I submitted perl-IO-stringy:2.113 for building yesterday. Three tasks were created: 12910, 12911, and 12912. The last two reached the final "ready" state, but the first one https://mbs.fedoraproject.org/module-build-service/2/module-builds/12910 is stuck in "done" state.
I cannot resubmit the build because:
$ fedpkg module-build Submitting the module build... Could not execute module_build: The build failed with: Module (state=5) already exists. Only a new build, resubmission of a failed build or build against new buildrequirements is allowed.
and I cannot kill the task because:
$ fedpkg module-build-cancel 12910 Cancelling module build #12910... Could not execute module_build_cancel: The cancellation of module build #12910 failed with: To cancel a module build, it must be in one of the following states: build, init, wait
Please kill the faulty task.
Pinging @breilly
Metadata Update from @mohanboddu: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: low-gain, low-trouble, ops
Any progress? Is there no Fedora admin available?
We simply provide space for this application, we don't develop it or have anyone who knows it's inner workings. ;(
I've mailed Brenden to see if he can take a look...
There are no tasks in MBS, it is entirely event driven. The only tasks are those in Koji, and those are complete.
The done state is, well, basically done. For scratch builds, that is the end, but this is not a scratch build.
The ready state is pretty much a non-op transition that only performed for non-scratch module builds. This transition is also blocked if greenwave gating fails, but I don't believe we are using that feature.
So, looks like somehow MBS missed the state transition message and never ran the done handler (which would have transitioned to ready).
The short version is, there is nothing left to cancel. If you want to rebuild, then you can bump the revision.
There are a handful of old module builds that are both non-scratch and done, but this is the only one in the past year and a half.
It looks like most (all but 2) of the cases are from a single day (2019-08-13). Those may be due to a past mbs bug.
$ curl -s 'https://mbs.fedoraproject.org/module-build-service/2/module-builds/?per_page=2000&state=done&scratch=False' | jq '.items | sort_by(.id) | .[] | {id, time_submitted} ' { "id": 5597, "time_submitted": "2019-08-13T19:44:50Z" } { "id": 5602, "time_submitted": "2019-08-13T19:45:15Z" } { "id": 5603, "time_submitted": "2019-08-13T19:45:15Z" } { "id": 5604, "time_submitted": "2019-08-13T19:45:15Z" } { "id": 5606, "time_submitted": "2019-08-13T19:45:33Z" } { "id": 5607, "time_submitted": "2019-08-13T19:45:33Z" } { "id": 5608, "time_submitted": "2019-08-13T19:45:33Z" } { "id": 5610, "time_submitted": "2019-08-13T19:50:01Z" } { "id": 5611, "time_submitted": "2019-08-13T19:50:01Z" } { "id": 5613, "time_submitted": "2019-08-13T19:50:01Z" } { "id": 5615, "time_submitted": "2019-08-13T19:50:12Z" } { "id": 5616, "time_submitted": "2019-08-13T19:50:12Z" } { "id": 5618, "time_submitted": "2019-08-13T19:50:13Z" } { "id": 5620, "time_submitted": "2019-08-13T19:50:23Z" } { "id": 5622, "time_submitted": "2019-08-13T19:50:24Z" } { "id": 5623, "time_submitted": "2019-08-13T19:50:24Z" } { "id": 5630, "time_submitted": "2019-08-13T19:51:00Z" } { "id": 5631, "time_submitted": "2019-08-13T19:51:00Z" } { "id": 5632, "time_submitted": "2019-08-13T19:51:01Z" } { "id": 5633, "time_submitted": "2019-08-13T19:51:15Z" } { "id": 5634, "time_submitted": "2019-08-13T19:51:15Z" } { "id": 7252, "time_submitted": "2019-12-22T19:41:37Z" } { "id": 12910, "time_submitted": "2021-09-07T08:25:15Z" }
Thanks for the explanation. Building a new version succeeded.
Metadata Update from @ppisar: - Issue close_status updated to: Will Not/Can Not fix - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.