bug(store): a headless worker that dies before RUNNING leaves the session PENDING forever; its stderr is lost
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 75/100
- Issue type
- Bug
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- python
- Domain
- backend, observability
Research direction
Start with _reap_dead in tether/store.py and trace how worker processes and session status are recorded. Check the existing tests for reaping dead workers, then add coverage for a PENDING session whose stored pid has exited and for preserving worker stderr in a log file. Done means dead PENDING sessions become FAILED with a useful error pointing to the captured stderr.
Written by the indexing model from the issue text.
Description
What happened
tether run --headless (model claude-softmedium/opus) stored pid 3540688 and status PENDING. The worker process died without ever setting RUNNING. tether wait --timeout 3600 blocked the full hour; the session is still PENDING with 0 steps (session ses_e6c13419). The worker's stdout/stderr go nowhere, so the cause cannot be found.
Expected
_reap_deadalso reaps PENDING sessions whose stored pid is gone (FAILED,Worker died (pid N) before starting).- The worker's stderr goes to a file (e.g.
.tether/workers/<id>.log), and the FAILED result points to it.
Notes
tether/store.py _reap_dead only checks rec.status == "RUNNING".
Found in the v0.2.18 real-world test (tether + Haiku / qwen3.5:9b as gru minions building 11 features; log .claude/gru/active/real-world-v0218.md).
north star: 45/68 (v0.2.18) -> not measured; reliability floor (#171)
- Dominant language
- Python
- Stars
- 3
- Forks
- 1
- Avg merge
- 52m
- Merged PRs (30d)
- 166
Getting set up
- No Dockerfile or Docker Compose file
- No pull request template
- Read the contributing 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.
More from igormcsouza/tether
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
igormcsouza/tether#278 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 64/100
igormcsouza/tether#269 ·
Maintainers usually reply within 1 day
-
bug
Difficulty 4/5 3-5 days Newbie friendliness 52/100
igormcsouza/tether#280 ·
Maintainers usually reply within 1 day
-
bug
Difficulty 4/5 3-5 days Newbie friendliness 49/100
igormcsouza/tether#279 ·
Maintainers usually reply within 1 day
-
Difficulty 3/5 1-2 days Newbie friendliness 65/100
igormcsouza/tether#273 ·
Maintainers usually reply within 1 day
All issues in igormcsouza/tether
Similar issues
-
Difficulty 1/5 Under an hour Newbie friendliness 91/100
aws-samples/sample-per-model-bedrock#803 ·
Maintainers usually reply within 1 day
-
documentation p1
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
huggingface/accelerate#4511 ·
Maintainers usually reply within 1 day
-
follow-up
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
lesson-submission pending-review
Difficulty 1/5 Under an hour Newbie friendliness 72/100
Ikalus1988/MisakaNet#4585 ·
Maintainers usually reply within 1 day