Checkpointing system
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 13
- Forks
- 21
- Avg merge
- 1d 11h
- Merged PRs (30d)
- 8
Description
When running regressions there can be ~45k jobs that are run as part of the regression. It can be a bit of a pain if the regression fails due to some intermittent external issue. Being able to retry failed jobs or restart a killed regression run would be nice.
- [ ] Store enough information to disk about the regression run being started
- [ ] Store metadata about the planned jobs up front
- [ ] Store the scheduler state, or enough information about the progress of the regression run so that it can be restored if killed mid run.
- [ ] Add the ability to continued a killed regression run
- [ ] Add the ability to retry failed jobs - this could be useful if the issue is due to resource dependent timeouts.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating DVSim’s Python entry points for regression runs and the scheduler state; no files or tests are named in the issue. Map what metadata and progress must be preserved before defining verification, with restoration of a killed run and retrying failed jobs as the completion checks.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- build-system, tooling
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100