Repair ward:release publish after promote loses Docker socket #1455

Closed
opened 2026-07-15 13:44:09 +00:00 by coilyco-ops · 4 comments
Owner

Merged PR #1450 for #1447 still does not satisfy the default-image acceptance: the post-merge promote path cannot publish forgejo.coilysiren.me/coilyco-flight-deck/ward:release from inside the current job container.

Evidence from QA/ops on 2026-07-15:

  • Source repair issue: #1447, labels headless, P2.
  • Merged repair PR: #1450 at head 280a06443e1fc36628ff984d8bf596c7a19c4df7.
  • Later main included #1450, #1451, and #1448 at a19f8fdd586e610cf80e41badc08535bc86d3eaf.
  • Promote run 8610 / UI run #2303 failed in Publish the ward agent image alias.
  • The raw job log shows the gate, tests, lint, and draft artifact publish succeeded, then the image alias step did:
Login Succeeded
Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
  • docker manifest inspect forgejo.coilysiren.me/coilyco-flight-deck/ward:release still returns manifest unknown.
  • Because the alias is missing, release.yml has not verified the configured default agent image and the current default remains unsafe for fresh warded launches.

Repair direction:

  • Make promote publish or otherwise create forgejo.coilysiren.me/coilyco-flight-deck/ward:release without assuming a Docker daemon is available inside the job container, or explicitly move that step to a runner context where the daemon socket is available.
  • Keep the release workflow check that fails if the alias is missing.
  • Rerun or retrigger the main promote path and confirm ward:release resolves before the release branch publish.
  • Confirm the subsequent release.yml run passes through the alias verification.
  • Preserve #1139/#1447 acceptance: the default Ward agent image must resolve, and release/promote must not fetch coilysiren/agentic-os actions from GitHub.

Related: #1139, #1446, #1447, #1450, run 8610 / #2303.

Merged PR #1450 for #1447 still does not satisfy the default-image acceptance: the post-merge promote path cannot publish `forgejo.coilysiren.me/coilyco-flight-deck/ward:release` from inside the current job container. Evidence from QA/ops on 2026-07-15: - Source repair issue: #1447, labels `headless`, `P2`. - Merged repair PR: #1450 at head `280a06443e1fc36628ff984d8bf596c7a19c4df7`. - Later main included #1450, #1451, and #1448 at `a19f8fdd586e610cf80e41badc08535bc86d3eaf`. - Promote run 8610 / UI run #2303 failed in `Publish the ward agent image alias`. - The raw job log shows the gate, tests, lint, and draft artifact publish succeeded, then the image alias step did: ```text Login Succeeded Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running? ``` - `docker manifest inspect forgejo.coilysiren.me/coilyco-flight-deck/ward:release` still returns `manifest unknown`. - Because the alias is missing, `release.yml` has not verified the configured default agent image and the current default remains unsafe for fresh warded launches. Repair direction: - Make promote publish or otherwise create `forgejo.coilysiren.me/coilyco-flight-deck/ward:release` without assuming a Docker daemon is available inside the job container, or explicitly move that step to a runner context where the daemon socket is available. - Keep the release workflow check that fails if the alias is missing. - Rerun or retrigger the main promote path and confirm `ward:release` resolves before the release branch publish. - Confirm the subsequent `release.yml` run passes through the alias verification. - Preserve #1139/#1447 acceptance: the default Ward agent image must resolve, and release/promote must not fetch `coilysiren/agentic-os` actions from GitHub. Related: #1139, #1446, #1447, #1450, run 8610 / #2303.
Author
Owner

DISPATCH-DEFERRED: QA/ops tried to dispatch this repair immediately, but the Ward repo engineer pool is at capacity.

Evidence:

warded coilyco-flight-deck/ward#1455 --harness codex
ward: dispatch broker: ward agent engineer --harness codex: repo engineer limit is reached for coilyco-flight-deck/ward: 3 active engineer(s) (limit 3); pass --override-capacity to exceed real running capacity (dispatch log: /Users/kai/.ward/agent-logs/dispatch/20260715T134414Z-director-codex-jp85-coilyco-flight-deck-ward-1455.log)

This issue remains ready for sibling dispatch as soon as capacity opens.

