Repair the isolated dev-base core image build failure #591

New issue

Closed

opened 2026-07-15 13:02:33 +00:00 by coilyco-ops · 3 comments

coilyco-ops commented

2026-07-15 13:02:33 +00:00

Member

Problem

PR #590 merged for #589 and successfully split publish-core into director-visible subphases, but post-merge dev-base publication run 8557 / repository run #1439 still failed and agentic-os-full:latest remains absent.

The new run page now isolates the failure precisely:

plan-draft: success
publish-core:
  Acquire core publish source: success
  Install uv: success
  Resolve docker host: success
  Derive core publish inputs: success
  Registry login: success
  Check whether the tier already landed: success
  Build core image: failure after 15m19s
  Verify published core image: failure
  Probe core buildcache write: failure

The expanded Forgejo step-log request returns cursor 0 with no lines for the failed build step, so the director can identify the failing phase but not the build engine exit cause. This is the focused continuation of #583 after #589 satisfied its visibility acceptance.

Acceptance

Inspect or reproduce the exact Build core image invocation from run 8557 / #1439 and identify its exit cause.
If repository-owned, land a focused fix while preserving bounded source acquisition, retry/resume, named workflow steps, and registry verification.
If external BuildKit, registry, storage, or runner state is the wall, expose the exact condition and required operator action on a director-readable surface.
PR CI is green.
A new dev-base-publish.yml run passes publish-core, then continues to publish agentic-os-full:latest, or leaves an exact external blocker.

Related: #583, #589, #590, run 8557 / #1439.

## Problem PR #590 merged for #589 and successfully split `publish-core` into director-visible subphases, but post-merge dev-base publication run 8557 / repository run #1439 still failed and `agentic-os-full:latest` remains absent. The new run page now isolates the failure precisely: ```text plan-draft: success publish-core: Acquire core publish source: success Install uv: success Resolve docker host: success Derive core publish inputs: success Registry login: success Check whether the tier already landed: success Build core image: failure after 15m19s Verify published core image: failure Probe core buildcache write: failure ``` The expanded Forgejo step-log request returns cursor 0 with no lines for the failed build step, so the director can identify the failing phase but not the build engine exit cause. This is the focused continuation of #583 after #589 satisfied its visibility acceptance. ## Acceptance * Inspect or reproduce the exact `Build core image` invocation from run 8557 / #1439 and identify its exit cause. * If repository-owned, land a focused fix while preserving bounded source acquisition, retry/resume, named workflow steps, and registry verification. * If external BuildKit, registry, storage, or runner state is the wall, expose the exact condition and required operator action on a director-readable surface. * PR CI is green. * A new `dev-base-publish.yml` run passes `publish-core`, then continues to publish `agentic-os-full:latest`, or leaves an exact external blocker. Related: #583, #589, #590, run 8557 / #1439.

coilyco-ops added the

headless

label

2026-07-15 13:02:33 +00:00

coilyco-ops commented

2026-07-15 13:02:50 +00:00

Author

Member

WARDED_WORKFLOW: reservation-held

reservation details

Holder: launch intent for container engineer-codex-agentic-os-591 on host 74a70515a491.

Accepted by ward agent --harness codex (reserved 2026-07-15T13:02:49Z). Concurrent ward agent runs are blocked until this intent becomes visible or the intent is released. The stale-intent fallback is still TTL-bounded (3h TTL). --override-reservation overrides.

Do not comment on or edit this issue to steer the run while it is reserved. The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a new issue, dispatched fresh. That is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).

run seed context — what this run is carrying (ward#609)

Resolved: coilyco-flight-deck/agentic-os#591 · branch issue-591 · harness codex · workflow pull-request-and-merge
Run: engineer-codex-agentic-os-591 · ward v0.710.0 · dispatched 2026-07-15T13:02:45Z
Comment thread: 0 included in the pre-flight read, 0 stripped (ward's own automated comments).

Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.710.0).

— Codex, via ward agent

WARDED_WORKFLOW: reservation-held <details><summary>reservation details</summary> Holder: launch intent for container `engineer-codex-agentic-os-591` on host `74a70515a491`. Accepted by `ward agent --harness codex` (reserved 2026-07-15T13:02:49Z). Concurrent `ward agent` runs are blocked until this intent becomes visible or the intent is released. The stale-intent fallback is still TTL-bounded (3h TTL). `--override-reservation` overrides. **Do not comment on or edit this issue to steer the run while it is reserved.** The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a **new issue, dispatched fresh**. That is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494). <details><summary>run seed context — what this run is carrying (ward#609)</summary> - **Resolved:** `coilyco-flight-deck/agentic-os#591` · branch `issue-591` · harness `codex` · workflow `pull-request-and-merge` - **Run:** `engineer-codex-agentic-os-591` · ward `v0.710.0` · dispatched `2026-07-15T13:02:45Z` - **Comment thread:** 0 included in the pre-flight read, 0 stripped (ward's own automated comments). Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.710.0). </details> </details>  — Codex, via `ward agent`

coilyco-ops commented

2026-07-15 13:03:18 +00:00

Author

Member

WARDED_WORKFLOW: dispatch-deferred

launch failure details

The sibling run never started. Ward v0.710 acquired the reservation and created engineer-codex-agentic-os-591 using agentic-os:release, then failed while seeding the container: docker cp to /opt/agentic-os/ward-shell-entrypoint.sh returned not a directory. The created container was stopped. No engineer code ran and no branch or PR was created.

This is the known stale director/image seeding incompatibility recorded on #583. The reservation is released by this explicit failed-prelaunch record for redispatch from a compatible director. #591 remains the precise repair for the Build core image failure isolated in run 8557 / #1439.

Codex, via ward agent

WARDED_WORKFLOW: dispatch-deferred <details><summary>launch failure details</summary> The sibling run never started. Ward v0.710 acquired the reservation and created engineer-codex-agentic-os-591 using agentic-os:release, then failed while seeding the container: docker cp to /opt/agentic-os/ward-shell-entrypoint.sh returned not a directory. The created container was stopped. No engineer code ran and no branch or PR was created. This is the known stale director/image seeding incompatibility recorded on #583. The reservation is released by this explicit failed-prelaunch record for redispatch from a compatible director. #591 remains the precise repair for the Build core image failure isolated in run 8557 / #1439. </details>  - Codex, via ward agent

coilyco-ops commented

2026-07-15 13:16:27 +00:00

Author

Member

Superseded by #592 and merged PR #593. The healthy engineer isolated the root cause to oversized tier build contexts and landed fix(dev-base): shrink tier build contexts with both required PR checks green. The next dev-base publication is the canonical validation; a second branch from #591 would duplicate the completed repair.

Superseded by #592 and merged PR #593. The healthy engineer isolated the root cause to oversized tier build contexts and landed `fix(dev-base): shrink tier build contexts` with both required PR checks green. The next dev-base publication is the canonical validation; a second branch from #591 would duplicate the completed repair.

coilyco-ops closed this issue

2026-07-15 13:16:41 +00:00