coilyco-flight-deck/agentic-os

Watch

Retry dev-base bootstrap downloads after transient network resets #630

New issue

Closed

opened 2026-07-23 08:22:11 +00:00 by coilyco-ops · 4 comments

coilyco-ops commented

2026-07-23 08:22:11 +00:00

Member

workflow: pull-request-and-merge

Current state

AOS #689 removed the core image. Every language image now starts directly from Ubuntu, and full remains the compatibility fan-in. Every published tier invokes actions/publish-dev-base-tier/scripts/install-docker.sh, which still downloads the Docker index, Docker archive, and buildx binary with plain curl -fsSL calls.

PR #633 predates that refactor and conflicts with current main. Its retry outcome remains valid, but the separate core-job bootstrap and the extra shared action are no longer required.

Scope

AOS adds bounded retry and backoff to every network download in the existing shared Docker and buildx installer.
AOS keeps permanent HTTP failures loud through curl fail mode.
AOS preserves the tier checkpoint and resume behavior.
Tests prove the common installer applies the retry contract to the index, archive, and buildx binary.
The updated PR drops obsolete core-job wiring and reconciles with the independent language matrix.

Done

Transient connection resets receive bounded retries.
Permanent download failures remain visible and fail the job.
Every language tier and full use the same corrected installer.
Repository validation passes.

workflow: pull-request-and-merge ## Current state AOS #689 removed the core image. Every language image now starts directly from Ubuntu, and `full` remains the compatibility fan-in. Every published tier invokes `actions/publish-dev-base-tier/scripts/install-docker.sh`, which still downloads the Docker index, Docker archive, and buildx binary with plain `curl -fsSL` calls. PR #633 predates that refactor and conflicts with current main. Its retry outcome remains valid, but the separate core-job bootstrap and the extra shared action are no longer required. ## Scope * AOS adds bounded retry and backoff to every network download in the existing shared Docker and buildx installer. * AOS keeps permanent HTTP failures loud through curl fail mode. * AOS preserves the tier checkpoint and resume behavior. * Tests prove the common installer applies the retry contract to the index, archive, and buildx binary. * The updated PR drops obsolete core-job wiring and reconciles with the independent language matrix. ## Done * Transient connection resets receive bounded retries. * Permanent download failures remain visible and fail the job. * Every language tier and `full` use the same corrected installer. * Repository validation passes.

coilyco-ops referenced this issue from a commit

2026-07-23 08:26:45 +00:00

fix(ci): retry dev-base bootstrap downloads (closes #630)

coilyco-ops referenced this issue from a pull request that will close it,

2026-07-23 08:26:53 +00:00

fix(ci): retry dev-base bootstrap downloads #633

coilyco-ops commented

2026-07-23 08:29:06 +00:00

Author

Member

WARDED_WORKFLOW: #633

details

Review gate: intentionally skipped because the in-container review gate is temporarily disabled pending brokered QA.

workflow: pull-request-and-merge; review summary: intentionally skipped pending brokered QA

Implementation felt direct: one shared bootstrap action removed the duplicated failure surface. Confidence: high. Surprise: none beyond the expected transient-egress failure mode. Follow-up: the director merge lane will trigger the immutable publish path, whose existing checkpoint skips already-landed tiers.

WARDED_WORKFLOW: https://forgejo.coilysiren.me/coilyco-flight-deck/agentic-os/pulls/633 <details><summary>details</summary> Review gate: intentionally skipped because the in-container review gate is temporarily disabled pending brokered QA. workflow: pull-request-and-merge; review summary: intentionally skipped pending brokered QA Implementation felt direct: one shared bootstrap action removed the duplicated failure surface. Confidence: high. Surprise: none beyond the expected transient-egress failure mode. Follow-up: the director merge lane will trigger the immutable publish path, whose existing checkpoint skips already-landed tiers. </details>

coilyco-ops commented

2026-07-23 08:29:16 +00:00

Author

Member

WARDED_WORKFLOW: reservation-released

release details

Run finished with WARDED_WORKFLOW: https://forgejo.coilysiren.me/coilyco-flight-deck/agentic-os/pulls/633.

ward container reap released container engineer-codex-agentic-os-630 (--harness codex): the terminal outcome supersedes the reservation, so a later redispatch no longer needs --override-reservation.

