Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Action hangs indefinitely if oidc-beacon malfunctions #117

Open
jku opened this issue Jan 3, 2024 · 8 comments
Open

Action hangs indefinitely if oidc-beacon malfunctions #117

jku opened this issue Jan 3, 2024 · 8 comments
Labels
bug Something isn't working

Comments

@jku
Copy link
Member

jku commented Jan 3, 2024

At some point after Dec 28 the sigstore-conformance action runs have started to not complete.

This is visible in both sigstore-python and sigstore-java PR workflow runs but will also happen in the selftest workflow here (I tried by rerunning a workflow that succeeded two weeks ago: it now hangs).

I have not identified a reason for this. I've looked at logs and the GitHub environment seems unchanged, as does Python. The only change I can see is that installing requirements.txt now leads to installing attrs 23.2.0 instead of 23.1.0 -- but pinning the older version does not change the situation so seems unrelated.

sigstore/sigstore-python#857

EDIT: this is because sigstore-conformance/extremely-dangerous-public-oidc-beacon was malfunctioning. Let's try to modify the action so that it gives some indication of what is going wrong in this case

@jku jku added the bug Something isn't working label Jan 3, 2024
@jku
Copy link
Member Author

jku commented Jan 3, 2024

strace says that after collecting the tests (timestamp 2024-01-03T13:19:03.4878640Z in the raw log)

  • pytest creates temporary copies of the assets (this finishes around 2024-01-03T13:19:03.5330507Z).
  • then it ends up in a loop where every 30 secs it reads some networking files (like reads ca certs) and makes a connection to a github IP... then starts over
  • I think it never starts an actual test

🤷

https://github.com/jku/sigstore-conformance/actions/runs/7397783000/job/20125543111 (this is slightly modified in addition to strace: only a single test gets collected here)

@jku
Copy link
Member Author

jku commented Jan 3, 2024

then it ends up in a loop

My guess: this is the while loop in identity_token() in test/conftest.py

@steiza
Copy link
Member

steiza commented Jan 3, 2024

This was probably caused by the OIDC beacon job which stopped running 5 days ago: https://github.com/sigstore-conformance/extremely-dangerous-public-oidc-beacon/actions

@jku
Copy link
Member Author

jku commented Jan 3, 2024

oh yeah, that makes sense. I've enabled the trigger action again.

Let's leave this open so we can add some sort of error that allows root-causing this failure

@jku jku changed the title Action hangs indefinitely Action hangs indefinitely if oidc-beacon malfunctions Jan 3, 2024
@woodruffw
Copy link
Member

The "scheduled actions stop running after 60 days of inactivity" feature strikes again 🙂

@woodruffw
Copy link
Member

Oh, maybe spoke too soon. I don't see the normal warning that appears when a scheduled workflow is disabled.

@jku
Copy link
Member Author

jku commented Jan 3, 2024

I don't see the normal warning that appears when a scheduled workflow is disabled.

I re-enabled it 7 minutes ago.

@woodruffw
Copy link
Member

For reference, tracker from when this last happened: sigstore-conformance/extremely-dangerous-public-oidc-beacon#4

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

3 participants