Fix bugs in CI workflow

.forgejo/workflows/ci.yml

@@ -31,6 +31,18 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - name: Resolve Podman user socket for this runner
+        run: |
+          SVC_USER="ci-service"   # CI runner user on the CI host
+          uid=$(id -u "$SVC_USER")
+          echo "PODMAN_SOCK=/run/user/${uid}/podman/podman.sock" >> "$GITHUB_ENV"
+
+      - name: Verify runner wiring to Podman
+        run: |
+          podman --version
+          test -S "${PODMAN_SOCK}" || { echo "Missing socket ${PODMAN_SOCK}"; exit 1; }
+          # Optional: sanity poke of the service via PiP later
+
       - name: Network/DNS sanity from job container
         run: |
           getent hosts git.gcdo.org || true
@@ -46,6 +58,7 @@ jobs:
       - name: Setup ephemeral PiP container
         env:
           PODMAN_CLIENT_IMG_DIGEST: ${{ env.PODMAN_CLIENT_IMG_DIGEST }}
+          SOCKET_PATH: ${{ env.PODMAN_SOCK }}
         run: |
           chmod +x ./secure_pip_setup.sh
           ./secure_pip_setup.sh
@@ -57,6 +70,9 @@ jobs:
           chmod +x ./pip_ready.sh
           ./pip_ready.sh
 
+      - name: Network sanity from PiP
+        run: podman exec ci-pip-${{ env.RUN_ID }} sh -lc 'getent hosts git.gcdo.org && curl -sS -o /dev/null -w "status=%{http_code}\n" https://git.gcdo.org/api/healthz'
+
       - name: Configure Git for token-based authentication
         run: |
           git config --global url."https://${{ secrets.REGISTRY_USERNAME }}:${{ secrets.REGISTRY_TOKEN }}@${{ env.REGISTRY }}".insteadOf "https://${{ env.REGISTRY }}"
@@ -85,7 +101,7 @@ jobs:
         env:
           WORKSPACE: ${{ github.workspace }}
         run: |
-          podman exec ci-pip-${{ env.RUN_ID }} sh -lc '
+          podman exec -e WORKSPACE="${WORKSPACE}" ci-pip-${{ env.RUN_ID }} sh -lc '
             podman run --rm \
               -v "$WORKSPACE":/workspace \
               -w /workspace \
@@ -95,11 +111,10 @@ jobs:
       - name: Run backend integration tests
         env:
           WORKSPACE: ${{ github.workspace }}
-          RUN_ID: ${{ env.RUN_ID }}
         run: |
-          podman exec ci-pip-${{ env.RUN_ID }} sh -lc '
+          podman exec -e WORKSPACE="${WORKSPACE}" ci-pip-${{ env.RUN_ID }} sh -lc '
             podman run --rm \
-              --network integ-'"$RUN_ID"' \
+              --network integ-${{ env.RUN_ID }} \
               -v "$WORKSPACE":/workspace \
               -w /workspace \
               -e DATABASE_URL=postgres://postgres:password@test-postgres:5432/sharenet_test \
@@ -122,6 +137,18 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - name: Resolve Podman user socket for this runner
+        run: |
+          SVC_USER="ci-service"   # CI runner user on the CI host
+          uid=$(id -u "$SVC_USER")
+          echo "PODMAN_SOCK=/run/user/${uid}/podman/podman.sock" >> "$GITHUB_ENV"
+
+      - name: Verify runner wiring to Podman
+        run: |
+          podman --version
+          test -S "${PODMAN_SOCK}" || { echo "Missing socket ${PODMAN_SOCK}"; exit 1; }
+          # Optional: sanity poke of the service via PiP later
+
       - name: Verify pinned digests provided
         run: |
           for v in NODE_IMG_DIGEST PODMAN_CLIENT_IMG_DIGEST; do
@@ -132,6 +159,7 @@ jobs:
       - name: Setup ephemeral PiP container
         env:
           PODMAN_CLIENT_IMG_DIGEST: ${{ env.PODMAN_CLIENT_IMG_DIGEST }}
+          SOCKET_PATH: ${{ env.PODMAN_SOCK }}
         run: |
           chmod +x ./secure_pip_setup.sh
           ./secure_pip_setup.sh
@@ -143,11 +171,14 @@ jobs:
           chmod +x ./pip_ready.sh
           ./pip_ready.sh
 