Outcome summary: #633

— Codex, via ward agent

WARDED_WORKFLOW: reservation-released <details><summary>release details</summary> Run finished with `WARDED_WORKFLOW: https://forgejo.coilysiren.me/coilyco-flight-deck/agentic-os/pulls/633`. `ward container reap` released container `engineer-codex-agentic-os-630` (`--harness codex`): the terminal outcome supersedes the reservation, so a later redispatch no longer needs `--override-reservation`. **Outcome summary:** https://forgejo.coilysiren.me/coilyco-flight-deck/agentic-os/pulls/633 </details>  — Codex, via `ward agent`

coilyco-ops referenced this issue

2026-07-24 06:48:14 +00:00

Releases are blocked: a dotnet image-push flake skips the release job, so no tag and no ward-specs bundle (stuck at v0.248.0) #490

coilyco-ops commented

2026-07-24 07:01:15 +00:00

Author

Member

WARD-OUTCOME: done
workflow: pull-request-and-merge

Review summary: PASS

Reviewer - Codex engineer completed a focused correctness, security, regression, and architecture review of PR #633 at 9f9efce.
Findings - The reviewer found no blocking or non-blocking defects.
Architecture - The branch now extends the single tier publisher used by the independent Ubuntu-based image family. The branch removes the stale core-image bootstrap action.
Validation - ward exec test -- tests/test_dev_base_publish_workflows.py passed 6 tests. The changed installer passed ward exec pre-commit-shellcheck-files -- actions/publish-dev-base-tier/scripts/install-docker.sh.
Catalog - Every repository contract reached by ward exec pre-commit-all passed. The all-files ShellCheck gate remains blocked by pre-existing CRLF content in .agents/skills/tooling-image-zoom/scripts/crop-image, which this branch does not change.

WARD-OUTCOME: done workflow: pull-request-and-merge Review summary: PASS * Reviewer - Codex engineer completed a focused correctness, security, regression, and architecture review of PR #633 at `9f9efce`. * Findings - The reviewer found no blocking or non-blocking defects. * Architecture - The branch now extends the single tier publisher used by the independent Ubuntu-based image family. The branch removes the stale core-image bootstrap action. * Validation - `ward exec test -- tests/test_dev_base_publish_workflows.py` passed 6 tests. The changed installer passed `ward exec pre-commit-shellcheck-files -- actions/publish-dev-base-tier/scripts/install-docker.sh`. * Catalog - Every repository contract reached by `ward exec pre-commit-all` passed. The all-files ShellCheck gate remains blocked by pre-existing CRLF content in `.agents/skills/tooling-image-zoom/scripts/crop-image`, which this branch does not change.

coilyco-ops commented

2026-07-24 07:58:33 +00:00

Author

Member

WARD-OUTCOME: done
workflow: pull-request-and-merge

verdict: pass
confidence: high
summary: PR #633 at 7005db5 cleanly adds bounded, atomic retries to the Docker and buildx downloads in the current single full-image publisher. The final diff contains only the shared installer and its topology-current regression test.
findings: []
conclusion: The reviewer passed PR #633 after refuting partial-download, permanent-failure, stale-topology, and current-main compatibility risks. Five focused workflow tests and focused ShellCheck passed.

WARD-OUTCOME: done workflow: pull-request-and-merge verdict: pass confidence: high summary: PR #633 at `7005db5` cleanly adds bounded, atomic retries to the Docker and buildx downloads in the current single full-image publisher. The final diff contains only the shared installer and its topology-current regression test. findings: [] conclusion: The reviewer passed PR #633 after refuting partial-download, permanent-failure, stale-topology, and current-main compatibility risks. Five focused workflow tests and focused ShellCheck passed.

coilysiren closed this issue

2026-07-25 13:29:30 +00:00

coilysiren referenced this issue from a commit

2026-07-25 13:29:31 +00:00

fix(ci): retry dev-base bootstrap downloads (#633)

No milestone

No project

No assignees

1 participant

Notifications

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/agentic-os#630

No description provided.

Rows
Columns