Merge branch 'dev' of https://gitea.airlabs.art/zyc/video-shuoshan into dev

.gitea/workflows/deploy.yaml

@@ -49,28 +49,45 @@ jobs:
         id: build_backend
         run: |
           set -o pipefail
-          DOCKER_BUILDKIT=0 docker build \
-            --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:${{ env.IMAGE_TAG }} \
-            --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:latest \
-            ./backend 2>&1 | tee /tmp/build.log
-          docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:${{ env.IMAGE_TAG }}
-          docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:latest
+          for attempt in 1 2 3; do
+            echo "Build backend attempt $attempt/3..."
+            DOCKER_BUILDKIT=0 docker build \
+              --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:${{ env.IMAGE_TAG }} \
+              --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:latest \
+              ./backend 2>&1 | tee /tmp/build.log && break
+            echo "Attempt $attempt failed, retrying in 10s..." && sleep 10
+          done
+          for attempt in 1 2 3; do
+            docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:${{ env.IMAGE_TAG }} && \
+            docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-backend:latest && break
+            echo "Push attempt $attempt failed, retrying in 10s..." && sleep 10
+          done
 
       - name: Build and Push Web
         id: build_web
         run: |
           set -o pipefail
-          DOCKER_BUILDKIT=0 docker build \
-            --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:${{ env.IMAGE_TAG }} \
-            --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:latest \
-            ./web 2>&1 | tee -a /tmp/build.log
-          docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:${{ env.IMAGE_TAG }}
-          docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:latest
+          for attempt in 1 2 3; do
+            echo "Build web attempt $attempt/3..."
+            DOCKER_BUILDKIT=0 docker build \
+              --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:${{ env.IMAGE_TAG }} \
+              --tag ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:latest \
+              ./web 2>&1 | tee -a /tmp/build.log && break
+            echo "Attempt $attempt failed, retrying in 10s..." && sleep 10
+          done
+          for attempt in 1 2 3; do
+            docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:${{ env.IMAGE_TAG }} && \
+            docker push ${{ env.CR_SERVER_ACTIVE }}/${{ env.CR_ORG }}/video-web:latest && break
+            echo "Push attempt $attempt failed, retrying in 10s..." && sleep 10
+          done
 
       - name: Setup Kubectl
         run: |
           if ! command -v kubectl &>/dev/null; then
-            curl -LO "https://mirrors.aliyun.com/kubernetes/kubectl/v1.28.0/bin/linux/amd64/kubectl"
+            for attempt in 1 2 3; do
+              curl -LO "https://files.m.daocloud.io/dl.k8s.io/release/v1.28.0/bin/linux/amd64/kubectl" && break
+              echo "Download attempt $attempt failed, retrying in 5s..." && sleep 5
+            done
             chmod +x kubectl && mv kubectl /usr/local/bin/
           fi
           kubectl version --client
@@ -79,11 +96,13 @@ jobs:
         run: |
           mkdir -p $HOME/.kube
           if [[ "${{ github.ref_name }}" == "master" ]]; then
-            echo "${{ secrets.VOLCANO_PROD_KUBE_CONFIG }}" > $HOME/.kube/config
+            printf '%s\n' '${{ secrets.VOLCANO_PROD_KUBE_CONFIG }}' > $HOME/.kube/config
           elif [[ "${{ github.ref_name }}" == "dev" ]]; then
-            echo "${{ secrets.VOLCANO_TEST_KUBE_CONFIG }}" > $HOME/.kube/config
+            printf '%s\n' '${{ secrets.VOLCANO_TEST_KUBE_CONFIG }}' > $HOME/.kube/config
           fi
           chmod 600 $HOME/.kube/config
+          echo "kubeconfig lines: $(wc -l < $HOME/.kube/config)"
+          grep server $HOME/.kube/config || echo "WARNING: no server found in kubeconfig"
 