+      - name: Network sanity from PiP
+        run: podman exec ci-pip-${{ env.RUN_ID }} sh -lc 'getent hosts git.gcdo.org && curl -sS -o /dev/null -w "status=%{http_code}\n" https://git.gcdo.org/api/healthz'
+
       - name: Run frontend tests (digest-pinned)
         env:
           WORKSPACE: ${{ github.workspace }}
         run: |
-          podman exec ci-pip-${{ env.RUN_ID }} sh -lc '
+          podman exec -e WORKSPACE="${WORKSPACE}" ci-pip-${{ env.RUN_ID }} sh -lc '
             podman run --rm \
               -v "$WORKSPACE":/workspace \
               -w /workspace \
@@ -164,9 +195,22 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - name: Resolve Podman user socket for this runner
+        run: |
+          SVC_USER="ci-service"   # CI runner user on the CI host
+          uid=$(id -u "$SVC_USER")
+          echo "PODMAN_SOCK=/run/user/${uid}/podman/podman.sock" >> "$GITHUB_ENV"
+
+      - name: Verify runner wiring to Podman
+        run: |
+          podman --version
+          test -S "${PODMAN_SOCK}" || { echo "Missing socket ${PODMAN_SOCK}"; exit 1; }
+          # Optional: sanity poke of the service via PiP later
+
       - name: Setup ephemeral PiP container
         env:
           PODMAN_CLIENT_IMG_DIGEST: ${{ env.PODMAN_CLIENT_IMG_DIGEST }}
+          SOCKET_PATH: ${{ env.PODMAN_SOCK }}
         run: |
           chmod +x ./secure_pip_setup.sh
           ./secure_pip_setup.sh
@@ -178,6 +222,9 @@ jobs:
           chmod +x ./pip_ready.sh
           ./pip_ready.sh
 
+      - name: Network sanity from PiP
+        run: podman exec ci-pip-${{ env.RUN_ID }} sh -lc 'getent hosts git.gcdo.org && curl -sS -o /dev/null -w "status=%{http_code}\n" https://git.gcdo.org/api/healthz'
+
       - name: Login to Forgejo Container Registry securely
         env:
           REGISTRY_HOST: ${{ env.REGISTRY }}
@@ -215,9 +262,22 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - name: Resolve Podman user socket for this runner
+        run: |
+          SVC_USER="ci-service"   # CI runner user on the CI host
+          uid=$(id -u "$SVC_USER")
+          echo "PODMAN_SOCK=/run/user/${uid}/podman/podman.sock" >> "$GITHUB_ENV"
+
+      - name: Verify runner wiring to Podman
+        run: |
+          podman --version
+          test -S "${PODMAN_SOCK}" || { echo "Missing socket ${PODMAN_SOCK}"; exit 1; }
+          # Optional: sanity poke of the service via PiP later
+
       - name: Setup ephemeral PiP container
         env:
           PODMAN_CLIENT_IMG_DIGEST: ${{ env.PODMAN_CLIENT_IMG_DIGEST }}
+          SOCKET_PATH: ${{ env.PODMAN_SOCK }}
         run: |
           chmod +x ./secure_pip_setup.sh
           ./secure_pip_setup.sh
@@ -229,6 +289,9 @@ jobs:
           chmod +x ./pip_ready.sh
           ./pip_ready.sh
 
+      - name: Network sanity from PiP
+        run: podman exec ci-pip-${{ env.RUN_ID }} sh -lc 'getent hosts git.gcdo.org && curl -sS -o /dev/null -w "status=%{http_code}\n" https://git.gcdo.org/api/healthz'
+
       - name: Login to Forgejo Container Registry securely
         env:
           REGISTRY_HOST: ${{ env.REGISTRY }}
@@ -292,11 +355,31 @@ jobs:
             fi
           }
           
