Red Hat Bugzilla – Bug 621179
tps-make-lists from comparative workflow fails on ppc64
Last modified: 2010-08-10 08:25:54 EDT
Description of problem:
tps-make-lists from comparative workflow fails on ppc64 platform.
I have executed comporative workflow in beaker and encounted the bug described in bug 580384. Please take a look at
and comments that folows for additional info.
In beaker there is the same problem with tps-make-lists 2010:9745 4AS ppc/ppc64 as in RHTS. Moreover, after scheduling a job
I have found that /mnt/redhat was not mounted and therefore packages in brewroot were not accessible.
in regard to failing on the ppc64 architecture. Can the tps setup/upgrade tests be modified to undefine ARCH when calling the tps commands?
ARCH="" tps-make-lists $ERRATUM
in regard to the failing to mount /mnt/redhat. Can you point out in the logs where it failed to mount /mnt/redhat? I can't find it myself.
> in regard to the failing to mount /mnt/redhat. Can you point out in the logs
> where it failed to mount /mnt/redhat? I can't find it myself.
When I tried to execute tps-make-lists command manualy I got an error regarding packages not available in brewroot. That's how I have realized that /mnt/redhat is not mounted. I do not know why because I didn't unmount/reboot the server. Anyway, I tried to schedule the same test again and the problem did not appear, /mnt/redhat is mounted as it should. I think you can ignore it.
I noticed in your previous job that one of the tests local watchdog expired. When this happens we reboot the machine before moving on to the next test. This would clear any nfs mounts.
Comment 2 and Comment 3 say it all: this is a problem with given test being killed by local-watchdog.
Karel, may be the duration is inadequate or there is a problem in itself.
Thanks Bill, for looking at this.
I have opened Bug 622392 for tracking comparative w/ reboot not working.
Reopening the bug. The main problem in this bug is with the tps-make-lists failing on ppc64 platform. The note about the missing mount point was just a remark.
I will ask pmuller to update the workflow test. I tried to test in beaker-stage but for some reason the test has not been updated.
It's in your hands now.
Is it still high priority?
Well, it is simple. We need working rcw for errata testing.
After the update of the test the only problem is with downgrade test, probably not an issue.
so is this all set then? Can we close?
> so is this all set then? Can we close?
I think so. Thank you for help.