The fedora 33 schedule[1] has a mass rebuild scheduled for Jul 22 2020, We need to plan and coordinate all tasks in preparation for it.
[1] https://fedorapeople.org/groups/schedule/f-33/f-33-key-tasks.html
list for mass rebuild:
https://pagure.io/releng/issue/9548 https://pagure.io/releng/issue/9539 https://pagure.io/releng/issue/9533 https://pagure.io/releng/issue/9524 https://pagure.io/releng/issue/9505 https://pagure.io/releng/issue/9347 https://pagure.io/releng/issue/9308 https://pagure.io/releng/issue/9286 https://pagure.io/releng/issue/9253 https://pagure.io/releng/issue/9119
Metadata Update from @humaton: - Issue assigned to humaton
FYI, FESCo is thinking about delaying mass rebuild for a few days due to not landed changes for LTO and AArch64 PAC/BTI.
https://pagure.io/fesco/issue/2451
Yeah, although fesco said we could delay up to 5 days without any approvals. ;)
@mohanboddu @humaton We should look and see where we are later tonight and decide how long we might want to delay and send out a delay announcement.
From what I see right now, I think thursday might be good, but you all probibly have checked all the changes more recently than I.
Thursday seems good, but I am not 100% confident.
Builds are failing on s390x with the error:
BuildrootError: Requested repo (1785306) is DELETED
https://koji.fedoraproject.org/koji/taskinfo?taskID=47924243 https://koji.fedoraproject.org/koji/taskinfo?taskID=47924757
This may be related to a change I made in kojira last night to try and free up space on the koji volume. Due to that it was deleting expired repos after 300 seconds... but with people building tons of things into f33, the buildroot is getting regenerated a bunch and the s390x builds might not be starting until after the repo they were going to use was deleted. I've reverted my change back, so it should stop doing this soon. We may want to resbumit those after we are done (or perhaps even just resubmit everything when we are done to give it a second chance to build)
This may be a silly question - but why did the "Second Attempt" for failed builds introduce a new commit and Release bump, instead of just resubmitting the failed packages as-is?
We were asked to do it that way as there were some eln fixes that didn't get in the first round of mass rebuild and they wanted to get those in when they did the followon eln rebuild.
I didn't think it would harm anything... are we low on integers? :)
That makes sense. Thanks for the explanation!
No, it was just unexpected, since up until now, only unmodified resubmissions were done for failed builds. But if that interferes with ELN rebuilds, then that's that :)
Closing the ticket as the work is done here.
Metadata Update from @mohanboddu: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.