-          # Substitute environment variables in prod-pod.yml
-          envsubst < deploy/prod-pod.yml > prod-pod-deploy.yml
+      - name: Resolve Podman user socket (prod)
+        run: |
+          SVC_USER="prod-service"
+          uid=$(id -u "$SVC_USER")
+          echo "PODMAN_SOCK=/run/user/${uid}/podman/podman.sock" >> "$GITHUB_ENV"
+
+      - name: Verify runner wiring to Podman
+        run: |
+          podman --version
+          test -S "${PODMAN_SOCK}" || { echo "Missing socket ${PODMAN_SOCK}"; exit 1; }
+          # Optional: sanity poke of the service via PiP later
+
+      - name: Render prod spec
+        run: envsubst < deploy/prod-pod.yml > prod-pod-deploy.yml
+
+      - name: Apply with Podman (no registry login needed)
+        env:
+          PODMAN_CLIENT_IMG_DIGEST: ${{ env.PODMAN_CLIENT_IMG_DIGEST }}
+        run: |
+          cat prod-pod-deploy.yml | \
+          podman run --rm -i \
+            -v "${PODMAN_SOCK}:${PODMAN_SOCK}" \
+            -e "CONTAINER_HOST=unix://${PODMAN_SOCK}" \
+            "${PODMAN_CLIENT_IMG_DIGEST}" \
+            sh -lc 'cat >/tmp/spec.yml; podman play kube --replace /tmp/spec.yml'
           
-          # Deploy the pod (replace existing if it exists)
-          podman play kube --replace prod-pod-deploy.yml
-          
-          # Clean up generated file
-          rm -f prod-pod-deploy.yml
+      - name: Clean up generated file
+        run: rm -f prod-pod-deploy.yml

deploy/prod-pod.yml

@@ -8,18 +8,15 @@ metadata:
     io.containers.no-new-privileges: "true"
 spec:
   hostname: sharenet-production-pod
-  # Security: run as non-root user with specific UID/GID (matches PROD_SERVICE_USER)
-  securityContext:
-    runAsNonRoot: true
-    runAsUser: 1000
-    runAsGroup: 1000
-    fsGroup: 1000
 
