[OpenAFS] multiplt kernel call traces

[ematlis@yahoo.com]

Simon-=0A=0Ayou are correct, I meant the rpmfusion repository, not the fedo=
ra.  Given what you've just said, I'll stick with your rpms!=0A=0AThanks ve=
ry much for your informative and detailed responses.=0A=0Aeric=0A=0Ap.s.-  =
the machine in question is sort of a beta tester for the Notre Dame campus =
in terms of Fedora 12; the official use of Linux is still RHEL5, which I fi=
nd too archaic for our department.  The research computing IT department ho=
wever is interested in my use of Fedora, so I'm reporting to them my anecdo=
tes.  I can happily say you have been very helpful.=0A=0A--- On Mon, 3/15/1=
0, Simon Wilkinson <sxw@inf.ed.ac.uk> wrote:=0A=0A> From: Simon Wilkinson <=
sxw@inf.ed.ac.uk>=0A> Subject: Re: [OpenAFS] multiplt kernel call traces=0A=
> To: ematlis@yahoo.com=0A> Cc: openafs-info@openafs.org=0A> Date: Monday, =
March 15, 2010, 9:29 AM=0A> =0A> On 15 Mar 2010, at 14:08, ematlis@yahoo.co=
m=0A> wrote:=0A> > As a follow up to your response- will this problem "go=
=0A> away" for kernel version 2.6.33?=A0 If so, perhaps I can=0A> upgrade t=
o that version.=0A> =0A> It will, yes.=0A> =0A> > As an alternative- can I =
fix this by downgrading the=0A> kernel to some lesser version (say 2.6.30)?=
=0A> =0A> Yes. The problematic code is only in 2.6.31 and 2.6.32. Of=0A> co=
urse, vendors may have backported that code into earlier=0A> versions (or n=
ot taken the fixes, in later versions), but=0A> that's one of the joys of t=
he way the Linux kernel gets=0A> distributed.=0A> =0A> > I found I needed p=
am-afs-session; is the library=0A> provided in the rpm going to be included=
 in the default=0A> openafs packages?=0A> =0A> pam-afs-session will always =
be a standalone package. There=0A> is an argument that we should possibly i=
nclude it in our=0A> repository, but we have yet to do so. OpenAFS 1.5.x do=
es=0A> have libkopenafs, which is similar to the library that=0A> pam-afs-s=
ession uses, so that will be available with that=0A> release.=0A> =0A> > La=
stly- since I run Fedora, which now provides openafs=0A> by their own repos=
itories, which repository would you=0A> recommend- theirs or yours?=0A> =0A=
> As far as I'm aware, Fedora doesn't package OpenAFS. Fedora=0A> doesn't a=
ccept packages which introduce new kernel modules=0A> as a matter of policy=
, so we could never get the OpenAFS=0A> client into=A0 Fedora itself. I gue=
ss it would be=0A> possible to get the OpenAFS server into Fedora, but=0A> =
significant changes would have to happen to the way it's=0A> package for it=
 to be acceptable to them.=0A> =0A> RPMFusion, on the other hand, does now =
appear to provide=0A> OpenAFS packages. These are significantly different f=
rom the=0A> ones that OpenAFS ship, so you need to decide fairly early=0A> =
one what you want to support for your site. I've never used=0A> the RPMfusi=
on packages, nor had much communication with=0A> their packager, so I can't=
 really comment on them.=0A> =0A> I'd love it if someone else was maintaini=
ng the OpenAFS=0A> RPMs, because it would be one less thing I need to do wi=
th=0A> each release. However, both my site and many others rely on=0A> the =
RPMs working the same way from release to release. We=0A> can't just switch=
 to a set of 3rd party RPMs at the drop of=0A> a hat, and I think it's a re=
al shame that the people doing=0A> third party RPM packaging never seem to =
come and talk to use=0A> before doing so.=0A> =0A> Given that we don't actu=
ally have contacts for 3rd party=0A> packagers, it's also unlikely that we'=
ll be able to tell=0A> them about impending security fixes, or patches that=
 are=0A> necessary for new security releases - so the OpenAFS RPMs=0A> are =
likely to get these before anyone else.=0A> =0A> This leaves the problem of=
 support. If you are using the=0A> OpenAFS RPMs, then I'm quite likely to k=
now how they were=0A> built, to have a set of kernel debug images floating =
around,=0A> and to be prepared to find the time to debug your problem.=0A> =
If you're using third party RPMs, then I'm as likely as not=0A> to ask you =
to report your problem to them, and to rule out=0A> anything to do with the=
 way they've configured their builds,=0A> and their build system, before I'=
ll take a look at it.=0A> =0A> So, for all of those reasons, I'd recommend =
using the=0A> OpenAFS rpms. Not that the rpmfusion ones are necessarily=0A>=
 bad, in any way, it's just that it's much easier for the=0A> community to =
help you if you're using RPMs that we know the=0A> origin of.=0A> =0A> Chee=
rs,=0A> =0A> Simon.=0A> =0A> =0A=0A=0A