codalab / codalab/codabench

Submissions stuck in "Scoring" when broker error occurs during compute_worker PATCH (non-transactional re-enqueue)

Open
#2,419 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Workers / Queues
Dominant language
Python
Stars
176
Forks
74
Avg merge
5d 3h
Merged PRs (30d)
21

Description

Problem

When the compute_worker PATCHes a submission with status=Scoring, the site re-enqueues the scoring task before committing the new status and outside any transaction. If the broker publish raises (RabbitMQ unreachable, ConnectionError, channel error, broker restart, …), no message ever reaches the compute-worker queue. The DB row is still moved
to Scoring on the next code path, so the submission stays in Scoring forever — no retry, no failover, no cleanup picks it up.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Trace the compute_worker PATCH handling for submissions entering Scoring and the scoring-task re-enqueue path. Reproduce or inspect behavior when broker publishing fails, then determine how status persistence and re-enqueue should be coordinated. Done means a broker error cannot leave a submission permanently stuck in Scoring.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, rabbitmq
Domain
backend, distributed-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.