Would it be possible to enable runroot plugin on some staging x86_64 builder please? I'm trying to test building ostree containers (pungi#1699), and the started task doesn't seem to ever start: https://koji.stg.fedoraproject.org/koji/taskinfo?taskID=110448742
Looking into ansible, the only stage host in the runroot list is s390x.
Metadata Update from @phsmoura: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: low-gain, low-trouble, ops
For reference, full runroot list:
buildvm-x86-01.iad2.fedoraproject.org buildvm-x86-02.iad2.fedoraproject.org buildvm-x86-03.iad2.fedoraproject.org buildvm-a64-01.iad2.fedoraproject.org buildvm-a64-02.iad2.fedoraproject.org buildvm-a64-03.iad2.fedoraproject.org buildvm-ppc64le-01.iad2.fedoraproject.org buildvm-ppc64le-09.iad2.fedoraproject.org buildvm-ppc64le-18.iad2.fedoraproject.org buildvm-ppc64le-27.iad2.fedoraproject.org buildvm-ppc64le-33.iad2.fedoraproject.org buildvm-s390x-01.stg.s390.fedoraproject.org buildvm-s390x-15.s390.fedoraproject.org buildvm-s390x-16.s390.fedoraproject.org buildvm-s390x-17.s390.fedoraproject.org
From these, the only staging buildvm is buildvm-s390x-01.stg.s390.fedoraproject.org
buildvm-s390x-01.stg.s390.fedoraproject.org
I created PR in ansible to enable runroot for one x86 buildvm. I'm just not sure if just moving it from one group to other is enough.
So, I added one builder of each arch to the runroot group, which should have been enough to enable the plugin (and we already had nfs setup right to rw mount in stg).
However, then I hit that one of our plugins (the scmpolicy one), wasn't in the new f39 version we had there. I built a patched koji package in f39-infra with the right patch for that plugin. Then updated them and... your task started! and failed. ;(
DEBUG util.py:448: rpm-md repo 'repo-0'; generated: 2023-11-14T12:11:51Z solvables: 151709 DEBUG util.py:448: error: Packages not found: bootc DEBUG util.py:448: error: compose tree failed: ExitStatus(unix_wait_status(256))
But thats a new error, and I think this is now solved. :)
Let us know if there's anything further needed from our side.
Metadata Update from @kevin: - Issue close_status updated to: Fixed with Explanation - Issue status updated to: Closed (was: Open)
I agree this is fixed from Koji perspective. The error is caused elsewhere. Out of curiosity, what's the timeout for cancelling stalled tasks? The x86_64 builder is currently handling two RPM builds that are taking over 30 hours and not showing any progress in logs.
It's 201600 (56 hours)...
https://koji.stg.fedoraproject.org/koji/taskinfo?taskID=110448742 failed! (so this is a success!) Thanks!
Log in to comment on this ticket.