       - name: Deploy to K3s
         id: deploy
@@ -113,38 +132,43 @@ jobs:
           sed -i "s|redis://zyc:Zyc188208@redis-shzlsczo52dft8mia.redis.ivolces.com:6379/0|${{ env.REDIS_URL }}|g" k8s/backend-deployment.yaml
           sed -i "s|redis://zyc:Zyc188208@redis-shzlsczo52dft8mia.redis.ivolces.com:6379/0|${{ env.REDIS_URL }}|g" k8s/celery-deployment.yaml
 
-          # Create/update image pull secret for CR
-          kubectl create secret docker-registry cr-pull-secret \
-            --docker-server="${{ env.CR_SERVER_ACTIVE }}" \
-            --docker-username="${{ env.CR_USERNAME_ACTIVE }}" \
-            --docker-password="${{ env.CR_PASSWORD_ACTIVE }}" \
-            --dry-run=client -o yaml | kubectl apply -f -
+          # All kubectl operations with retry (K3s 内网连接可能抖动)
+          for attempt in 1 2 3; do
+            echo "Deploy attempt $attempt/3..."
+            {
+              # Create/update image pull secret for CR
+              kubectl create secret docker-registry cr-pull-secret \
+                --docker-server="${{ env.CR_SERVER_ACTIVE }}" \
+                --docker-username="${{ env.CR_USERNAME_ACTIVE }}" \
+                --docker-password="${{ env.CR_PASSWORD_ACTIVE }}" \
+                --dry-run=client -o yaml | kubectl apply -f -
 
-          # Create/update secrets (业务密钥，DB 已写在 yaml 里)
-          kubectl create secret generic video-backend-secrets \
-            --from-literal=ARK_API_KEY='${{ secrets.ARK_API_KEY }}' \
-            --from-literal=TOS_ACCESS_KEY='${{ secrets.TOS_ACCESS_KEY }}' \
-            --from-literal=TOS_SECRET_KEY='${{ secrets.TOS_SECRET_KEY }}' \
-            --from-literal=DJANGO_SECRET_KEY='${{ secrets.DJANGO_SECRET_KEY }}' \
-            --from-literal=ALIYUN_SMS_ACCESS_KEY='${{ secrets.ALIYUN_SMS_ACCESS_KEY }}' \
-            --from-literal=ALIYUN_SMS_ACCESS_SECRET='${{ secrets.ALIYUN_SMS_ACCESS_SECRET }}' \
-            --dry-run=client -o yaml | kubectl apply -f -
+              # Create/update secrets (业务密钥，DB 已写在 yaml 里)
+              kubectl create secret generic video-backend-secrets \
+                --from-literal=ARK_API_KEY='${{ secrets.ARK_API_KEY }}' \
+                --from-literal=TOS_ACCESS_KEY='${{ secrets.TOS_ACCESS_KEY }}' \
+                --from-literal=TOS_SECRET_KEY='${{ secrets.TOS_SECRET_KEY }}' \
+                --from-literal=DJANGO_SECRET_KEY='${{ secrets.DJANGO_SECRET_KEY }}' \
+                --from-literal=ALIYUN_SMS_ACCESS_KEY='${{ secrets.ALIYUN_SMS_ACCESS_KEY }}' \
+                --from-literal=ALIYUN_SMS_ACCESS_SECRET='${{ secrets.ALIYUN_SMS_ACCESS_SECRET }}' \
+                --dry-run=client -o yaml | kubectl apply -f -
 
-          # Apply manifests
-          set -o pipefail
-          {
-            kubectl apply -f k8s/backend-deployment.yaml
-            kubectl apply -f k8s/celery-deployment.yaml
-            kubectl apply -f k8s/web-deployment.yaml
-            kubectl apply -f k8s/ingress.yaml
+              # Apply manifests
+              kubectl apply -f k8s/backend-deployment.yaml
+              kubectl apply -f k8s/celery-deployment.yaml
+              kubectl apply -f k8s/web-deployment.yaml
+              kubectl apply -f k8s/ingress.yaml
 
-            # Preserve real client IP
-            kubectl patch svc traefik -n kube-system -p '{"spec":{"externalTrafficPolicy":"Local"}}' 2>/dev/null || true
+              # Preserve real client IP
+              kubectl patch svc traefik -n kube-system -p '{"spec":{"externalTrafficPolicy":"Local"}}' 2>/dev/null || true
 
