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

Alias don't get updated properly #224

Closed
albertz opened this issue Nov 28, 2024 · 0 comments · Fixed by #225
Closed

Alias don't get updated properly #224

albertz opened this issue Nov 28, 2024 · 0 comments · Fixed by #225

Comments

@albertz
Copy link
Member

albertz commented Nov 28, 2024

It seems some of the aliases don't get updated properly anymore. I have this case now for some alias which already existed from some earlier run, but I changed the config, thus it got a new job hash, and should update this alias.

This has always worked fine. But now it does not anymore. This might be a result of #214.

In main output, I see:

...
[2024-11-28 06:08:27,767] INFO: running: Job<alias/ctc/time4-n12-spm512-blankSep-auxAED-b150k-ctcFixGrad/train work/i6_core/returnn/training/ReturnnTrainingJob.aJAdtFpzqsEc> {ep 20/100} 
[2024-11-28 06:08:27,767] INFO: running(1) waiting(88)

However, the symlink points to an old job:

alias/ctc/time4-n12-spm512-blankSep-auxAED-b150k-ctcFixGrad/train -> /rwthfs/rz/cluster/hpcwork/az668407/setups-data/combined/2021-05-31/work/i6_core/returnn/training/ReturnnTrainingJob.9qiRvGpHLhgz

I also waited now a couple of hours, in case the create alias thread was hanging or so.

I also restarted the manager now. It still does not update the alias.

(cc @Atticus1806)

albertz added a commit that referenced this issue Nov 28, 2024
Fix #224

What I don't understand:
Why did I not get any exception?
Without this change, it just did not call the func
(or I don't know, but at least it did not work, it did nothing).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant