[OpenAFS] 1.4.0rc5 SRPMS with amd64 problems...
Derek Atkins
warlord@MIT.EDU
Thu, 06 Oct 2005 09:28:05 -0400
Thanks.. I don't know exactly why it was stripping off that
extra bit originally. I'll make this change for the next
release. Either rc7 or 1.4.0-final.
Thank you for testing the RPMS. Any other issues you've found
with them?
-derek
"Matthew N. Andrews" <matt@slackers.net> writes:
> I modded openafs-kernel-version.sh to do the echo, and exit before final
> case clause where it strips of the smp, and surely enough it seems to do
> the right thing now.
>
> -Matt
>
>
> Derek Atkins wrote:
>> Replying to my own post... What happens if you modify this file so
>> that it
>> returns the kernel version before stripping off the kernel config
>> mnemonic? Does that cause it to build the correct kernel module for you?
>>
>> -derek
>>
>> Quoting Derek Atkins <warlord@MIT.EDU>:
>>
>>> Yes, the idea is that 'rpmbuild -ba' without any arguments should
>>> create a
>>> module for the currently-running kernel.. But I think there's a
>>> problem with
>>> that. Can you run:
>>>
>>> /usr/src/redhat/SOURCES/openafs-kernel-version.sh
>>>
>>> This will tell me what it's printing out for the kernel. I suspect
>>> this script
>>> is ripping off the extra parts of the kernel so it's getting confused.
>>>
>>> -derek
>>>
>>> Quoting "Matthew N. Andrews" <matt@slackers.net>:
>>>
>>>> I'm also trying to build the 1.4.0rc5 SRPMS on an RHEL 4 amd64
>>>> machine(actually centos4.1, but basically the same). and am seeing that
>>>> only one openafs-kernel rpm is built, and that it contains only
>>>> /lib/modules/2.6.9-11.EL/kernel/fs/openafs/openafs.ko
>>>>
>>>> I am running the 2.6.9-11.ELsmp kernel, but have bothe the smp, and
>>>> uniprocessor kernels installed.
>>>>
>>>> the first 20 lines of the rpmbuild command follow:
>>>>
>>>> Executing(%prep): /bin/sh -e /var/tmp/rpm-tmp.86138
>>>> + umask 022
>>>> + cd /home/matt/rpmbuild/BUILD
>>>> + LANG=C
>>>> + export LANG
>>>> + unset DISPLAY
>>>> + : @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
>>>> + : @@@
>>>> + : @@@ kernel version: 2.6.9-11.EL
>>>> + : @@@ base kernel version:2.6.9-11.EL
>>>> + : @@@ kernel modules dir: /lib/modules/2.6.9-11.EL
>>>> + : @@@ kernel source dir: /lib/modules/2.6.9-11.EL/build
>>>> + : @@@ kernel type: up
>>>> + : @@@ PAM modules dir: /lib64/security
>>>> + : @@@ build userspace: 1
>>>> + : @@@ build modules: 1
>>>> + : @@@ arch: x86_64
>>>> + : @@@ target cpu: x86_64
>>>> + : @@@
>>>> + : @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
>>>>
>>>> if I explicitly specify that I want to build for the smp kernel by
>>>> running 'rpmbuild -ba rpmbuild/SPECS/openafs.spec --define "kernvers
>>>> `uname -r`"' I get a smp kernel module in it's own rpm and that works
>>>> great, but if your goal is to make the default be to build an rpm for
>>>> the running kernel, then something's missing
>>>>
>>>>
>>>> -Matt
>>>>
>>>>
>>>>
>>>> Derek Atkins wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> Let's focus on RHEL4 for the moment, okay?
>>>>> 1) What rpmbuild command do you give?
>>>>> 2) What's the first 20 lines of output from rpmbuild?
>>>>>
>>>>> -derek
>>>>>
>>>>> Kurt Seiffert <seiffert@indiana.edu> writes:
>>>>>
>>>>>
>>>>>> I'm trying to work with the SRPM's for 1.4.0rc5.
>>>>>>
>>>>>> We are compiling IBM xSeries 346 with Xeon cpu's. We have RHEL AS
>>>>>> v4 64 AMD64/
>>>>>> Intel EM64T installed on one system and RHEL AS v3 for AMD64/Intel
>>>>>> EM64T on
>>>>>> the other system.
