AGENTS.md post-push step references retired systemctl service, should be k3s rollout #16

Open
opened 2026-05-23 20:55:23 +00:00 by coilysiren · 0 comments
Owner

Originally filed by @coilysiren on 2026-05-20T07:25:48Z - https://github.com/coilysiren/repo-recall/issues/221

Problem

AGENTS.md line 49 (Release + post-push section):

Post-push: verify CI at +300s; brew outdated -> brew upgrade; coily ssh systemctl start repo-recall-update.service. Skip for docs-only.

That post-push step is stale. repo-recall is Kubernetes-only on kai-server now (infrastructure/deploy/repo-recall.yml: three containers per pod, tailscale sidecar, tailnet-only). There is no repo-recall-update.service to systemctl-start.

Fix

Rewrite the post-push step to:

  • For Mac (local brew install): brew outdated -> brew upgrade coilysiren/repo-recall/repo-recall -> brew services restart repo-recall. (Mac is the staging-shape consumer of the new release, ref coilysiren/session-lattice#21 cabling.)
  • For kai-server (prod): the deploy lands via the container image build (docker.yml) and a k3s rollout. Document the actual rollout command (coily ops kubectl rollout restart deployment/repo-recall -n repo-recall or whatever the canonical wrapper is). Cross-link infrastructure/deploy/repo-recall.yml so the next reader knows where the manifest lives.

Why it matters

The agent will literally try to run the systemctl command on every push. Quiet failure mode (systemctl returns 5/non-existent unit, agent shrugs and moves on) means prod-on-kai-server doesn't actually update.

_Originally filed by @coilysiren on 2026-05-20T07:25:48Z - [https://github.com/coilysiren/repo-recall/issues/221](https://github.com/coilysiren/repo-recall/issues/221)_ **Problem** `AGENTS.md` line 49 (Release + post-push section): > Post-push: verify CI at +300s; `brew outdated` -> `brew upgrade`; `coily ssh systemctl start repo-recall-update.service`. Skip for docs-only. That post-push step is stale. repo-recall is Kubernetes-only on kai-server now (`infrastructure/deploy/repo-recall.yml`: three containers per pod, tailscale sidecar, tailnet-only). There is no `repo-recall-update.service` to systemctl-start. **Fix** Rewrite the post-push step to: - For Mac (local brew install): `brew outdated` -> `brew upgrade coilysiren/repo-recall/repo-recall` -> `brew services restart repo-recall`. (Mac is the staging-shape consumer of the new release, ref coilysiren/session-lattice#21 cabling.) - For kai-server (prod): the deploy lands via the container image build (`docker.yml`) and a k3s rollout. Document the actual rollout command (`coily ops kubectl rollout restart deployment/repo-recall -n repo-recall` or whatever the canonical wrapper is). Cross-link `infrastructure/deploy/repo-recall.yml` so the next reader knows where the manifest lives. **Why it matters** The agent will literally try to run the systemctl command on every push. Quiet failure mode (systemctl returns 5/non-existent unit, agent shrugs and moves on) means prod-on-kai-server doesn't actually update.
coilysiren added
P3
and removed
P2
labels 2026-05-31 07:01:19 +00:00
Sign in to join this conversation.
No labels
P0
P1
P2
P3
P4
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/repo-recall#16
No description provided.