Restarting workflow runs of older versions hang forever and cannot be canceled

[Bas94]

Description

Currently the Counter of active jobs hangs at 2 at our instance:

Further investigations shows, that this counter is 2 because 2 Workflow runs are hanging in "waiting"

Those runs cannot be canceled. Clicking on cancel won't lead to any reaction of the system. These runs remain waiting.

In contrast: The actual UI when clicking at the run shows the last run that was done. There nothing shows anything scheduled in the job list. Thus the run is marked for waiting but no job is schedules for a run.

Gitea Version

1.25.0

Can you reproduce the bug on the Gitea demo site?

No

Log Gist

No response

Screenshots

No response

Git Version

No response

Operating System

No response

How are you running Gitea?

We use the current package delivered with cloudron. The cloudron system has updated our system over the weekend. They seem to use the official docker image.

Database

None

[ln-12]

I am having the same issue. Already tried gitea manager flush-queues but that did not help. I am running docker.io/gitea/gitea:1.25.0 together with gitea/act_runner:0.2.13 via docker compose.

[R4nmaru314]

same issue here

[jared-opus]

I have observed this issue as well (gitea 1.25.0 and act_runner 0.2.13)

[Antidote2151]

Same here

[lunny]

Is this a duplicated of #35780 ?

[ln-12]

I don't think so. Re-running itself was also not working for me on 1.25.0 before. Then I tried the current nightly build and could trigger actions again. But the problem with the remaining actions from this issue is still present.

[ChristopherHX]

Possible regression of change #35703 , if no job has effectively been restarted the regular workflow stop event is not triggered

[lunny]

#35840 should fix this?

[Zettat123]

#35840 should fix this?

#35840 can fix the status of remaining runs. Then these runs can be rerun correctly.

[ankurk91]

Can some give us a temporary fix unless 1.25.1 gets release, i can lots of our pipeline stuck due to broken re-run feature.

[ln-12]

I can confirm that updating to the latest main-nightly docker image and running gitea doctor check --run fix-actions-unfinished-run-status --fix inside the container fixed the issue for me. Thanks!