GHA failures in aiidalab-launch

[danielhollas]

Tests seem to be timing out randomly on GHA (e.g. after 30 minutes).
Interestingly, the pytest tests seem to always finish fine, but then the workflow step gets stuck? Very strange...

See for example

https://github.com/aiidalab/aiidalab-launch/actions/runs/3550589631/jobs/5972824480

[danielhollas]

Maybe duplicate of #154, see discussion there.

[unkcpz]

thanks for reporting this @danielhollas. When #154 happened, it is stuck in the command line rather than printing an error message and exiting so you can type new on the terminal, correct?

[danielhollas]

It is stuck and I needed to the Ctrl+C and only then the error was printed.

[unkcpz]

I see, I'll take a look at #154. The error seems like the async function is called from another coroutine directly without await.

[danielhollas]

Yeah. Just note from the discussion there that Simon already spent some time on it and couldn't figure it out, so it might be a pain, maybe not the biggest priority now, although it is annoying.

btw: Could you make a release of aiidalab-launch. I'd like to start using the two minor improvements that I implemented. Thanks!

[unkcpz]

Sure, I'll make the release. Thanks a lot for the fixes!

[danielhollas]

In failures here and here),the test suite no longer times out, but actually fails in a reasonable amount of time. The failures are too generic to pinpoint the issue, since it says the container fails to start. Probably more debug logging is needed to debug this, we need the pytest to somehow print the container logs when it fails to start.