Forgejo Actions CI never runs for planish — no runner covers the repo (repo_id=100) #16

New issue

Closed

opened 2026-08-14 06:19:55 +00:00 by Mark · 1 comment

Mark commented

2026-08-14 06:19:55 +00:00

Owner

Source: Kanban QA card t_24babe4e (post-deploy probe for AuWoMe/planish#15)

Summary

No Forgejo Actions runner is registered for AuWoMe/planish (repo_id=100), so every CI run sits waiting until it is cancelled. Planish has never had a successful CI run.

Actual behavior

Latest push run 5659 (head=874bb24d) stuck status=waiting, started=0 since 2026-08-14 03:27Z.
The #15 fix's run 5631 (head=c058cb87) was created 2026-08-13 23:21Z, never started, cancelled 23:47Z.
All 17 visible runs: failures (May), cancelled/waiting (Aug). Zero successful runs.
Forgejo DB action_runner shows no runner with repo_id=100. Online runners are repo-scoped to other repos (nodedesk -> repo 60, kinapp -> repo 160, mcheroes -> 118, sandboxed-sh -> 35). The user-level runners with the docker label (ids 3,4,11,12,15) belong to zsus, which has been offline on the tailnet for 13+ days.

Expected behavior

Pushes to main should trigger a CI run that is picked up by a runner covering this repo. Register a repo-scoped runner (or bring a user-level runner online) with a label the workflow's runs-on matches (e.g. linux), then confirm a push dispatches within seconds.

Reproduction steps

GET /api/v1/repos/AuWoMe/planish/actions/runs?limit=10 — observe runs stuck waiting/cancelled, never started.
Forgejo DB: SELECT id, repo_id, last_online FROM action_runner WHERE repo_id=100; — returns no rows.
Push a commit to main; watch the new run stay waiting indefinitely.

Impact

The #15 fix (and every other change) ships without CI verification; regressions are caught only by local runs.

Ref: kanban card t_24babe4e.

**Source:** Kanban QA card `t_24babe4e` (post-deploy probe for AuWoMe/planish#15) ## Summary No Forgejo Actions runner is registered for AuWoMe/planish (repo_id=100), so every CI run sits `waiting` until it is cancelled. Planish has never had a successful CI run. ## Actual behavior - Latest push run 5659 (head=874bb24d) stuck `status=waiting`, `started=0` since 2026-08-14 03:27Z. - The #15 fix's run 5631 (head=c058cb87) was created 2026-08-13 23:21Z, never started, cancelled 23:47Z. - All 17 visible runs: failures (May), cancelled/waiting (Aug). Zero successful runs. - Forgejo DB `action_runner` shows no runner with `repo_id=100`. Online runners are repo-scoped to other repos (nodedesk -> repo 60, kinapp -> repo 160, mcheroes -> 118, sandboxed-sh -> 35). The user-level runners with the `docker` label (ids 3,4,11,12,15) belong to zsus, which has been offline on the tailnet for 13+ days. ## Expected behavior Pushes to `main` should trigger a CI run that is picked up by a runner covering this repo. Register a repo-scoped runner (or bring a user-level runner online) with a label the workflow's `runs-on` matches (e.g. `linux`), then confirm a push dispatches within seconds. ## Reproduction steps 1. `GET /api/v1/repos/AuWoMe/planish/actions/runs?limit=10` — observe runs stuck `waiting`/`cancelled`, never started. 2. Forgejo DB: `SELECT id, repo_id, last_online FROM action_runner WHERE repo_id=100;` — returns no rows. 3. Push a commit to `main`; watch the new run stay `waiting` indefinitely. ## Impact The #15 fix (and every other change) ships without CI verification; regressions are caught only by local runs. Ref: kanban card `t_24babe4e`.

Mark added the

bug

label

2026-08-14 06:19:55 +00:00

Mark added the

triage-approved

label

2026-09-15 15:36:49 +00:00

Mark commented

2026-09-18 23:23:13 +00:00

Author

Owner

Stale - closing: runner coverage is working

Filed 2026-08-14 ("no runner covers the repo"). Re-verified live on 2026-09-18:

Runner localhost-planish-runner (id=35) is registered, labeled ubuntu-latest:docker://node:20-bookworm, container forgejo-runner-planish Up 3 days, image data.forgejo.org/forgejo/runner:6.2.2.
.forgejo/workflows/ci.yml runs-on ubuntu-latest - label match correct.
Planish CI has run 45+ times since the issue was filed. Most recent on remote:
- run 6887 (run_number=46, commit bae00a12, "fix(ts): bump Stripe apiVersion ...") -> success (2026-09-14T23:13:41Z).
- run 6886 (run_number=45, commit 02758362, "fix(ci): pin esbuild ...") -> ran 2026-09-14T23:08:55Z (failure content-side, not runner-coverage).
- run 6834 (run_number=42, commit 91afb5c, current main, "docs(changelog) ...") -> ran 2026-09-14T15:55:08Z.
- run 7107 (run_number=47, commit b7216bd6, "docs(readme): correct false NodeHub deploy-on-push claim (fixes #17)") -> status=waiting as of 2026-09-18T23:20:58Z (queued for the runner).
Failures are content-side (workflow step failures) - not "runner doesn't cover this repo". The runner is unambiguously picking up planish jobs.

Same disposition as Mark/chainbound#59 (closed earlier today, same systemic-backlog wave).

Closed by triage-runner via kanban fixit lane.

## Stale - closing: runner coverage is working Filed 2026-08-14 ("no runner covers the repo"). Re-verified live on 2026-09-18: - Runner `localhost-planish-runner` (id=35) is registered, labeled `ubuntu-latest:docker://node:20-bookworm`, container `forgejo-runner-planish` Up 3 days, image `data.forgejo.org/forgejo/runner:6.2.2`. - `.forgejo/workflows/ci.yml` runs-on `ubuntu-latest` - label match correct. - Planish CI has run 45+ times since the issue was filed. Most recent on remote: - **run 6887** (run_number=46, commit `bae00a12`, "fix(ts): bump Stripe apiVersion ...") -> **success** (2026-09-14T23:13:41Z). - **run 6886** (run_number=45, commit `02758362`, "fix(ci): pin esbuild ...") -> ran 2026-09-14T23:08:55Z (failure content-side, not runner-coverage). - **run 6834** (run_number=42, commit `91afb5c`, current main, "docs(changelog) ...") -> ran 2026-09-14T15:55:08Z. - **run 7107** (run_number=47, commit `b7216bd6`, "docs(readme): correct false NodeHub deploy-on-push claim (fixes #17)") -> **status=waiting** as of 2026-09-18T23:20:58Z (queued for the runner). - Failures are content-side (workflow step failures) - not "runner doesn't cover this repo". The runner is unambiguously picking up planish jobs. Same disposition as `Mark/chainbound#59` (closed earlier today, same systemic-backlog wave). Closed by triage-runner via kanban fixit lane.

Mark closed this issue

2026-09-18 23:23:13 +00:00