From 4cc2ee853d2ab5b16388ec95e38a7ac7f1fc1ec2 Mon Sep 17 00:00:00 2001
From: xiaoxia <xiaoxia@xiaoxia.com>
Date: Mon, 10 Aug 2026 12:48:11 +0800
Subject: [PATCH] fix: add graceful shutdown to CI deploy script

- scripts/ci_staging_deploy.sh: replace docker rm -f with docker stop + rm
  Worker: 300s, API: 30s, Web: 10s
- infra/docker/compose.yml: add stop_grace_period and stop_signal
  for api (30s) and worker (300s)
- infra/docker/entrypoint-worker.sh: add -W (warm shutdown) flag
  so Celery waits for current tasks before exiting on SIGTERM

Rollback function docker rm -f intentionally kept (speed priority)
---
 infra/docker/compose.yml          | 4 ++++
 infra/docker/entrypoint-worker.sh | 3 ++-
 scripts/ci_staging_deploy.sh      | 9 ++++++---
 3 files changed, 12 insertions(+), 4 deletions(-)
 mode change 100644 => 100755 infra/docker/compose.yml
 mode change 100644 => 100755 infra/docker/entrypoint-worker.sh

diff --git a/infra/docker/compose.yml b/infra/docker/compose.yml
old mode 100644
new mode 100755
index b9e7178f5..5c7164c7e
--- a/infra/docker/compose.yml
+++ b/infra/docker/compose.yml
@@ -47,6 +47,8 @@ services:
     
     container_name: xiaoxia-api-${ENV:-staging}
     restart: unless-stopped
+    stop_grace_period: 30s
+    stop_signal: SIGTERM
     
     # 环境变量文件（包含数据库密码等敏感信息）
     env_file:
@@ -102,6 +104,8 @@ services:
     
     container_name: xiaoxia-worker-${ENV:-staging}
     restart: unless-stopped
+    stop_grace_period: 300s
+    stop_signal: SIGTERM
     
     env_file:
       - ../../.env
diff --git a/infra/docker/entrypoint-worker.sh b/infra/docker/entrypoint-worker.sh
old mode 100644
new mode 100755
index d03eb4bd1..9bbb6f898
--- a/infra/docker/entrypoint-worker.sh
+++ b/infra/docker/entrypoint-worker.sh
@@ -10,4 +10,5 @@ exec celery \
   -A worker_app.celery_app \
   worker \
   --loglevel=info \
-  "--concurrency=${CONCURRENCY}"
+  "--concurrency=${CONCURRENCY}" \
+  -W
diff --git a/scripts/ci_staging_deploy.sh b/scripts/ci_staging_deploy.sh
index 71444545f..22a34a502 100755
--- a/scripts/ci_staging_deploy.sh
+++ b/scripts/ci_staging_deploy.sh
@@ -311,9 +311,12 @@ else
 fi
 
 echo "Stopping old containers..."
-docker rm -f xiaoxia-api-staging 2>/dev/null || true
-docker rm -f xiaoxia-worker-staging 2>/dev/null || true
-docker rm -f xiaoxia-web-staging 2>/dev/null || true
+# 优雅关闭：先 stop（发 SIGTERM，等待），再 rm
+# Worker 需要更长时间（视频任务最长可能5分钟）
+docker stop -t 300 xiaoxia-worker-staging 2>/dev/null || true
+docker stop -t 30 xiaoxia-api-staging 2>/dev/null || true
+docker stop -t 10 xiaoxia-web-staging 2>/dev/null || true
+docker rm xiaoxia-worker-staging xiaoxia-api-staging xiaoxia-web-staging 2>/dev/null || true
 
 LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"