From: SourceForge.net <no...@so...> - 2005-10-19 18:20:53
|
Bugs item #1277363, was opened at 2005-08-31 05:28 Message generated for change (Comment added) made by bernardli You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=109368&aid=1277363&group_id=9368 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Packages Group: 4.2 Status: Open Resolution: Fixed Priority: 9 Submitted By: Erich Focht (efocht) Assigned to: Erich Focht (efocht) Summary: pvm test fails on rhel3/4-ia64 Initial Comment: There is no error message, I'd suggest the test script is improved as well. ---------------------------------------------------------------------- >Comment By: Bernard Li (bernardli) Date: 2005-10-19 11:20 Message: Logged In: YES user_id=879102 I guess if the older version works with x86, then perhaps we can leave the rebuilding until 5.0 - if you guys agree, then let's close this bug ASAP. ---------------------------------------------------------------------- Comment By: Erich Focht (efocht) Date: 2005-10-14 02:56 Message: Logged In: YES user_id=338721 Merged fixes to branch (svn 3802). Should we rebuild the other arch/distros, too? Then they would have the same RPM versions as on ia64 and x86_64. If yes, then we need urgently rebuild for i686 on rhel3, rhel4, fc, mdk. ---------------------------------------------------------------------- Comment By: John (muglerj) Date: 2005-10-13 13:17 Message: Logged In: YES user_id=505737 bumped ---------------------------------------------------------------------- Comment By: Erich Focht (efocht) Date: 2005-10-12 10:37 Message: Logged In: YES user_id=338721 Checked in new SRPM where the 64bit recognition is built into the spec file. This should save us from the same mistake at the next rebuild. I've rebuilt rhel3/4 ia64 pvm packages and can confirm that this fixes the problem. Removed the SRPMS/README file, as that info isn't valid any more. ---------------------------------------------------------------------- Comment By: Thomas Naughton (naughtont) Date: 2005-10-11 13:42 Message: Logged In: YES user_id=288102 I'll have to look back at the tests, but there is a README in the SRPMS dir and the ia64 rpms *must* be built in that manner. The x86_64 rpms would just fall back to 32-bit mode and not fail as brilliantly....good/bad? :) ---------------------------------------------------------------------- Comment By: Erich Focht (efocht) Date: 2005-10-11 13:38 Message: Logged In: YES user_id=338721 That's it, probably. The module file says: setenv PVM_ARCH LINUX and append-path PATH /opt/pvm3/lib/LINUX append-path PATH /opt/pvm3/bin/LINUX This is the same for ia64 and x86_64. I thought for x86_64 the test was passing, are you sure this is needed there, too? ---------------------------------------------------------------------- Comment By: Thomas Naughton (naughtont) Date: 2005-10-11 10:27 Message: Logged In: YES user_id=288102 I agree that the test should be improved, very minimalistic. In the mean time, can you see if this was built with LINUX64 as the arch. Should also be in the module file. ---------------------------------------------------------------------- Comment By: Erich Focht (efocht) Date: 2005-10-11 03:09 Message: Logged In: YES user_id=338721 Tested with rhel3 and rhe4 ia64 (oscar4.2beta4). Now I have an error message from the rhel3 ia64 pvm test failure: /var/spool/pbs/mom_priv/jobs/9.t2rhel3.SC: line 36: 2740 Segmentation fault pvmd pvm_nodes (wd: ~/pvm) libpvm [pid2754] /var/spool/pbs/mom_priv/jobs/9.t2rhel3.SC: line 40: 2754 Segmentation fault ./master1 pvmd3: no process killed ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=109368&aid=1277363&group_id=9368 |