Commit Graph

2 Commits

Author SHA1 Message Date
e3mrah
cadc7b5cea
fix(sandbox-ci): mcp-server Dockerfile repo-root context + pty/mcp auto-bump wiring (chart was half-deployable) (#1667)
Sandbox chart was un-deployable end-to-end because three CI-side gaps
compounded after PR #1658 wired the mcp-server module to depend on
core/controllers + core/services/shared via `replace` directives:

1. **mcp-server Dockerfile built against a too-narrow context**. The
   workflow passed `context: products/sandbox/mcp-server` and the
   Dockerfile assumed `COPY . .` could see everything it needed, but
   the `replace ../../../core/controllers` line in the module's go.mod
   only resolves when the build can actually reach those paths. Result:
   every push after #1658 failed at `go build` with `module not found`.
   Fix mirrors core/controllers/sandbox/Dockerfile (Slice-CC1 layout):
   COPY the replace targets' module roots + sources, then build with
   WORKDIR set to the dependent module. Static binary still produced
   into a distroless/static-debian12:nonroot final stage.

2. **mcp-server workflow had no chart auto-bump step**. Even after a
   green build, `runtime.mcpImage` in platform/sandbox/chart/values.yaml
   stayed empty so the chart's `required` guard
   (deployment.yaml line 72) refused to render. Added the same
   yq-bump + bot-commit pattern build-sandbox-controller.yaml already
   uses, targeting `.runtime.mcpImage` and writing a fully-qualified
   `<repo>:<sha>` string (consumer reads it as one image reference,
   not a {repository,tag} pair). Also widened paths-filter to include
   core/controllers/** + core/services/shared/** so changes to the
   replace targets re-trigger the build.

3. **pty-server workflow had no auto-bump either**. Same surgery:
   yq-bump `.runtime.ptyServerImage` + commit-and-push. Context stays
   narrow (pty-server has no cross-tree `replace` directives).

4. **Stop-gap pin values for runtime.{ptyServerImage,mcpImage}** so the
   next chart roll out doesn't fail-fast before the rebuilt workflows
   land their first bumps:
   - ptyServerImage → ad5163e6 (current latest pty-server)
   - mcpImage → 1b0e86c (last pre-#1658 green build; the rebuilt
     workflow will land the next real SHA on the next push to main).

Verified locally:
- `go build ./products/sandbox/mcp-server/...` clean (43.8 MB static
  binary at /tmp/openova-sandbox-mcp; `file` confirms statically
  linked ELF).
- `helm template test platform/sandbox/chart --set enabled=true …`
  renders cleanly; both env vars carry the SHA-pinned image refs.

No Chart.yaml bump. Read-only clusters.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-18 13:22:17 +04:00
e3mrah
1b0e86cb1a
ci(sandbox): build workflows for controller + pty-server + mcp-server (so chart can actually deploy) (#1632)
PR #1622 shipped the sandbox-controller binary + chart, and PR #1618
shipped pty-server + mcp-server scaffolds, but neither came with CI
build workflows — meaning the chart's image.repository points at a
GHCR package that no workflow ever publishes (ImagePullBackOff on
every install). Per docs/INVIOLABLE-PRINCIPLES.md #4a every runtime
image MUST be produced by a GitHub Actions workflow from a committed
git SHA; this PR closes that gap.

Three new workflows, all event-driven (push paths-filter + PR +
workflow_dispatch, no cron):

- build-sandbox-controller.yaml — mirrors build-application-controller
  (shared core/controllers go.mod, go vet + race tests, Buildx push,
  cosign keyless sign, SBOM attest, auto-bump platform/sandbox/chart/
  values.yaml image.tag back to main so the next install picks up the
  SHA-pinned image without operator action).

- build-sandbox-pty-server.yaml — separate go module under
  products/sandbox/pty-server (own go.mod/go.sum), Dockerfile uses
  COPY . . so build context is the server directory. Same Buildx +
  cosign + SBOM flow as the controller. No values.yaml bump yet:
  Wave-2 wiring of the StatefulSet template will land in a follow-up.

- build-sandbox-mcp-server.yaml — stdlib-only stdio MCP sidecar
  (no go.sum yet), same shape as pty-server.

Per `feedback_no_mvp_no_workarounds.md` rule 1 (target-state, never
"manual follow-up bump") the controller workflow auto-bumps the chart
values.yaml so a Sovereign overlay flipping `enabled: true` Just Works.
Per the user's hard rule for this PR, no Chart.yaml bump and no
blueprint-release dispatch — the Sandbox chart's publication cadence
is gated by Wave-2 readiness, not per-image builds.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-18 10:11:28 +04:00