Prevent stale rescue from overwriting finished jobs - #1373
Open
bgentry wants to merge 1 commit into
Open
Conversation
A worker can complete a job after the rescuer fetches it but before the rescue update executes. Updating by ID alone lets that stale snapshot replace the completed state and mutate errors, metadata, and timestamps. Require jobs to still be `running` in the PostgreSQL and SQLite rescue updates so jobs that have left that state are skipped atomically. Preserve normal rescue behavior for eligible jobs in the same batch. Add shared driver coverage that completes a job between fetch and rescue, then checks the entire row remains unchanged for each rescue outcome. Fixes #1302.
bgentry
marked this pull request as ready for review
September 8, 2026 00:34
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A worker can complete a job after
JobRescuerfetches it but before the rescue update executes. Because rescue updates match only by ID, the stale rescue can overwrite the completed state and change errors, metadata, and timestamps.Require the job to still be
runningin the PostgreSQL and SQLite rescue updates. Jobs that have left that state are skipped atomically, while eligible jobs in the same batch are rescued normally.Add shared driver regression coverage that completes a job between fetch and rescue, then verifies the entire completed row remains unchanged for each rescue outcome. Include an unreleased changelog entry.
Fixes #1302.