tests: spread tests fail intermittently
Metadata
Current evaluation
Spread tests flake when the docker snap daemon never starts. Still present in-tree: spread.yaml does 'snap install docker' then a bare 'retry -n 10 --wait 2 docker run --rm hello-world' with no failure diagnostics. Fix proposed (log capture, longer retry) not yet implemented.
Suggested action: keep open
Reason: Confirmed unfixed in-tree: spread.yaml:150/161 still runs 'snap install docker' followed by 'retry -n 10 --wait 2 sh -c "docker run --rm hello-world"' with no diagnostics on failure, and a repo-wide grep for 'snap logs docker' returns zero matches, proving the PR #96 mitigation was never merged. The maintainer triage comment explicitly states 'this is still present in-tree (not obsolete)' and the Jira sync (ROCKCRAFT-296) shows it is tracked work, so it must stay open. Impact is moderate (CI flake wasting runs, not user-facing), and the fix is a small, well-understood diagnostics/retry tweak; actionability is slightly reduced because the original log is from the retired Google backend and ubuntu-18.04, so the exact original trigger cannot be reproduced against today's openstack config.
Evaluated by: mr-cal-bot
Issue body
Evaluation history
| Date | Model | Scores | Action | Summary |
|---|---|---|---|---|
| stealth/space-bunny-alpha |
Impact:
42
Quick Win:
22
Complexity:
22
Confidence:
82
Actionability:
68
|
keep open | Spread tests flake when the docker snap daemon never starts. Still present in-tree: spread.yaml does 'snap install docker' then a bare 'retry -n 10 --wait 2 docker run --rm hello-world' with no failure diagnostics. Fix proposed (log capture, longer retry) not yet implemented. | |
| qwen/qwen3.6-35b-a3b |
Impact:
30
Quick Win:
13
Complexity:
40
Confidence:
75
Actionability:
70
|
keep open | Intermittent spread test failures due to Docker daemon not starting after installation. Triaged, still present in current spread.yaml with unchanged retry logic. | |
| qwen/qwen3.8-27b |
Impact:
35
Quick Win:
21
Staleness:
50
Complexity:
40
Confidence:
85
Support Request:
5
|
keep open | Rockcraft spread tests intermittently fail because the docker snap daemon never starts in test VMs, and the retry loop is insufficient. Feb 2026 triage confirmed the same pattern is still in spread.yaml; a fix (diagnostics on failure) is suggested but no PR proposed. | |
| qwen/qwen3.6-35b-a3b |
Staleness:
55
Complexity:
30
Confidence:
85
Support Request:
10
|
keep open | Intermittent spread test failures due to Docker daemon startup timeout. Triaged bug with recent maintainer suggestions for improved diagnostics, awaiting implementation. | |
| qwen3.6-35b-a3b-mtp-q6 |
Staleness:
55
Complexity:
40
Confidence:
75
Support Request:
10
|
keep open | Intermittent spread test failures due to Docker daemon not starting during CI setup. Triaged and open, with recent suggestions to add diagnostic logging and increase retry timeouts. |
Update history
No update history recorded yet.
Related issues
| Issue | Project | State | Summary | Similarity |
|---|---|---|---|---|
| #534 test(spread): workaround for flaky test | fetch-service | merged | Merged PR #534 (commit 4f89bb7) worked around #533 by skipping the flaky snap names-query assertion while retaining the other spread-test checks. |