-            kubectl rollout restart deployment/video-backend
-            kubectl rollout restart deployment/celery-worker
-            kubectl rollout restart deployment/video-web
-          } 2>&1 | tee /tmp/deploy.log
+              kubectl rollout restart deployment/video-backend
+              kubectl rollout restart deployment/celery-worker
+              kubectl rollout restart deployment/video-web
+            } 2>&1 | tee /tmp/deploy.log && break
+            echo "Attempt $attempt failed, retrying in 10s..."
+            sleep 10
+          done
 
       # ===== Log Center: failure reporting =====
       - name: Report failure to Log Center
@@ -204,3 +228,13 @@ jobs:
                 \"run_url\": \"https://gitea.airlabs.art/${{ github.repository }}/actions/runs/${{ github.run_number }}\"
               }
             }" || true
+
+      # ===== Cleanup: remove unused Docker resources =====
+      - name: Docker Cleanup
+        if: always()
+        run: |
+          docker container prune -f
+          docker image prune -a -f
+          docker builder prune -a -f
+          echo "Disk usage after cleanup:"
+          df -h / | tail -1

backend/Dockerfile

@@ -30,4 +30,4 @@ RUN chmod +x /app/entrypoint.sh
 EXPOSE 8000
 
 ENTRYPOINT ["/app/entrypoint.sh"]
-CMD ["gunicorn", "--bind", "0.0.0.0:8000", "--workers", "2", "--timeout", "120", "--access-logfile", "-", "--error-logfile", "-", "config.wsgi:application"]
+CMD ["gunicorn", "--bind", "0.0.0.0:8000", "--workers", "2", "--worker-class", "gevent", "--worker-connections", "200", "--timeout", "120", "--access-logfile", "-", "--error-logfile", "-", "config.wsgi:application"]

backend/apps/generation/tasks.py

@@ -21,19 +21,29 @@ def poll_video_task(self, record_id):
     from apps.generation.models import GenerationRecord
     from utils.airdrama_client import query_task, map_status
 
+    # 防重复：同一 record 同一时刻只允许一个 poll 在执行
+    from django.core.cache import cache
+    lock_key = f'poll_lock:{record_id}'
+    if not cache.add(lock_key, '1', timeout=POLL_INTERVAL * 3):
+        logger.info('poll_video_task: record %s already being polled, skipping', record_id)
+        return
+
     try:
         record = GenerationRecord.objects.get(pk=record_id)
     except GenerationRecord.DoesNotExist:
         logger.warning('poll_video_task: record %s not found', record_id)
+        cache.delete(lock_key)
         return
 
     ark_task_id = record.ark_task_id
     if not ark_task_id:
         logger.warning('poll_video_task: record %s has no ark_task_id', record_id)
+        cache.delete(lock_key)
         return
 
     if record.status not in ('queued', 'processing'):
         logger.info('poll_video_task: record %s already in terminal state: %s', record_id, record.status)
+        cache.delete(lock_key)
         return
 
     # Poll Volcano API
@@ -42,12 +52,14 @@ def poll_video_task(self, record_id):
         new_status = map_status(ark_resp.get('status', ''))
     except Exception:
         logger.exception('poll_video_task: API query failed for %s, will retry', ark_task_id)
+        cache.delete(lock_key)
         raise self.retry(countdown=POLL_INTERVAL)
 
     if new_status in ('queued', 'processing'):
         # Still running — update status, then re-enqueue
         record.status = new_status
         record.save(update_fields=['status', 'updated_at'])
+        cache.delete(lock_key)
         raise self.retry(countdown=POLL_INTERVAL)
 
     # Terminal state reached — process result

backend/config/settings.py

@@ -189,7 +189,7 @@ CELERY_BEAT_SCHEDULE = {
 LANGUAGE_CODE = 'zh-hans'
 TIME_ZONE = 'Asia/Shanghai'
 USE_I18N = True
-USE_TZ = True
+USE_TZ = False
 
 STATIC_URL = 'static/'
 STATIC_ROOT = BASE_DIR / 'staticfiles'