smoke test all linux target architectures (#2016)

* smoke test all linux target architectures * go mod update * use 'real' target cluster * better logging * newer alpine
2026-04-15 07:16:34 +00:00 · 2026-04-14 18:30:26 +04:00
parent 6d1b1eba49
commit 69743729a3
4 changed files with 160 additions and 7 deletions
--- a/.github/workflows/build-test.yaml
+++ b/.github/workflows/build-test.yaml
@@ -104,6 +104,134 @@ jobs:
          path: bin/
          retention-days: 1

+  # Minimal linux architecture smoke tests
+  linux-arch-smoke:
+    if: needs.changes.outputs.go-files == 'true'
+    needs: [changes, lint]
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - goarch: amd64
+            docker_platform: linux/amd64
+            expected_machine: x86_64
+          - goarch: arm64
+            docker_platform: linux/arm64
+            expected_machine: aarch64
+          - goarch: arm
+            goarm: "7"
+            docker_platform: linux/arm/v7
+            expected_machine: armv7l
+          - goarch: riscv64
+            docker_platform: linux/riscv64
+            expected_machine: riscv64
+    steps:
+      - uses: actions/checkout@v6
+      - uses: actions/setup-go@v6
+        with:
+          go-version-file: 'go.mod'
+      - name: Setup K3s
+        uses: replicatedhq/action-k3s@main
+        with:
+          version: v1.31.2-k3s1
+      - uses: docker/setup-qemu-action@v3
+        with:
+          platforms: arm64,arm,riscv64
+
+      - name: Build linux/${{ matrix.goarch }} binaries
+        env:
+          CGO_ENABLED: "0"
+          GOOS: linux
+          GOARCH: ${{ matrix.goarch }}
+          GOARM: ${{ matrix.goarm }}
+        run: |
+          if [ -n "${GOARM}" ]; then
+            echo "Building for GOARCH=${GOARCH} GOARM=${GOARM}"
+          else
+            echo "Building for GOARCH=${GOARCH}"
+          fi
+          make build
+
+      - name: Collect a minimal support bundle on ${{ matrix.docker_platform }}
+        env:
+          PREFLIGHT_AUTO_UPDATE: "false"
+          TROUBLESHOOT_AUTO_UPDATE: "false"
+        run: |
+          set -euo pipefail
+          tmpdir="$(mktemp -d)"
+          trap 'rm -rf "$tmpdir"' EXIT
+          kubeconfig_path="${KUBECONFIG:-$HOME/.kube/config}"
+
+          assert_tar_member() {
+            local pattern="$1"
+            local label="$2"
+            local member
+            member="$(grep -m1 "$pattern" "$tmpdir/archive-members.txt" || true)"
+            if [ -z "$member" ]; then
+              echo "Expected ${label} was not found in ${archive}"
+              exit 1
+            fi
+            echo "Found ${label}: ${member}"
+          }
+
+          echo "Using kubeconfig at ${kubeconfig_path}"
+          test -f "${kubeconfig_path}"
+          cp "${kubeconfig_path}" "$tmpdir/kubeconfig"
+
+          echo "Verifying K3s cluster is reachable on the host"
+          kubectl --kubeconfig "$tmpdir/kubeconfig" get nodes -o wide
+
+          docker run --rm \
+            --platform ${{ matrix.docker_platform }} \
+            --network host \
+            -v "$PWD:/src" \
+            -v "$tmpdir:/tmp/linux-arch-smoke" \
+            -w /src \
+            alpine:3.23 \
+            sh -ec '
+              export PREFLIGHT_AUTO_UPDATE=false
+              export TROUBLESHOOT_AUTO_UPDATE=false
