fix(spur-cloud-api): adapt GPU allocation to CDI-based device API (#63) #79
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: CI | |
| on: | |
| push: | |
| branches: [main] | |
| pull_request: | |
| branches: [main] | |
| workflow_dispatch: | |
| env: | |
| CARGO_TERM_COLOR: always | |
| RUSTFLAGS: "-D warnings" | |
| jobs: | |
| deny: | |
| runs-on: ubuntu-latest | |
| env: | |
| CARGO_DENY_VERSION: "0.19.6" | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Install Rust toolchain | |
| uses: dtolnay/rust-toolchain@stable | |
| - name: Install cargo-deny | |
| run: | | |
| curl -sL "https://github.com/EmbarkStudios/cargo-deny/releases/download/${CARGO_DENY_VERSION}/cargo-deny-${CARGO_DENY_VERSION}-x86_64-unknown-linux-musl.tar.gz" | tar xz -C /usr/local/bin --strip-components=1 | |
| - name: Check dependencies | |
| run: cargo deny check | |
| build-and-test: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Install protobuf compiler | |
| run: sudo apt-get install -y protobuf-compiler | |
| - name: Install Rust toolchain | |
| uses: dtolnay/rust-toolchain@stable | |
| - name: Cache cargo registry and build | |
| uses: actions/cache@v4 | |
| with: | |
| path: | | |
| ~/.cargo/registry | |
| ~/.cargo/git | |
| target | |
| key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }} | |
| restore-keys: | | |
| ${{ runner.os }}-cargo- | |
| - name: Build backend | |
| run: cargo build --all-targets | |
| - name: Test backend | |
| run: cargo test | |
| fmt: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - uses: dtolnay/rust-toolchain@stable | |
| with: | |
| components: rustfmt | |
| - name: Check Rust formatting | |
| run: cargo fmt --all --check | |
| clippy: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Install protobuf compiler | |
| run: sudo apt-get install -y protobuf-compiler | |
| - uses: dtolnay/rust-toolchain@stable | |
| with: | |
| components: clippy | |
| - name: Cache cargo registry and build | |
| uses: actions/cache@v4 | |
| with: | |
| path: | | |
| ~/.cargo/registry | |
| ~/.cargo/git | |
| target | |
| key: ${{ runner.os }}-cargo-clippy-${{ hashFiles('**/Cargo.lock') }} | |
| restore-keys: | | |
| ${{ runner.os }}-cargo-clippy- | |
| - name: Clippy | |
| run: cargo clippy --workspace --all-targets --locked -- -W clippy::all | |
| frontend: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Setup Node.js | |
| uses: actions/setup-node@v4 | |
| with: | |
| node-version: '20' | |
| cache: 'npm' | |
| cache-dependency-path: frontend/package-lock.json | |
| - name: Install dependencies | |
| working-directory: frontend | |
| run: npm ci | |
| - name: Build frontend | |
| working-directory: frontend | |
| run: npm run build | |
| integration: | |
| name: Integration Test (MI300X) | |
| runs-on: [self-hosted, mi300x] | |
| needs: [build-and-test, fmt, clippy, frontend] | |
| concurrency: | |
| group: spur-cloud-${{ github.ref }} | |
| cancel-in-progress: true | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Build release binary | |
| run: | | |
| export PATH="$HOME/bin:$HOME/.cargo/bin:$PATH" | |
| export PROTOC="$HOME/bin/protoc" | |
| export PROTOC_INCLUDE="$HOME/protoc-install/include" | |
| cargo build --release -j 4 | |
| - name: Start test services | |
| run: | | |
| export PATH="$HOME/bin:$HOME/.cargo/bin:$PATH" | |
| # Start spurctld if not running | |
| if ! pgrep -x spurctld >/dev/null; then | |
| ~/spur/etc/start-controller.sh 2>/dev/null || true | |
| sleep 3 | |
| fi | |
| # Start PostgreSQL test database | |
| if ! pg_isready -q 2>/dev/null; then | |
| echo "PostgreSQL not available, skipping integration test" | |
| exit 0 | |
| fi | |
| # Create test database if needed | |
| psql -U postgres -tc "SELECT 1 FROM pg_database WHERE datname='spur_cloud_test'" 2>/dev/null | grep -q 1 || \ | |
| psql -U postgres -c "CREATE DATABASE spur_cloud_test" 2>/dev/null || true | |
| - name: Run API smoke test | |
| run: | | |
| export PATH="$HOME/bin:$HOME/.cargo/bin:$PATH" | |
| # Skip if PostgreSQL is not available | |
| if ! pg_isready -q 2>/dev/null; then | |
| echo "SKIP: PostgreSQL not available on this runner" | |
| echo "Build succeeded — integration test requires PostgreSQL" | |
| exit 0 | |
| fi | |
| # Generate test config | |
| cat > /tmp/spur-cloud-test.toml << 'EOF' | |
| public_url = "http://localhost:9090" | |
| [server] | |
| listen_addr = "127.0.0.1:9090" | |
| backend = "native_host" | |
| [native_host] | |
| agent_port = 6828 | |
| [database] | |
| url = "postgresql://postgres@localhost/spur_cloud_test" | |
| [spur] | |
| controller_addr = "http://127.0.0.1:6817" | |
| [auth] | |
| jwt_secret = "test-secret-for-ci-only-do-not-use-in-prod" | |
| EOF | |
