Hello there,
I''ve applied two kinds of patches rhel 5.1 kernel. The first one is
perfctr
patch from the package papi-3.6.2. And the second one is lustre''s
kernel
patch series for rhel 5.1 kernels. Then built the kernel rpms without any
error and reboot the host. After rebooting the host, I''ve build ofed
1.4
software stack against the patched kernel source. And again, there are no
problems about ofed stuff. All the things works great. And finally when try
to build lustre rpms against patched kernel using the ofed 1.4 ofa_kernel
for o2ib things I encountered problems during rpm build stage. Belov, you
can find my configure command:
./configure
--with-linux=/root/customLustreAndKernelPackagesForUybhmCns/try1/rpmbuild/BUILD/kernel-2.6.18k3di_lustre.1
--enable-uoss --enable-posix-osd --enable-adaptive-timeouts --enable-efence
--enable-libwrap --disable-snmp --enable-quota --enable-libpthread
--enable-panic_dumplog --with-o2ib=/usr/src/openib --disable-server
--enable-modules
I know that some of the options enabled by default, but i tried lots of
configure options to got it work. But no luck... The configure and make
commands did not return any errors. But the make rpms command returns the
following error:
...
...
...
Executing(%doc): /bin/sh -e /var/tmp/rpm-tmp.68313
+ umask 022
+ cd /root/customLustreAndKernelPackagesForUybhmCns/try1/rpmbuild/BUILD
+ cd lustre-1.6.5.1
+ DOCDIR=/var/tmp/lustre-1.6.5.1-root/usr/share/doc/lustre-modules-1.6.5.1
+ export DOCDIR
+ rm -rf /var/tmp/lustre-1.6.5.1-root/usr/share/doc/lustre-modules-1.6.5.1
+ /bin/mkdir -p
/var/tmp/lustre-1.6.5.1-root/usr/share/doc/lustre-modules-1.6.5.1
+ cp -pr COPYING
/var/tmp/lustre-1.6.5.1-root/usr/share/doc/lustre-modules-1.6.5.1
+ exit 0
Requires(interp): /bin/sh /bin/sh
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
Requires(post): /bin/sh
Requires(postun): /bin/sh
Requires: modutils >= 2.4.10
Processing files: lustre-source-1.6.5.1-2.6.18_k3di_lustre.1
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
Requires: /bin/bash /bin/sh /usr/bin/expect /usr/bin/perl perl(Cwd)
perl(File::Basename) perl(File::Copy) perl(FileHandle) perl(Getopt::Long)
perl(Getopt::Std) perl(POSIX) perl(Socket) perl(Sys::Hostname) perl(strict)
perl(vars)
Processing files: lustre-tests-1.6.5.1-2.6.18_k3di_lustre.1
error: File not found by glob:
/var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/*
error: File not found by glob:
/var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/kernel/fs/lustre/llog_test.*
Processing files: lustre-debuginfo-1.6.5.1-2.6.18_k3di_lustre.1
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
RPM build errors:
File not found by glob:
/var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/*
File not found by glob:
/var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/kernel/fs/lustre/llog_test.*
make[1]: *** [rpms-real] Error 1
make[1]: Leaving directory
`/root/customLustreAndKernelPackagesForUybhmCns/try1/lustre-1.6.5.1''
make: *** [rpms] Error 2
What can be the problem? What kind of info you need for identifying this
problem? Thanks in advance.
Ender GULER
System Administrator
UYBHM
Istanbul Techincal University
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.lustre.org/pipermail/lustre-discuss/attachments/20090326/380e12cd/attachment.html
On Thu, 2009-03-26 at 09:08 +0200, Ender G?ler wrote:> > RPM build errors: > File not found by > glob: /var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/* > File not found by > glob: /var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/kernel/fs/lustre/llog_test.* > make[1]: *** [rpms-real] Error 1 > make[1]: Leaving directory > `/root/customLustreAndKernelPackagesForUybhmCns/try1/lustre-1.6.5.1'' > make: *** [rpms] Error 2 > > > What can be the problem?Some files failed to build/distribute.> What kind of info you need for identifying this problem? Thanks in > advance.Please try with the latest release, 1.6.7. I don''t know that/if that will fix it, but if it does not, then we can at least chase a bug in the current release rather than a past release. b. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part Url : http://lists.lustre.org/pipermail/lustre-discuss/attachments/20090326/78e8ef7e/attachment.bin