+              /src/bin/support-bundle \
+                --kubeconfig /tmp/linux-arch-smoke/kubeconfig \
+                --interactive=false \
+                /src/test/e2e/support-bundle/spec/linuxArchSmokeLocalHostCollectors.yaml \
+                --output /tmp/linux-arch-smoke/linux-arch-support-bundle.tar.gz
+            '
+
+          archive="$tmpdir/linux-arch-support-bundle.tar.gz"
+          test -f "$archive"
+          echo "Generated support bundle archive: $archive"
+          tar -tf "$archive" > "$tmpdir/archive-members.txt"
+
+          assert_tar_member '/host-collectors/system/cpu.json$' 'cpu.json'
+          assert_tar_member '/host-collectors/system/memory.json$' 'memory.json'
+          assert_tar_member '/host-collectors/run-host/uname.txt$' 'uname.txt'
+          assert_tar_member '/cluster-resources/nodes.json$' 'cluster-resources nodes.json'
+          assert_tar_member '/cluster-resources/resources.json$' 'cluster-resources resources.json'
+          assert_tar_member '/analysis.json$' 'analysis.json'
+
+          echo "Checking uname -m output"
+          uname_path="$(grep -m1 '/host-collectors/run-host/uname.txt$' "$tmpdir/archive-members.txt")"
+          uname_machine="$(tar -xOf "$archive" "$uname_path" | tr -d '\r\n')"
+          if [ "$uname_machine" != '${{ matrix.expected_machine }}' ]; then
+            echo "Unexpected uname -m output: ${uname_machine}"
+            exit 1
+          fi
+          echo "Analyzer input check passed: uname -m reported ${uname_machine}"
+
+          echo "Checking analysis.json"
+          analysis_path="$(grep -m1 '/analysis.json$' "$tmpdir/archive-members.txt")"
+          tar -xOf "$archive" "$analysis_path" > "$tmpdir/analysis.json"
+          jq -e '
+            .[]
+            | select(.insight.primary == "Linux arch smoke node count")
+            | select(.insight.detail == "This cluster has exactly 1 node")
+            | select(.severity == "debug")
+          ' "$tmpdir/analysis.json" >/dev/null
+          echo "Found expected analyzer result: Linux arch smoke node count -> This cluster has exactly 1 node"
+
  # E2E tests
  e2e:
    if: needs.changes.outputs.go-files == 'true' || github.event_name == 'push'
@@ -143,7 +271,7 @@ jobs:
  # Success summary
  success:
    if: always()
-    needs: [lint, test, build, e2e]
+    needs: [lint, test, build, linux-arch-smoke, e2e]
    runs-on: ubuntu-latest
    steps:
      - name: Check results
@@ -152,6 +280,7 @@ jobs:
          if [[ "${{ needs.lint.result }}" == "failure" ]] || \
             [[ "${{ needs.test.result }}" == "failure" ]] || \
             [[ "${{ needs.build.result }}" == "failure" ]] || \
+             [[ "${{ needs.linux-arch-smoke.result }}" == "failure" ]] || \
             [[ "${{ needs.e2e.result }}" == "failure" ]]; then
            echo "::error::Some jobs failed or were cancelled"
            exit 1
@@ -161,6 +290,7 @@ jobs:
          if [[ "${{ needs.lint.result }}" == "cancelled" ]] || \
             [[ "${{ needs.test.result }}" == "cancelled" ]] || \
             [[ "${{ needs.build.result }}" == "cancelled" ]] || \
+             [[ "${{ needs.linux-arch-smoke.result }}" == "cancelled" ]] || \
             [[ "${{ needs.e2e.result }}" == "cancelled" ]]; then
            echo "::error::Some jobs failed or were cancelled"
            exit 1
