-homed clients? Isn't there something > from that work that can be leveraged here? Hm. The server notices a new** IP address, and makes an InitCallbackState call. The client discards all callbacks from that server. So that clears things up just fine. The problem is that there is an interval between (1) the time the NAT box's IP address changes and (2) the time the CM decides it needs to call the server (possibly up to 10 minutes or more). During that interval, changes by other clients, to files held by this CM, will generate callbacks to the old, and now "wrong" IP address. Nothing that you can do at either endpoint can detect an IP address change until some contact is made. Therefore, we have only two choices: A. ignore this race condition B. implement something on the NAT box itself. One cheap hack would be for the NAT box to keep track of the file server IP addresses used, and then, when its IP address changes, make InitCallBack calls. It would work for the OP. ------------- ** one small problem. If the address isn't actually new, but one that had been in use by some other client until just recently, then we get screwed. The fileserver won't make initcallbackstate calls in that case. I'm not too happy with that scenario -- it's the sort of thing that causes weird cache inconsistencies that are impossible to debug. From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa12130; 4 Jan 2001 21:47 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id VAA26576; Thu, 4 Jan 2001 21:47:06 -0500 To: Sascha Silbe Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> <20010105000833.F606@cube.sascha.silbe.org> From: Derek Atkins Date: 04 Jan 2001 21:47:06 -0500 In-Reply-To: Sascha Silbe's message of "Fri, 5 Jan 2001 00:08:33 +0100" Message-ID: Lines: 39 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sascha Silbe writes: > DA> All output from the client cache manager exits via UDP port 7001, and= > =20 > DA> all responses to the cache manager (including callbacks) return via UDP > DA> port 7001. > Fine. How about the source port of the callback packets? Is it the same as > the destination port of the original request? Yes. > DA> Callbacks can pretty much occur at any time. HOWEVER, the client > DA> cache manager 'pings' each server periodically to make sure it is > DA> still there. > Wonderful. Does it use the same destination ports for both requests and pin= > gs? Yes. > DA> 1) stop AFS > Did that. > > DA> 2) remove the AFS module > /etc/init.d/afs seems to already do that for me. > > DA> 3) restart AFS > Did that, but no effect. :( Hrm.. That's WEIRD.. When you restart AFS it should build new connections to the AFS Servers, which should re-initialize any NAT mappings. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa12138; 4 Jan 2001 21:50 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id VAA26580; Thu, 4 Jan 2001 21:50:00 -0500 To: Mitch Collinsworth Cc: Sascha Silbe , openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall References: From: Derek Atkins Date: 04 Jan 2001 21:49:59 -0500 In-Reply-To: Mitch Collinsworth's message of "Thu, 4 Jan 2001 19:05:38 -0500 (EST)" Message-ID: Lines: 35 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Mitch Collinsworth writes: > I agree that NAT is not the greatest thing to come down the road, but > it's already here. The battle to prevent it was lost a long time ago. > It's going to be with us for a while and there are lots of people for > whom it's going to be a fact of life. Some of them are even going to > turn out to be important to your or my or someone else's livelihood > here and it would be better for most of us if we figured out a way to > make it work. I'd like to see us take this sort of thing as a > challenge to improve the software rather than write off the whole idea. The problem is that OpenAFS _must_ remain wire-compatible with Transarc/IBM AFS. This implies that we can't just go change the wire protocol or how it behaves. > Ok, the client can't know when the NAT-box changes IP. But certainly > the server can notice a new IP is talking to it. And if we add a > method for client authentication that's not based on IP then surely we > can weather the NAT-box IP change without losing our marbles. What > would it take to implement something like this? Didn't Transarc > recently add support for multi-homed clients? Isn't there something > from that work that can be leveraged here? Yes, the server can notice a new IP.. But do the NAT mappings still work? Does the server respond to the same port numbers? > -Mitch -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE327F9.dip.t-dialin.net by penn.central.org id aa12314; 4 Jan 2001 22:08 EST Received: (qmail 3037 invoked from network); 5 Jan 2001 03:08:11 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 5 Jan 2001 03:08:11 -0000 Received: (qmail 2677 invoked by uid 1000); 5 Jan 2001 03:08:10 -0000 Date: Fri, 5 Jan 2001 04:08:09 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall Message-ID: <20010105040809.B554@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> <20010105000833.F606@cube.sascha.silbe.