upload times out after 6hrs

[adrianinsaval]

has happened twice now:
https://github.com/adrianinsaval/pacman-repo/actions/runs/4262521615/jobs/7418125145
https://github.com/adrianinsaval/pacman-repo/actions/runs/4242812930/attempts/1

[adrianinsaval]

third time: https://github.com/adrianinsaval/pacman-repo/actions/runs/4282485820/jobs/7456820854
how to know where it hangs? is it happening only to me?

[svenstaro]

I can't really reproduce this problem on my own repositories. I wonder what the difference is.

[adrianinsaval]

this keeps happening every now and then seemingly at random

[svenstaro]

Would you like to try to research this in depth? I'm not sure where the timeout is coming from and without a proper reproducible test case it's going to be rather hard. We already have retries but it doesn't seem to trigger for you error case.

[paddatrapper]

We are experiencing this too - https://github.com/flanksource/karina/actions/runs/4729694790/jobs/8392508080 and https://github.com/flanksource/karina/actions/runs/4363985067/jobs/7630860313 for example. Not sure what causes it - it is random and doesn't affect a specific artefact reliably.

[adrianinsaval]

Would you like to try to research this in depth? I'm not sure where the timeout is coming from and without a proper reproducible test case it's going to be rather hard. We already have retries but it doesn't seem to trigger for you error case.

what could I do to research this?

[svenstaro]

@adrianinsaval I'd do this in your case: Clone the action and add a ton of debug messages. Change your workflow to use your fork. Run the workflow in a loop to try to eventually hit the timeout. Check which of your debug messages came through and then fix the code accordingly.

This is tedious but I don't really know of a better approach. We might have to configure some connect timeouts and retry on timeout or something like that.