   containers:
   - name: postgres
     image: ${REGISTRY_HOST}/${APP_NAME}/postgres:${IMAGE_TAG}
     # Security: drop all capabilities, read-only root filesystem except data volume
     securityContext:
+      runAsNonRoot: true
+      runAsUser: 1000
+      runAsGroup: 1000
       readOnlyRootFilesystem: true
       allowPrivilegeEscalation: false
       capabilities:
@@ -29,26 +26,27 @@ spec:
       value: ${POSTGRES_DATABASE_NAME}
     - name: POSTGRES_USER
       value: ${POSTGRES_USERNAME}
+    - name: POSTGRES_PASSWORD
+      value: ${POSTGRES_PASSWORD}
     - name: PGDATA
       value: /var/lib/postgresql/data/pgdata
     ports:
     - containerPort: ${POSTGRES_PORT}
       protocol: TCP
     volumeMounts:
-    - name: pgdata
-      mountPath: /var/lib/postgresql/data
-      readOnly: false
+      - { name: pgdata,       mountPath: /var/lib/postgresql/data,  readOnly: false }
+      - { name: postgres-run, mountPath: /var/run/postgresql,       readOnly: false }
     # Health checks
     livenessProbe:
       exec:
-        command: ["pg_isready", "-U", "sharenet"]
+        command: ["pg_isready", "-U", "${POSTGRES_USERNAME}"]
       initialDelaySeconds: 30
       periodSeconds: 30
       timeoutSeconds: 10
       failureThreshold: 3
     readinessProbe:
       exec:
-        command: ["pg_isready", "-U", "sharenet"]
+        command: ["pg_isready", "-U", "${POSTGRES_USERNAME}"]
       initialDelaySeconds: 5
       periodSeconds: 10
       timeoutSeconds: 5
@@ -65,23 +63,29 @@ spec:
     image: ${REGISTRY_HOST}/${APP_NAME}/backend:${IMAGE_TAG}
     # Security: drop all capabilities, read-only root filesystem
     securityContext:
+      runAsNonRoot: true
+      runAsUser: 1000
+      runAsGroup: 1000
       readOnlyRootFilesystem: true
       allowPrivilegeEscalation: false
       capabilities:
         drop: ["ALL"]
     env:
     - name: DATABASE_URL
-      value: postgres://${POSTGRES_USERNAME}:${POSTGRES_PASSWORD}@{POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DATABASE_NAME}?sslmode=disable
+      value: "postgres://${POSTGRES_USERNAME}:${POSTGRES_PASSWORD}@localhost:${POSTGRES_PORT}/${POSTGRES_DATABASE_NAME}?sslmode=disable"
     - name: PORT
-      value: ${PROD_BACKEND_PORT}
+      value: "${PROD_BACKEND_PORT}"
     ports:
     - containerPort: ${PROD_BACKEND_PORT}
       protocol: TCP
+    volumeMounts:
+      - { name: backend-tmp, mountPath: /tmp }
+      - { name: backend-tmp, mountPath: /run }
     # Health checks
     livenessProbe:
       httpGet:
         path: /health
-        port: ${PROD_BACKEND_PORT}
+        port: "${PROD_BACKEND_PORT}"
         scheme: HTTP
       initialDelaySeconds: 30
       periodSeconds: 30
@@ -90,7 +94,7 @@ spec:
     readinessProbe:
       httpGet:
         path: /health
-        port: ${PROD_BACKEND_PORT}
+        port: "${PROD_BACKEND_PORT}"
         scheme: HTTP
       initialDelaySeconds: 10
       periodSeconds: 10
@@ -108,6 +112,9 @@ spec:
     image: ${REGISTRY_HOST}/${APP_NAME}/frontend:${IMAGE_TAG}
     # Security: drop all capabilities, read-only root filesystem
     securityContext:
+      runAsNonRoot: true
+      runAsUser: 1000
+      runAsGroup: 1000
       readOnlyRootFilesystem: true
       allowPrivilegeEscalation: false
       capabilities:
@@ -116,12 +123,15 @@ spec:
     - name: NEXT_PUBLIC_API_HOST
       value: ${PROD_BACKEND_HOST}
     - name: NEXT_PUBLIC_API_PORT
-      value: ${PROD_BACKEND_PORT}
+      value: "${PROD_BACKEND_PORT}"
     - name: PORT
-      value: ${PROD_FRONTEND_PORT}
+      value: "${PROD_FRONTEND_PORT}"
     ports:
     - containerPort: ${PROD_FRONTEND_PORT}
       protocol: TCP
+    volumeMounts:
+      - { name: frontend-tmp, mountPath: /tmp }
+      - { name: frontend-tmp, mountPath: /run }
     # Resource limits
     resources:
       requests:
@@ -141,25 +151,21 @@ spec:
         drop: ["ALL"]
     ports:
     - containerPort: 80
+      hostPort: 8080
       protocol: TCP
     - containerPort: 443
+      hostPort: 8443
       protocol: TCP
     volumeMounts:
-    - name: nginx-conf
-      mountPath: /etc/nginx/nginx.conf
-      subPath: nginx.conf
-      readOnly: true
-    - name: nginx-cache
-      mountPath: /var/cache/nginx
-      readOnly: false
-    - name: letsencrypt
-      mountPath: /etc/letsencrypt
-      readOnly: true
+      - { name: nginx-run,    mountPath: /var/run,              readOnly: false }
+      - { name: nginx-conf,   mountPath: /etc/nginx/nginx.conf, readOnly: true, subPath: nginx.conf }
+      - { name: nginx-cache,  mountPath: /var/cache/nginx,      readOnly: false }
+      - { name: letsencrypt,  mountPath: /etc/letsencrypt,      readOnly: true }
     # Health check
     livenessProbe:
       httpGet:
         path: /healthz
-        port: 80
+        port: 8090
         scheme: HTTP
       initialDelaySeconds: 10
       periodSeconds: 30
@@ -179,6 +185,14 @@ spec:
     hostPath:
       path: /opt/sharenet/volumes/pgdata
       type: DirectoryOrCreate
+  - name: postgres-run
+    emptyDir: {}
+  - name: backend-tmp
+    emptyDir: { medium: Memory }
+  - name: frontend-tmp
+    emptyDir: { medium: Memory }
+  - name: nginx-run
+    emptyDir: {}
   - name: nginx-conf
     hostPath:
       path: /opt/sharenet/nginx