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="zx4FCpZtqtKETZ7O" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: ; from warlord@MIT.EDU on Thu, Jan 04, 2001 at 09:47:06PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --zx4FCpZtqtKETZ7O Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Jan 04, 2001 at 09:47:06PM -0500, Derek Atkins [DA] wrote: [Ports are exactly the same] That's great. So it is going to work without any further modifications. [restarting AFS client after IP change] DA> Hrm.. That's WEIRD.. When you restart AFS it should build new DA> connections to the AFS Servers, which should re-initialize any NAT DA> mappings. Just got a new IP (while writing this mail :) ) and tried again: =3D=3D=3D begin xterm1 =3D=3D=3D root@cube:/home/sascha# /etc/init.d/afs stop Stopping AFS services.....=20 inetd.afs: no process killed root@cube:/home/sascha# lsmod Module Size Used by 3dfx 3760 0 (unused) emu10k1 40768 5=20 ac97_codec 7760 0 [emu10k1] root@cube:/home/sascha# /etc/init.d/afs start Starting AFS services.....=20 [skipped legal notes] afsd: All AFS daemons started. root@cube:/home/sascha#=20 =3D=3D=3D end xterm1 =3D=3D=3D =3D=3D=3D begin xterm2 =3D=3D=3D sascha@cube:~$ ls $AFSHOME ls: /afs/informatik.uni-tuebingen.de/home/silbe: Connection timed out sascha@cube:~$=20 =3D=3D=3D end xterm2 =3D=3D=3D :( CU/Lnx Sascha --zx4FCpZtqtKETZ7O Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6VTqZa8hvv2E1w1sRAoTsAJ9HZL2qEp5tK9Ox+KcqPvl+4CoVVACeKNCd +tUrPCPD/nJfMPAIJDip0Hs= =rhup -----END PGP SIGNATURE----- --zx4FCpZtqtKETZ7O-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3EE327F9.dip.t-dialin.net by penn.central.org id aa12355; 4 Jan 2001 22:11 EST Received: (qmail 3241 invoked from network); 5 Jan 2001 03:11:36 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 5 Jan 2001 03:11:36 -0000 Received: (qmail 2699 invoked by uid 1000); 5 Jan 2001 03:11:35 -0000 Date: Fri, 5 Jan 2001 04:11:35 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall Message-ID: <20010105041135.C554@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org References: <200101050245.VAA09391@oo.yi.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="rQ2U398070+RC21q" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: <200101050245.VAA09391@oo.yi.org>; from aeneous@speakeasy.org on Thu, Jan 04, 2001 at 09:45:22PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --rQ2U398070+RC21q Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Thu, Jan 04, 2001 at 09:45:22PM -0500, aeneous@speakeasy.org [a] wrote: a> B. implement something on the NAT box itself. How about a native AFS proxy? CU/Lnx Sascha --rQ2U398070+RC21q Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE6VTtna8hvv2E1w1sRAvKMAJ951OloWZzKFH+uWOOa6m5UkaZ2hQCdE4y1 V9usY9i2K8Rl4uCEZ6DhOZQ= =wBsi -----END PGP SIGNATURE----- --rQ2U398070+RC21q-- From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from ORANGE-TOUR.IHTFP.ORG by penn.central.org id aa12561; 4 Jan 2001 23:02 EST Received: (from warlord@localhost) by rcn.ihtfp.org (8.9.3) id XAA26611; Thu, 4 Jan 2001 23:02:01 -0500 To: Sascha Silbe Cc: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> <20010105000833.F606@cube.sascha.silbe.org> <20010105040809.B554@cube.sascha.silbe.org> From: Derek Atkins Date: 04 Jan 2001 23:02:01 -0500 In-Reply-To: Sascha Silbe's message of "Fri, 5 Jan 2001 04:08:09 +0100" Message-ID: Lines: 14 X-Mailer: Gnus v5.5/Emacs 20.3 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: Sascha Silbe writes: > sascha@cube:~$ ls $AFSHOME > ls: /afs/informatik.uni-tuebingen.de/home/silbe: Connection timed out > sascha@cube:~$=20 Wait 10 minutes, then unlog and aklog. -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH warlord@MIT.EDU PGP key available From somebody Thu Jan 01 00:00:00 UTC 1970 Received: from p3E9E126C.dip.t-dialin.net by penn.central.org id aa12719; 4 Jan 2001 23:29 EST Received: (qmail 9104 invoked from network); 5 Jan 2001 04:29:32 -0000 Received: from cube.sascha.silbe.org (192.168.1.2) by sphere.sascha.silbe.org with SMTP; 5 Jan 2001 04:29:32 -0000 Received: (qmail 5615 invoked by uid 1000); 5 Jan 2001 04:29:30 -0000 Date: Fri, 5 Jan 2001 05:29:29 +0100 From: Sascha Silbe To: openafs-info@openafs.org Subject: Re: [OpenAFS] AFS-Client behind masquerading firewall Message-ID: <20010105052929.D554@cube.sascha.silbe.org> Mail-Followup-To: openafs-info@openafs.org References: <20001219043252.A2595@cube.sascha.silbe.org> <20010104222914.B606@cube.sascha.silbe.org> <20010105000833.F606@cube.sascha.silbe.org> <20010105040809.B554@cube.sascha.silbe.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="LTeJQqWS0MN7I/qa" Content-Disposition: inline User-Agent: Mutt/1.3.9i In-Reply-To: ; from warlord@MIT.EDU on Thu, Jan 04, 2001 at 11:02:01PM -0500 Sender: openafs-info-admin@openafs.org Errors-To: openafs-info-admin@openafs.org X-BeenThere: openafs-info@openafs.org X-Mailman-Version: 2.0rc1 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: OpenAFS Info/Discussion List-Unsubscribe: , List-Archive: --LTeJQqWS0MN7I/qa Content-Type: text/plain; charset=us-as