Thanks Brian, I''ll try the latest release. Are there any possibility to run servers with lustre v1.6.5.1 and the clients with lustre version 1.6.7? If not, we have to upgrade all the cluster (servers and clients) and it''s not very feasible in our situation. Regards, Ender On Thu, Mar 26, 2009 at 5:59 PM, Brian J. Murrell <Brian.Murrell at sun.com>wrote:> On Thu, 2009-03-26 at 09:08 +0200, Ender G?ler wrote: > > > > > RPM build errors: > > File not found by > > glob: /var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/* > > File not found by > > glob: > /var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/kernel/fs/lustre/llog_test.* > > make[1]: *** [rpms-real] Error 1 > > make[1]: Leaving directory > > `/root/customLustreAndKernelPackagesForUybhmCns/try1/lustre-1.6.5.1'' > > make: *** [rpms] Error 2 > > > > > > What can be the problem? > > Some files failed to build/distribute. > > > What kind of info you need for identifying this problem? Thanks in > > advance. > > Please try with the latest release, 1.6.7. I don''t know that/if that > will fix it, but if it does not, then we can at least chase a bug in the > current release rather than a past release. > > b. > > > _______________________________________________ > Lustre-discuss mailing list > Lustre-discuss at lists.lustre.org > http://lists.lustre.org/mailman/listinfo/lustre-discuss > >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.lustre.org/pipermail/lustre-discuss/attachments/20090326/5ef92f0e/attachment.html
On Thu, 2009-03-26 at 18:28 +0200, Ender G?ler wrote:> Are there any possibility to run servers with lustre v1.6.5.1 and the > clients with lustre version 1.6.7?Yes. Officially it''s outside of our +/-1 version interoperability guarantee, however there have been no changes between those versions that would introduce interoperability issues as far as I recall. Ideally you can always have all of your nodes on the latest release though. b. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part Url : http://lists.lustre.org/pipermail/lustre-discuss/attachments/20090326/e38f0bed/attachment.bin
On Mar 26, 2009, at 8:59 AM, Brian J. Murrell wrote:> On Thu, 2009-03-26 at 09:08 +0200, Ender G?ler wrote: > >> >> RPM build errors: >> File not found by >> glob: /var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/* >> File not found by >> glob: /var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/ >> kernel/fs/lustre/llog_test.* >> make[1]: *** [rpms-real] Error 1 >> make[1]: Leaving directory >> `/root/customLustreAndKernelPackagesForUybhmCns/try1/lustre-1.6.5.1'' >> make: *** [rpms] Error 2 >> >> >> What can be the problem? > > Some files failed to build/distribute. > >> What kind of info you need for identifying this problem? Thanks in >> advance. > > Please try with the latest release, 1.6.7. I don''t know that/if that > will fix it, but if it does not, then we can at least chase a bug in > the > current release rather than a past release.We recently upgraded from 1.6.6 to 1.6.7 and experienced corruption on the MDT. At this point, we have reason to believe that Lustre is the source of the corruption, as there were no clean unmounts of the MDT between instances of corruption, but neither we nor Sun have been able to find the source yet. See bz 18889 and bz18695 for more details. Sun is actively working the issue. We have since downgraded to 1.6.6 on the servers, but are still running 1.6.7 on our clients. j> > > b. > > <signature.asc><ATT00001.txt>-- Jason Rappleye System Administrator NASA Advanced Supercomputing Division NASA Ames Research Center Moffett Field, CA 94035 (650) 604-2875 jason.rappleye at nasa.gov
On Mar 26, 2009, at 3:32 PM, Jason Rappleye <jason.rappleye at nasa.gov> wrote:> > On Mar 26, 2009, at 8:59 AM, Brian J. Murrell wrote: > >> On Thu, 2009-03-26 at 09:08 +0200, Ender G?ler wrote: >> >>> >>> RPM build errors: >>> File not found by >>> glob: /var/tmp/lustre-1.6.5.1-root/usr/lib64/lustre/tests/* >>> File not found by >>> glob: /var/tmp/lustre-1.6.5.1-root/lib/modules/2.6.18-k3di_lustre.1/ >>> kernel/fs/lustre/llog_test.* >>> make[1]: *** [rpms-real] Error 1 >>> make[1]: Leaving directory >>> `/root/customLustreAndKernelPackagesForUybhmCns/try1/lustre-1.6.5.1'' >>> make: *** [rpms] Error 2 >>> >>> >>> What can be the problem? >> >> Some files failed to build/distribute. >> >>> What kind of info you need for identifying this problem? Thanks in >>> advance. >> >> Please try with the latest release, 1.6.7. I don''t know that/if that >> will fix it, but if it does not, then we can at least chase a bug in >> the >> current release rather than a past release. > > We recently upgraded from 1.6.6 to 1.6.7 and experienced corruption on > the MDT. At this point, we have reason to believe that Lustre is the > source of the corruption, as there were no clean unmountsOops...I meant to say no *unclean* umounts,> of the MDT > between instances of corruption, but neither we nor Sun have been able > to find the source yet. See bz 18889 and bz18695 for more details. Sun > is actively working the issue. > > We have since downgraded to 1.6.6 on the servers, but are still > running 1.6.7 on our clients. > > j > >> >> >> b. >> >> <signature.asc><ATT00001.txt> > > -- > Jason Rappleye > System Administrator > NASA Advanced Supercomputing Division > NASA Ames Research Center > Moffett Field, CA 94035 > (650) 604-2875 > jason.rappleye at nasa.gov > > > > > _______________________________________________ > Lustre-discuss mailing list > Lustre-discuss at lists.lustre.org > http://lists.lustre.org/mailman/listinfo/lustre-discuss