--- a/go.mod
+++ b/go.mod
@@ -47,7 +47,7 @@ require (
 	go.opentelemetry.io/otel v1.43.0
 	go.opentelemetry.io/otel/sdk v1.43.0
 	go.podman.io/image/v5 v5.39.2
-	golang.org/x/exp v0.0.0-20241217172543-b2144cdd0a67
+	golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f
 	golang.org/x/mod v0.35.0
 	golang.org/x/sync v0.20.0
 	gopkg.in/yaml.v2 v2.4.0
@@ -166,7 +166,7 @@ require (
 	go.opentelemetry.io/otel/trace v1.43.0 // indirect
 	go.yaml.in/yaml/v2 v2.4.3 // indirect
 	go.yaml.in/yaml/v3 v3.0.4 // indirect
-	golang.org/x/tools v0.43.0 // indirect
+	golang.org/x/tools v0.44.0 // indirect
 	google.golang.org/genproto/googleapis/api v0.0.0-20260226221140-a57be14db171 // indirect
 	google.golang.org/genproto/googleapis/rpc v0.0.0-20260226221140-a57be14db171 // indirect
 	gopkg.in/evanphx/json-patch.v4 v4.13.0 // indirect
--- a/go.sum
+++ b/go.sum
@@ -650,8 +650,8 @@ golang.org/x/crypto v0.0.0-20180904163835-0709b304e793/go.mod h1:6SG95UA2DQfeDnf
 golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
 golang.org/x/crypto v0.50.0 h1:zO47/JPrL6vsNkINmLoo/PH1gcxpls50DNogFvB5ZGI=
 golang.org/x/crypto v0.50.0/go.mod h1:3muZ7vA7PBCE6xgPX7nkzzjiUq87kRItoJQM1Yo8S+Q=
-golang.org/x/exp v0.0.0-20241217172543-b2144cdd0a67 h1:1UoZQm6f0P/ZO0w1Ri+f+ifG/gXhegadRdwBIXEFWDo=
-golang.org/x/exp v0.0.0-20241217172543-b2144cdd0a67/go.mod h1:qj5a5QZpwLU2NLQudwIN5koi3beDhSAlJwa67PuM98c=
+golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f h1:W3F4c+6OLc6H2lb//N1q4WpJkhzJCK5J6kUi1NTVXfM=
+golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f/go.mod h1:J1xhfL/vlindoeF/aINzNzt2Bket5bjo9sdOYzOsU80=
 golang.org/x/mod v0.35.0 h1:Ww1D637e6Pg+Zb2KrWfHQUnH2dQRLBQyAtpr/haaJeM=
 golang.org/x/mod v0.35.0/go.mod h1:+GwiRhIInF8wPm+4AoT6L0FA1QWAad3OMdTRx4tFYlU=
 golang.org/x/net v0.0.0-20180906233101-161cd47e91fd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
@@ -686,8 +686,8 @@ golang.org/x/text v0.36.0/go.mod h1:NIdBknypM8iqVmPiuco0Dh6P5Jcdk8lJL0CUebqK164=
 golang.org/x/time v0.15.0 h1:bbrp8t3bGUeFOx08pvsMYRTCVSMk89u4tKbNOZbp88U=
 golang.org/x/time v0.15.0/go.mod h1:Y4YMaQmXwGQZoFaVFk4YpCt4FLQMYKZe9oeV/f4MSno=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.43.0 h1:12BdW9CeB3Z+J/I/wj34VMl8X+fEXBxVR90JeMX5E7s=
-golang.org/x/tools v0.43.0/go.mod h1:uHkMso649BX2cZK6+RpuIPXS3ho2hZo4FVwfoy1vIk0=
+golang.org/x/tools v0.44.0 h1:UP4ajHPIcuMjT1GqzDWRlalUEoY+uzoZKnhOjbIPD2c=
+golang.org/x/tools v0.44.0/go.mod h1:KA0AfVErSdxRZIsOVipbv3rQhVXTnlU6UhKxHd1seDI=
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk=
 gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E=
--- a/test/e2e/support-bundle/spec/linuxArchSmokeLocalHostCollectors.yaml
+++ b/test/e2e/support-bundle/spec/linuxArchSmokeLocalHostCollectors.yaml
@@ -0,0 +1,23 @@
+apiVersion: troubleshoot.sh/v1beta2
+kind: SupportBundle
+metadata:
+  name: linux-arch-smoke-local-host-collectors
+spec:
+  collectors:
+    - clusterResources: {}
+  hostCollectors:
+    - cpu: {}
+    - memory: {}
+    - run:
+        collectorName: uname
+        command: uname
+        args: ["-m"]
+  analyzers:
+    - nodeResources:
+        checkName: Linux arch smoke node count
+        outcomes:
+          - pass:
+              when: "= 1"
+              message: This cluster has exactly 1 node
+          - fail:
+              message: Expected exactly 1 node in the cluster