sandboxed-sh runner container down — CI dead again (run 5676 waiting since 2026-08-14T06:00Z) #80

New issue

Closed

opened 2026-08-14 09:31:04 +00:00 by Mark · 0 comments

Mark commented

2026-08-14 09:31:04 +00:00

Owner

Summary

sandboxed-sh has a registered per-repo runner (~/forgejo-runner-sandboxed-sh, .runner present since Aug 12) but its container is NOT running. Newest CI run is stuck waiting; sandboxed-sh CI is dead again since ~06:00Z.

Evidence

Run id=5676 (push, commit 49c9fe61 "fix(build): commit Cargo.lock for reproducible builds (#9)")
- created: 2026-08-14T06:00:02Z
- status: waiting, started=1970-01-01T00:00:00Z (never picked up)
- URL: https://forgejo.tail018ac4.ts.net/Mark/sandboxed-sh/actions/runs/772
Prior runs 5624/5625/5635 (same commit, Aug 13 12:00 / 18:00 / Aug 14 00:00) were picked up and failed in ~40s (run 5635: started 2026-08-14T00:00:02Z, stopped 00:00:40Z) — a runner WAS online then.
On blackhole: ~/forgejo-runner-sandboxed-sh/data/.runner present, but podman ps shows no runner container for sandboxed-sh.

Steps to reproduce

On blackhole: podman ps | grep sandboxed-sh → no container.
Check ~/forgejo-runner-sandboxed-sh/data/.runner → present (registered but not running).
Push to Mark/sandboxed-sh → run created, status waiting, started=1970, never dispatched.

Expected

Per the #78 goal, a live runner for sandboxed-sh claims jobs within seconds of a push.

Actual

Runner registered but container down; run 5676 has been waiting since 2026-08-14T06:00:02Z. CI dead again.

Action needed

Restart the sandboxed-sh runner container (or add supervision so it comes back automatically — see the supervisor issue filed alongside this one).

Filed by reporter from kanban card t_86cbfbe7 (probe of closed issue #78).

## Summary sandboxed-sh has a registered per-repo runner (`~/forgejo-runner-sandboxed-sh`, `.runner` present since Aug 12) but its container is **NOT running**. Newest CI run is stuck `waiting`; sandboxed-sh CI is dead again since ~06:00Z. ## Evidence - Run **id=5676** (push, commit `49c9fe61` "fix(build): commit Cargo.lock for reproducible builds (#9)") - created: 2026-08-14T06:00:02Z - status: `waiting`, started=1970-01-01T00:00:00Z (never picked up) - URL: https://forgejo.tail018ac4.ts.net/Mark/sandboxed-sh/actions/runs/772 - Prior runs 5624/5625/5635 (same commit, Aug 13 12:00 / 18:00 / Aug 14 00:00) were picked up and failed in ~40s (run 5635: started 2026-08-14T00:00:02Z, stopped 00:00:40Z) — a runner WAS online then. - On blackhole: `~/forgejo-runner-sandboxed-sh/data/.runner` present, but `podman ps` shows no runner container for sandboxed-sh. ## Steps to reproduce 1. On blackhole: `podman ps | grep sandboxed-sh` → no container. 2. Check `~/forgejo-runner-sandboxed-sh/data/.runner` → present (registered but not running). 3. Push to Mark/sandboxed-sh → run created, status `waiting`, started=1970, never dispatched. ## Expected Per the #78 goal, a live runner for sandboxed-sh claims jobs within seconds of a push. ## Actual Runner registered but container down; run 5676 has been waiting since 2026-08-14T06:00:02Z. CI dead again. ## Action needed Restart the sandboxed-sh runner container (or add supervision so it comes back automatically — see the supervisor issue filed alongside this one). Filed by reporter from kanban card t_86cbfbe7 (probe of closed issue #78).