Files
buzz/.github/workflows/mesh-lifecycle.yml
Wes 69b1225923 fix(ci): prevent poisoned Rust caches (#6618)
## Summary

- roll every `Swatinem/rust-cache` use back from v2.9.2 to the last
known-good v2.9.1 digest
- give Unit Tests a new `sherpa-cache-v1` key so it cannot restore the
existing poisoned artifact
- pin Renovate to v2.9.1 and add CI contracts that reject unsafe cache
actions or a misplaced generation key

## Why

After #5441 upgraded rust-cache to v2.9.2, warm-cache `main` Unit Tests
runs began failing while linking `buzz-voice` with `could not find
native static library sherpa-onnx-c-api`. The failed run at `db5617dd1`
restored the same 1.4 KB cache generation that had already failed at
`01091c15a`; the preceding cold run at `26f4c3ed3` downloaded sherpa
1.13.4 and passed.

v2.9.2 changed target cleanup, while `sherpa-onnx-sys` treats its
prebuilt `lib/` directory as proof that the native archive exists.
Rolling back the action and invalidating the affected key removes both
sides of that failure state without disabling target caching.

## Validation

At `6da0037a0407fc498cd482fcbbb74c7a15907e9f`:

- `scripts/test-rust-cache-contract.sh`
- `scripts/test-rust-cache-contract-regressions.sh`
- negative fixtures reject a bad digest in a newly named `.yaml`
workflow and a generation key moved outside the cache action's `with`
block
- YAML parse for all workflows
- release, desktop candidate, mobile release, mobile candidate, and
mobile worktree source contracts
- `just file-size-check`
- pre-commit and pre-push hooks

The PR Unit Tests run proves the cold-cache path because pull requests
restore but do not save Rust caches. The first successful `main` run
after merge will save the new Unit Tests key; the following `main` run
will exercise the warm restore.

## Related issue

None found.

Signed-off-by: Wes <wesbillman@users.noreply.github.com>
Co-authored-by: Carl <32a2e2c9d428ee08902cab75d956da2c1d235a22d4766b0dd4138bf6e2e5db1d@buzz.block.builderlab.xyz>
2026-08-24 16:21:31 -06:00

112 lines
4.2 KiB
YAML

name: Mesh Lifecycle
# Relay-driven mesh lifecycle smoke: membership → signed discovery notes →
# relay-derived allowlist → join → CPU inference over QUIC → stranger denied
# (relay membership rejection + no routed inference, with a differential
# trusted-inference health proof so a dead serve node can't fake a denial).
# Runs the full Buzz "shared compute" join story with three real mesh-llm
# node processes on one runner, using the Buzz relay as the control plane
# (no hand-carried invite tokens). Mirrors the shape mesh-llm's own CI uses
# for its two-node smokes (tiny CPU model, one runner, real QUIC mesh).
on:
push:
branches: [main]
paths:
- 'crates/buzz-relay/examples/mesh_*.rs'
- 'crates/buzz-relay/Cargo.toml'
- 'crates/buzz-admin/**'
- 'crates/buzz-test-client/**'
- 'crates/buzz-ws-client/**'
- 'Cargo.lock'
- 'desktop/src-tauri/src/mesh_llm/**'
- 'scripts/ci-mesh-lifecycle-smoke.sh'
- 'scripts/start-relay-for-tests.sh'
- '.github/workflows/mesh-lifecycle.yml'
pull_request:
paths:
- 'crates/buzz-relay/examples/mesh_*.rs'
- 'crates/buzz-relay/Cargo.toml'
- 'crates/buzz-admin/**'
- 'crates/buzz-test-client/**'
- 'crates/buzz-ws-client/**'
- 'Cargo.lock'
- 'desktop/src-tauri/src/mesh_llm/**'
- 'scripts/ci-mesh-lifecycle-smoke.sh'
- 'scripts/start-relay-for-tests.sh'
- '.github/workflows/mesh-lifecycle.yml'
workflow_dispatch:
concurrency:
group: mesh-lifecycle-${{ github.event_name == 'pull_request' && github.ref || github.sha }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
env:
CARGO_TERM_COLOR: always
jobs:
lifecycle-smoke:
name: Relay-Driven Mesh Lifecycle Smoke
runs-on: ubuntu-24.04
timeout-minutes: 45
permissions:
contents: read
steps:
- uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
- uses: cashapp/activate-hermit@cea9af7913204a965fd488637a8d1811bba2e616 # v1
- uses: Swatinem/rust-cache@e18b497796c12c097a38f9edb9d0641fb99eee32 # v2.9.1
with:
save-if: ${{ github.event_name != 'pull_request' }}
# The mesh-llm SDK downloads a signed native runtime (llama.cpp CPU
# build) on first init, and the serve node downloads the smoke model
# from HuggingFace on first run. Key on the lockfile so a mesh pin bump
# rolls the runtime cache; the model ref is stable.
- name: Restore mesh runtime + model caches
id: mesh-caches
uses: actions/cache/restore@caa296126883cff596d87d8935842f9db880ef25 # v5
with:
path: |
~/.cache/mesh-llm/native-runtimes
~/.cache/huggingface/hub
key: mesh-lifecycle-${{ runner.os }}-smollm2-135m-${{ hashFiles('Cargo.lock') }}
restore-keys: |
mesh-lifecycle-${{ runner.os }}-smollm2-135m-
- name: Start integration services
run: |
for attempt in 1 2 3; do
if docker compose up -d postgres redis minio minio-init; then
break
fi
if [ "$attempt" -eq 3 ]; then
echo "docker compose up failed after 3 attempts" >&2
exit 1
fi
echo "docker compose up failed (attempt $attempt), retrying in $((attempt * 5))s..." >&2
sleep $((attempt * 5))
done
- name: Run relay-driven mesh lifecycle smoke
run: ./scripts/ci-mesh-lifecycle-smoke.sh 2>&1 | tee /tmp/mesh-lifecycle-harness.log
- name: Save mesh runtime + model caches
if: github.ref == 'refs/heads/main' && steps.mesh-caches.outputs.cache-hit != 'true'
uses: actions/cache/save@caa296126883cff596d87d8935842f9db880ef25 # v5
with:
path: |
~/.cache/mesh-llm/native-runtimes
~/.cache/huggingface/hub
key: mesh-lifecycle-${{ runner.os }}-smollm2-135m-${{ hashFiles('Cargo.lock') }}
- name: Upload relay + harness logs
if: failure()
uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
with:
name: mesh-lifecycle-logs
path: |
/tmp/buzz-relay.log
/tmp/mesh-lifecycle-harness.log
if-no-files-found: ignore