codalab / codalab/codabench

MOT17 / MOT20 submissions stuck in Running for ~1 hour, then fail with "Child task failed or non-zero return code" and no logs

Open
#2,548 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

Description

Submissions to the MOT17 and MOT20 benchmarks stay in the Running state for nearly 1 hour, and then fail with:

Submission failed: Child task failed or non-zero return code. See logs for more details.

However, no logs are available on the submission page (scoring/ingestion stdout and stderr are all empty), so it is impossible to tell whether the failure comes from the submission content or from the compute worker.

Details

  • Benchmarks affected: MOT17, MOT20
  • Submission format: ZIP with the result .txt files directly in the root directory, standard MOTChallenge 10-column format (frame,id,x,y,w,h,conf,-1,-1,-1)
  • Local checks on the MOT17 submission: all 42 sequence files are present, no NaN/Inf, no duplicate IDs within a frame, no non-positive width/height, frame indices within the sequence length
  • Behavior: status stays Running for ~1 hour → Failed with the message above → no logs

Expected behavior

  • The submission is scored, or
  • If it fails, the logs show the actual error so participants can fix their submission

Related

Possibly related to #2384 (MOT20 submission failed: child task failed) and #2465 (task repeatedly fails with no logs).

Could you please check the compute worker / backend logs for these benchmarks? Thank you!

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

Start by reproducing the MOT17 or MOT20 submission failure and inspect the compute worker/backend logs for the scoring task. Done means the submission either completes successfully or exposes the actual failure in the submission-page logs, including relevant stdout and stderr.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
40/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.