DISPATCH-DEFERRED: QA/ops tried to dispatch this repair immediately, but the Ward repo engineer pool is at capacity. Evidence: ```text warded coilyco-flight-deck/ward#1455 --harness codex ward: dispatch broker: ward agent engineer --harness codex: repo engineer limit is reached for coilyco-flight-deck/ward: 3 active engineer(s) (limit 3); pass --override-capacity to exceed real running capacity (dispatch log: /Users/kai/.ward/agent-logs/dispatch/20260715T134414Z-director-codex-jp85-coilyco-flight-deck-ward-1455.log) ``` This issue remains ready for sibling dispatch as soon as capacity opens.
Author
Owner

WARDED_WORKFLOW: blocked 🛑

details

Review gate intentionally skipped because the temporary Ward default is pending brokered QA.

workflow: pull-request-and-merge; review summary: intentionally skipped because the temporary Ward default is pending brokered QA.

PR #1514 merged after its retry check passed. Local ward exec test, ward exec vet, ward exec lint, ward exec lint-workflows, and pre-commit passed. The daemon-free manifest verifier feels appropriately small and direct; confidence in the repair is high.

Blocker: main promote runs 2368 and 2369 both fail before alias publication at the unrelated #1516 test TestRunHostDispatchBrokerRequestReportsLaterLaunchFailureThroughArtifact. That test reads the dispatch artifact while the detached worker still reports outcome: in-progress; this is reproducible on the manual rerun. Repair that broker-test race, then rerun promote so ward:release can publish and release.yml can verify it.

— Codex, via ward agent

<!-- ward-agent-signature --> WARDED_WORKFLOW: blocked 🛑 <details><summary>details</summary> Review gate intentionally skipped because the temporary Ward default is pending brokered QA. workflow: pull-request-and-merge; review summary: intentionally skipped because the temporary Ward default is pending brokered QA. PR #1514 merged after its retry check passed. Local `ward exec test`, `ward exec vet`, `ward exec lint`, `ward exec lint-workflows`, and pre-commit passed. The daemon-free manifest verifier feels appropriately small and direct; confidence in the repair is high. Blocker: main promote runs 2368 and 2369 both fail before alias publication at the unrelated #1516 test `TestRunHostDispatchBrokerRequestReportsLaterLaunchFailureThroughArtifact`. That test reads the dispatch artifact while the detached worker still reports `outcome: in-progress`; this is reproducible on the manual rerun. Repair that broker-test race, then rerun promote so `ward:release` can publish and `release.yml` can verify it. </details> — Codex, via `ward agent`
Author
Owner

WARDED_WORKFLOW: reservation-released

release details

Run finished with WARDED_WORKFLOW: blocked 🛑.

ward container reap released container engineer-codex-ward-1455 (--harness codex): the terminal outcome supersedes the reservation, so a later redispatch no longer needs --override-reservation.

— Codex, via ward agent

<!-- ward-agent-reservation-released --> WARDED_WORKFLOW: reservation-released <details><summary>release details</summary> Run finished with `WARDED_WORKFLOW: blocked 🛑`. `ward container reap` released container `engineer-codex-ward-1455` (`--harness codex`): the terminal outcome supersedes the reservation, so a later redispatch no longer needs `--override-reservation`. </details> <!-- ward-agent-signature --> — Codex, via `ward agent`
Author
Owner

WARDED_WORKFLOW: reservation-released

release details

Run never started. ward container reap released container engineer-codex-ward-1455 (--harness codex): it exited without launching the agent (smoke-test death, ward#222/#264/#595), so it did no work and the launch intent it took is retracted. Nothing is running on this issue. It needs re-dispatch. A ward agent director re-queues it automatically. A manual ward agent retry no longer needs --override-reservation.

— Codex, via ward agent

<!-- ward-agent-reservation-released --> <!-- ward-needs-redispatch --> WARDED_WORKFLOW: reservation-released <details><summary>release details</summary> Run never started. `ward container reap` released container `engineer-codex-ward-1455` (`--harness codex`): it exited without launching the agent (smoke-test death, ward#222/#264/#595), so it did no work and the launch intent it took is retracted. Nothing is running on this issue. It needs re-dispatch. A `ward agent director` re-queues it automatically. A manual `ward agent` retry no longer needs `--override-reservation`. </details> <!-- ward-agent-signature --> — Codex, via `ward agent`
Commenting is not possible because the repository is archived.
No milestone
No project
No assignees
1 participant
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
coilyco-flight-deck/ward#1455
No description provided.