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

bash runner reports that "jobs were successfully submitted" but does not submit anything #51

Open
bhaddow opened this issue Aug 17, 2024 · 3 comments
Assignees
Labels
bug Something isn't working need investigation Unknown scope

Comments

@bhaddow
Copy link
Contributor

bhaddow commented Aug 17, 2024

For the bash runner, it doesn't really make sense to talk about submitting jobs. Also jobs running detached are harder to track. What should be the behaviour? I would be happy for the bash runner to block whilst jobs are executed.

@varisd
Copy link
Contributor

varisd commented Aug 19, 2024

The idea is that with bash runner, we don't have any scheduler process (if we had, there would be a risk of it being killed by the HPC job timeout, for example). For this reason, each submitted job process (running as detached) is in charge of tracking whether their dependencies are still being executed or finished and then running its own code. OpusPocus only takes care of submitting (i.e running) the job processes.

We can rework bash runner to be similar to the DebugRunner, specifically, the job processes will be submitted (in parallel, if feasible) whenever possible and then OpusPocus will be waiting until the processes finished and continue submitting processes which have their dependencies satisfied (successfully finished executing).

@bhaddow
Copy link
Contributor Author

bhaddow commented Aug 19, 2024

"Rework" sounds like it's a significant change. Maybe the primary issue is one of documentation/communication. At the moment, when I launch bashrunner, it is not clear what has happened, where my processes are. There's just a mysterious message about jobs being "submitted" and then it exits.

Maybe if the message says "jobs have been launched in the background". Using ./go.py status to check their progress." it would be more useful?

@varisd
Copy link
Contributor

varisd commented Aug 19, 2024

Maybe if the message says "jobs have been launched in the background". Using ./go.py status to check their progress." it would be more useful?
I think we can add at least report of under which jobid (process id) the job was submitted. I will check whether we can add the job/process info to the "status" subcommand that is, if the step is SUBMITTED/RUNNING it will also list pid/jid info.

@rggdmonk rggdmonk added bug Something isn't working need investigation Unknown scope labels Aug 20, 2024
@rggdmonk rggdmonk added this to the MVP - August 31, 2024 milestone Aug 23, 2024
@varisd varisd self-assigned this Aug 23, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working need investigation Unknown scope
Projects
None yet
Development

No branches or pull requests

3 participants