Avoid repeatedly restarting from bad checkpoint
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 108
- Forks
- 37
- PR merge metrics
- No merged PRs in 30d
Description
We may want to add support to give up restarting from a checkpoint if it is attempted too many times in a row. Consider a case where some checkpoint file has been corrupted, which then leads to a crash in the application while reading its file. If we mark that a restart was attempted from a checkpoint but not completed, we could identify this situation on the next restart attempt. We could then mark that checkpoint as bad and fall back to try the next most recent checkpoint.
This should be supported for both checkpoints in cache (which would then be deleted) and those fetched from the file system (then marked as bad but perhaps not deleted).
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
No files, tests, or entry points are named. Start by tracing checkpoint restart handling for cached and filesystem checkpoints, then define how repeated incomplete restarts identify a bad checkpoint and select the next most recent one; done means both storage paths support this behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c
- Domain
- distributed-systems
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 30/100