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

integrate with dask-ctl #4

Open
keewis opened this issue Mar 3, 2022 · 2 comments
Open

integrate with dask-ctl #4

keewis opened this issue Mar 3, 2022 · 2 comments

Comments

@keewis
Copy link
Collaborator

keewis commented Mar 3, 2022

There are at least a few related packages that have significant overlap:

Since in particular dask-ctl seems to be a bit more mature, it would be good to see if we can rewrite this package to use dask-ctl underneath, and to upstream any features that are missing.

This would then become a collection of hpc configuration files, and maybe a thin wrapper that would translate names to paths and feed that to dask_ctl.lifecycle.create_cluster.

@keewis
Copy link
Collaborator Author

keewis commented Mar 3, 2022

looking at dask/dask-jobqueue#544 it seems we took the easy way out: we print the scheduler address to stdout or into a file, and we don't really care if the scheduler is on the current machine (because in the most common workflow we're usually on a compute node anyways)

cc @jacobtomlinson for reference, but I will open a new issue on dask-ctl to see what the best way forward would be.

@jacobtomlinson
Copy link

Sounds great, let me know how I can help! For reference dask-jobqueue currently does not fully support dask-ctl for the reasons you mention about the scheduler being a subprocess of the Python code that created the cluster. But for cluster creation from a spec file you should be all good!

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

No branches or pull requests

2 participants