>>>>>>
>>>>>> Both systems are smp systems, though uname -r on the RHEL4 system
>>>>>> responds
>>>>>> with:
>>>>>> Linux rfs2.uits.iupui.edu 2.6.9-11.ELsmp #1 SMP Fri May 20 18:25:30
>>>>>> EDT 2005
>>>>>> x86_64 x86_64 x86_64 GNU/Linux
>>>>>>
>>>>>> while the RHEL3 responds with:
>>>>>> Linux rfs-gw2.iu.edu 2.4.21-37.EL #1 SMP Wed Sep 7 13:32:18 EDT
>>>>>> 2005 x86_64
>>>>>> x86_64 x86_64 GNU/Linux
>>>>>>
>>>>>> The afs system type for both appears to be amd64_linux.
>>>>>>
>>>>>> On the RHEL4 system the RPM's refuse to build a kernel-smp module
>>>>>> to install.
>>>>>>
>>>>>> [root@rfs2 openafs-1.4.0-rc5]# ls /usr/src/redhat/RPMS/x86_64/
>>>>>> openafs-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-kernel-1.4.0rc5-2.6.9_11.EL_1.x86_64.rpm
>>>>>> openafs-client-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-kernel-source-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-compat-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-kpasswd-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-debuginfo-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-krb5-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-devel-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-server-1.4.0rc5-1.1.x86_64.rpm
>>>>>> openafs-docs-1.4.0rc5-1.1.x86_64.rpm
>>>>>>
>>>>>> If I install the openafs-kernel rpm, the /etc/init.d/openafs-client
>>>>>> start
>>>>>> fails because it can't find an smp kernel module:
>>>>>>
>>>>>> [root@rfs2 openafs-1.4.0-rc5]# /etc/init.d/openafs-client start
>>>>>> Starting openafs-client: FATAL: Module openafs not found.
>>>>>> failed to load openafs kernel module. [FAILED]
>>>>>> [root@rfs2 openafs-1.4.0-rc5]# rpm -ql openafs-kernel
>>>>>> /lib/modules/2.6.9-11.EL/kernel/fs/openafs/openafs.ko
>>>>>>
>>>>>> Under RHEL3, I have a different problem. The module loads but then
>>>>>> afsd fails
>>>>>> to start. I get an error in the /var/log/messages that seems to
>>>>>> indicate that
>>>>>> the openafs module is built for 32 bit rather than 64 bit.
>>>>>>
>>>>>> [root@rfs-gw1 x86_64]# /etc/init.d/openafs-client start
>>>>>> Starting openafs-client: Warning: loading
>>>>>> /lib/modules/2.4.21-37.EL/kernel
>>>>>> /fs/openafs/openafs.o will taint the kernel: non-GPL license -
>>>>>> http://
>>>>>> www.openafs.org/dl/license10.html
>>>>>> See http://www.tux.org/lkml/#export-tainted for information about
>>>>>> tainted modules
>>>>>> Module openafs loaded, with warnings
>>>>>> afsd: Error -1 in basic initialization.
>>>>>> Adding cell 'openafs.org': error -1
>>>>>> afsd: No check server daemon in client.
>>>>>> afsd: All AFS daemons started.
>>>>>> afsd: Can't mount AFS on /afs(22)
>>>>>>
>>>>>> [root@rfs-gw1 x86_64]# tail /var/log/messages
>>>>>> Oct 3 15:46:52 rfs-gw1 kernel: Found system call table at
>>>>>> 0xffffffff805e78c0 (scan: close+chdir+write)
>>>>>> Oct 3 15:46:52 rfs-gw1 kernel: Found 32-bit system call table at
>>>>>> 0xffffffff804412e0 (exported)
>>>>>>
>>>>>> Any ideas?
>>>>>>
>>>>>> Thanks.
>>>>>>
>>>>>> -KAS
>>>>>>
>>>>>> Kurt A. Seiffert | seiffert@indiana.edu
>>>>>>
>>>>>> UITS Distributed Storage Services Group | C: 812-345-1892
>>>>>>
>>>>>> Indiana University, Bloomington | W: 1 812-855-5089
>>>>>>
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> OpenAFS-info mailing list
>>>> OpenAFS-info@openafs.org
>>>> https://lists.openafs.org/mailman/listinfo/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
>>>
>>> _______________________________________________
>>> OpenAFS-info mailing list
>>> OpenAFS-info@openafs.org
>>> https://lists.openafs.org/mailman/listinfo/openafs-info
>>>
>>
>>
>>
>
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/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