CI coverage for batch launchers
- Dominant language
- Jupyter Notebook
- Stars
- 2.6k
- Forks
- 1k
- PR merge metrics
- No merged PRs in 30d
Description
We have CI coverage for batch through SLURM, but not the other batch launchers (beyond template generation).
[dask-jobqueue][] has CI for most of our supported batch systems, deployed via docker-compose. We could adapt their setup to get proper exercise of all of most of our supported cluster types.
[dask-jobqueue]: https://github.com/dask/dask-jobqueue/tree/0.7.3/ci
It should mostly be a find-replace of installing dask to installing ipyparallel, since all the cluster setup itself should be reusable.
Contributor guide
Research direction
Start with the dask-jobqueue ci directory linked in the issue and compare its docker-compose batch-system setup with ipyparallel's existing SLURM CI coverage. Adapt the setup to install ipyparallel and exercise the other supported batch launchers, with CI runs demonstrating coverage beyond template generation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- docker-compose, python
- Domain
- ci-cd, distributed-systems
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100