Record successful forge runner smoke
All checks were successful
Forge Runner Smoke / compatibility-smoke (push) Successful in 0s

This commit is contained in:
2026-06-08 09:36:54 +02:00
parent 367f777d13
commit a43a66853d

View File

@@ -1,10 +1,11 @@
# Gitea Actions Runner Evidence
Last updated: 2026-06-07
Last updated: 2026-06-08
Status: initial non-secret inventory for `FORGE-WP-0003`. Haskelseed now has a
running `act_runner` OpenRC service with labels matching the first compatibility
contract. Smoke proof is still pending.
Status: haskelseed is registered as a `coulomb` organization Gitea Actions
runner and the forge smoke workflow has passed. Inter-hub now fails after runner
scheduling, so the remaining blocker is workload-specific rather than runner
visibility.
## Workstream
@@ -80,8 +81,9 @@ Observed on 2026-06-07:
| registration labels after activation | `self-hosted:host`, `haskelseed:host`, `linux:host`, `linux_amd64:host`, `x86_64:host`, `container-build:host`, `registry-publish:host` |
| ephemeral | `false` |
| runner backup | `/root/.runner.bak-20260607225905` |
| smoke workflow run | `forge-runner-smoke.yaml #1`, run `/coulomb/railiance-forge/actions/runs/1`, commit `19ee47fe82` |
| smoke workflow status | `Waiting`, duration `0s` as of 2026-06-08 00:40 Europe/Berlin |
| org registration | re-registered under the `coulomb` organization on 2026-06-08 using an attended token from the Gitea UI; token was not recorded |
| smoke workflow run | `forge-runner-smoke.yaml #4`, run `/coulomb/railiance-forge/actions/runs/4`, commit `de6178764c` |
| smoke workflow status | `Success` as of 2026-06-08 |
Activation evidence:
@@ -162,31 +164,30 @@ Haskelseed inventory on 2026-06-08:
This means the blocked inter-hub workflow did not publish the expected image
tags to the current Gitea registry path.
## Smoke Evidence Slot
## Smoke Evidence
No smoke job has passed yet. Current haskelseed registration labels now match
the proposed first compatibility contract, and Gitea created run
`forge-runner-smoke.yaml #1`, but the run is still `Waiting` at `0s`.
The next gate needs authenticated Gitea Actions UI/API inspection to determine
whether the waiting run needs approval, rerun, or runner assignment repair.
Expected evidence after `.gitea/workflows/forge-runner-smoke.yaml` passes:
The smoke workflow passed after haskelseed was re-registered under the `coulomb`
organization runner scope. The earlier waiting runs were caused by Gitea not
having an eligible org/repo runner record even though a local haskelseed
registration file existed.
| Field | Value |
| --- | --- |
| Date | `TODO` |
| Runner name | `TODO` |
| Labels used | `TODO` |
| Workflow id/url | `TODO` |
| Repo commit | `TODO` |
| Docker availability | `TODO` |
| Cluster credential absent | `TODO` |
| Registry credential absent in smoke | `TODO` |
| Date | 2026-06-08 |
| Runner name | `haskelseed` |
| Labels used | `self-hosted`, `haskelseed` |
| Workflow id/url | `/coulomb/railiance-forge/actions/runs/4` |
| Repo commit | `de6178764c` |
| Docker availability | smoke step completes; public run status is `Success` |
| Cluster credential absent | smoke workflow asserts `KUBECONFIG` is unset |
| Registry credential absent in smoke | smoke workflow asserts `REGISTRY_PASSWORD` and `GITEA_RUNNER_REGISTRATION_TOKEN` are unset |
## Inter-Hub Unblock Slot
Do not rerun production push probes until the smoke workflow passes.
The runner substrate is no longer the primary blocker: the forge smoke workflow
passes and Gitea can schedule jobs on haskelseed. The first inter-hub job picked
up after org registration failed, according to the authenticated UI observation
reported by the operator on 2026-06-08.
Expected evidence after inter-hub rerun:
@@ -195,6 +196,6 @@ Expected evidence after inter-hub rerun:
| Inter-hub commit | `TODO` |
| Workflow id/url | `TODO` |
| Image tag/digest | `TODO` |
| Deployment result | `TODO` |
| Deployment result | failed after scheduling to the runner |
| `https://hub.coulomb.social/api/v2/hubs` | `TODO` |
| Remaining blocker | `TODO` |
| Remaining blocker | inspect inter-hub job logs for the workload-specific failure |