| # Start the API server in background | |
| timeout 30 ./target/release/spur-cloud-api --config /tmp/spur-cloud-test.toml & | |
| API_PID=$! | |
| sleep 3 | |
| # Health check | |
| curl -sf http://localhost:9090/healthz || { echo "FAIL: healthz"; kill $API_PID 2>/dev/null; exit 1; } | |
| echo "PASS: healthz" | |
| # Register + login | |
| REGISTER=$(curl -sf -X POST http://localhost:9090/api/auth/register \ | |
| -H 'Content-Type: application/json' \ | |
| -d '{"email":"ci@test.com","username":"citest","password":"testpassword123"}') | |
| TOKEN=$(echo "$REGISTER" | python3 -c "import sys,json; print(json.load(sys.stdin)['token'])" 2>/dev/null) | |
| [ -n "$TOKEN" ] && echo "PASS: register + login" || { echo "FAIL: register"; kill $API_PID 2>/dev/null; exit 1; } | |
| # Issue #44: Test case-insensitive email login | |
| LOGIN_UPPER=$(curl -s -w '\n%{http_code}' -X POST http://localhost:9090/api/auth/login \ | |
| -H 'Content-Type: application/json' \ | |
| -d '{"email":"CI@TEST.COM","password":"testpassword123"}') | |
| LOGIN_UPPER_CODE=$(echo "$LOGIN_UPPER" | tail -1) | |
| if [ "$LOGIN_UPPER_CODE" = "200" ]; then | |
| echo "PASS: case-insensitive email login (CI@TEST.COM → ci@test.com)" | |
| else | |
| echo "FAIL: case-insensitive login returned $LOGIN_UPPER_CODE"; kill $API_PID 2>/dev/null; exit 1 | |
| fi | |
| # Issue #44: Test wrong password returns specific error | |
| WRONG_PW=$(curl -s -X POST http://localhost:9090/api/auth/login \ | |
| -H 'Content-Type: application/json' \ | |
| -d '{"email":"ci@test.com","password":"wrongpassword"}') | |
| echo "$WRONG_PW" | grep -q "incorrect password" && echo "PASS: wrong password error message" || \ | |
| { echo "FAIL: expected 'incorrect password', got: $WRONG_PW"; kill $API_PID 2>/dev/null; exit 1; } | |
| # Issue #44: Test unknown email returns specific error | |
| UNKNOWN_EMAIL=$(curl -s -X POST http://localhost:9090/api/auth/login \ | |
| -H 'Content-Type: application/json' \ | |
| -d '{"email":"nobody@test.com","password":"testpassword123"}') | |
| echo "$UNKNOWN_EMAIL" | grep -q "no account found" && echo "PASS: unknown email error message" || \ | |
| { echo "FAIL: expected 'no account found', got: $UNKNOWN_EMAIL"; kill $API_PID 2>/dev/null; exit 1; } | |
| # Issue #36: Test GPU quota enforcement | |
| # Register a second user, set quota, try to exceed it | |
| REGISTER2=$(curl -sf -X POST http://localhost:9090/api/auth/register \ | |
| -H 'Content-Type: application/json' \ | |
| -d '{"email":"quota@test.com","username":"quotauser","password":"testpassword123"}') | |
| TOKEN2=$(echo "$REGISTER2" | python3 -c "import sys,json; print(json.load(sys.stdin)['token'])" 2>/dev/null) | |
| # Set quota to 2 GPUs via admin endpoint | |
| curl -sf -X PUT "http://localhost:9090/api/admin/users/quota" \ | |
| -H 'Content-Type: application/json' \ | |
| -H "Authorization: Bearer $TOKEN" \ | |
| -d "{\"email\":\"quota@test.com\",\"max_gpus\":2}" >/dev/null 2>&1 && \ | |
| echo "PASS: set GPU quota" || echo "WARN: admin quota endpoint not available (spurctld may be down)" | |
| # Auth providers | |
| PROVIDERS=$(curl -sf http://localhost:9090/api/auth/providers) | |
| echo "$PROVIDERS" | grep -q "local" && echo "PASS: providers" || echo "WARN: providers response unexpected" | |
| # Issue #48: Test CPU-only session creation (gpu_count=0) | |
| CPU_SESSION=$(curl -s -w '\n%{http_code}' -X POST http://localhost:9090/api/sessions \ | |
| -H 'Content-Type: application/json' \ | |
| -H "Authorization: Bearer $TOKEN" \ | |
| -d '{"name":"cpu-test","container_image":"ubuntu:22.04","gpu_count":0,"gpu_type":"none"}') | |
| CPU_CODE=$(echo "$CPU_SESSION" | tail -1) | |
| if [ "$CPU_CODE" = "201" ]; then | |
| echo "PASS: CPU-only session creation (gpu_count=0)" | |
| else | |
| echo "WARN: CPU-only session returned $CPU_CODE (spurctld may be down)" | |
| fi | |
| # Issue #48: Test cross-validation rejects gpu_count=0 with gpu_type set | |
| INVALID=$(curl -s -w '\n%{http_code}' -X POST http://localhost:9090/api/sessions \ | |
| -H 'Content-Type: application/json' \ | |
| -H "Authorization: Bearer $TOKEN" \ | |
| -d '{"name":"bad","container_image":"ubuntu:22.04","gpu_count":0,"gpu_type":"mi300x"}') | |
| INVALID_CODE=$(echo "$INVALID" | tail -1) | |
| if [ "$INVALID_CODE" = "400" ]; then | |
| echo "PASS: cross-validation rejects gpu_count=0 + gpu_type=mi300x" | |
| else | |
| echo "FAIL: expected 400, got $INVALID_CODE"; kill $API_PID 2>/dev/null; exit 1 | |
| fi | |
| # GPU capacity (requires spurctld) | |
| GPU=$(curl -sf http://localhost:9090/api/gpus -H "Authorization: Bearer $TOKEN" 2>/dev/null) | |
| echo "GPU capacity: $GPU" | |
| echo "PASS: API smoke test complete" | |
| kill $API_PID 2>/dev/null || true |