coder / coder/internal

flake: test-go-pg (macos) job timeout/cancelled

Open
#1,365 5 comments 0 reactions 1 assignee Claimed by @BobbyHo View on GitHub
flake
Dominant language
No language data
Stars
3
Forks
0
PR merge metrics
No merged PRs in 30d

Description

CI Run Link: https://github.com/coder/coder/actions/runs/22303873776
Failed Job: test-go-pg (macos-latest) https://github.com/coder/coder/actions/runs/22303873776/job/64518157527
Aggregator failure: required job failed because test-go-pg was cancelled

Commit Info:
- SHA: b776a14b46fc14a972d15d59473daa5f150d452c
- Author: Thomas Kosiewski
- Message: fix(coderd): harden OAuth2 provider security (#22194)

Failure evidence:
- required job output:
```
- test-go-pg: cancelled
One of the required checks has failed or has been cancelled
```
- macOS job log (test step):
```
2026-02-23T11:24:37Z ok github.com/coder/coder/v2/cli 100.025s
2026-02-23T11:44:25Z ##[error]The operation was canceled.
```

Timeline / timeout context:
- test-go-pg job has `timeout-minutes: 25` in .github/workflows/ci.yaml.
- Job started at ~11:19:13Z and was cancelled at 11:44:25Z (≈25 minutes).
- No further test output appeared between 11:24:37Z and cancellation, suggesting the `make test` step hung or stalled and hit the job timeout.

Root cause classification: Infrastructure / test suite hang
- The macOS test-go-pg job exceeded the 25-minute job timeout during `make test` (gotestsum over ./cli/... ./enterprise/cli/... ./agent/...).
- No `WARNING: DATA RACE`, panic, or OOM indicators found in the job log; cancellation is consistent with the workflow timeout rather than a specific test failure.

Assignment analysis (component / infra ownership):
- No specific test function identified (gotestsum standard-quiet output stops before hang), so test-level blame isn’t possible.
- Using infra ownership fallback:
- `git log --oneline -10 .github/actions/test-go-pg/action.yaml`
- 7d558e7 fix: make `make test` runnable again (Danny Kopping)
- 84b7a03 chore: abstract pg test logic and double runner sizes (Danny Kopping)
- Assigning to @dannykopping as the most recent meaningful maintainer of the test-go-pg action and parallelism logic.

Related issues search:
- Searched coder/internal for: "test-go-pg macos cancelled", "operation was canceled", "test-go-pg timeout", "job timeout" — no matching open issues found.

Reproduction:
- Re-run the CI workflow on main and watch test-go-pg (macos-latest).
- Locally on macOS: `TEST_PACKAGES="./cli/... ./enterprise/cli/... ./agent/..." make test` (observe for hang or unusually long runtime).

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.