xiaoxia/xiaoxia-saas
1
0
Fork 0
Code Issues 11 Pull Requests 2 Actions Packages Projects Releases 7 Wiki Activity

ci: Staging部署并行优化，镜像pull+容器启动+健康检查全部并行 #696

Merged
xiaoxia merged 1 commits from ci/deploy-staging-parallel into develop 2026-07-22 08:42:38 +08:00
Conversation 5 Commits 1 Files Changed 1
+151 -93
1 changed files with 151 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/ci_staging_deploy.sh
+151 -93
View File
@@ -1,22 +1,9 @@
#!/bin/sh
# ===========================================
# Staging 部署脚本（SSH 模式，支持自动回滚）
# Staging 部署脚本（SSH 模式，并行优化版）
# ===========================================
# 通过 SSH 在 staging 服务器上执行
#
# 环境变量:
# IMAGE_TAG - 镜像版本 tag（如 commit SHA 或分支名）
# REGISTRY_TOKEN - Registry 访问令牌
# REGISTRY - Registry 地址（默认 xiaoxia-registry.cn-hangzhou.cr.aliyuncs.com/xiaoxiakeji）
# REGISTRY_USER - Registry 用户名（默认 xiaoxia）
# ENV_FILE - 环境变量文件路径
# GENERATED_DIR - 生成文件目录
# SKIP_MIGRATION - 跳过数据库迁移（true/false，默认 false）
# SKIP_ROLLBACK - 失败时跳过自动回滚（true/false，默认 false）
set -eu
# ---- 重试工具函数 ----
retry_cmd() {
local max_attempts=$1
local backoff=$2
@@ -73,10 +60,9 @@ mkdir -p "$GENERATED_DIR"
mkdir -p "$LEGACY_ASSETS_DIR"
echo "==========================================="
echo " Staging 部署 - $IMAGE_TAG"
echo " Staging 部署 - $IMAGE_TAG (并行优化版)"
echo "==========================================="
# ---- 记录当前运行的镜像版本（用于回滚） ----
echo "Recording current image versions for rollback..."
PREV_API_IMAGE=""
PREV_WORKER_IMAGE=""
@@ -95,7 +81,6 @@ for c in xiaoxia-api-staging xiaoxia-worker-staging xiaoxia-web-staging; do
fi
done
# ---- 回滚函数 ----
rollback() {
echo ""
echo "!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
@@ -108,7 +93,6 @@ rollback() {
exit 1
fi
# 停止当前（失败的）新容器
echo "Stopping new containers..."
docker rm -f xiaoxia-api-staging 2>/dev/null || true
docker rm -f xiaoxia-worker-staging 2>/dev/null || true
@@ -116,7 +100,6 @@ rollback() {
LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
# 恢复 API
if [ -n "$PREV_API_IMAGE" ]; then
echo "Rolling back API to: $PREV_API_IMAGE"
docker run -d \
@@ -137,12 +120,9 @@ rollback() {
--health-retries 3 \
--health-start-period 40s \
$LOG_OPTS \
"$PREV_API_IMAGE"
else
echo "No previous API image to roll back to"
"$PREV_API_IMAGE" &
fi
# 恢复 Worker
if [ -n "$PREV_WORKER_IMAGE" ]; then
echo "Rolling back Worker to: $PREV_WORKER_IMAGE"
docker run -d \
@@ -164,12 +144,9 @@ rollback() {
--health-retries 3 \
--health-start-period 30s \
$LOG_OPTS \
"$PREV_WORKER_IMAGE"
else
echo "No previous Worker image to roll back to"
"$PREV_WORKER_IMAGE" &
fi
# 恢复 Web
if [ -n "$PREV_WEB_IMAGE" ]; then
echo "Rolling back Web to: $PREV_WEB_IMAGE"
LEGACY_VOLUME=""
@@ -187,12 +164,11 @@ rollback() {
--health-timeout 5s \
--health-retries 3 \
$LOG_OPTS \
"$PREV_WEB_IMAGE"
else
echo "No previous Web image to roll back to"
"$PREV_WEB_IMAGE" &
fi
# 等待 API 回滚后恢复健康
wait
if [ -n "$PREV_API_IMAGE" ]; then
echo "Waiting for rolled-back API to become healthy..."
i=0
@@ -224,7 +200,6 @@ rollback() {
exit 1
}
# ---- 登录 Registry ----
if [ -n "$REGISTRY_TOKEN" ]; then
echo "=========================================="
echo " Login to Registry (with retries)"
@@ -233,28 +208,64 @@ if [ -n "$REGISTRY_TOKEN" ]; then
retry_docker_login
fi
# ---- Pull 新版本镜像 ----
# ---- 并行 Pull 三个镜像 ----
REGISTRY_API="${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}"
REGISTRY_WORKER="${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}"
REGISTRY_WEB="${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}"
echo "=========================================="
echo " Pull images (with retries)"
echo " Pull images (parallel, up to 3 retries each)"
echo "=========================================="
retry_docker_pull "$REGISTRY_API"
retry_docker_pull "$REGISTRY_WORKER"
retry_docker_pull "$REGISTRY_WEB"
PULL_LOG_DIR="/tmp/staging-pull-$$"
mkdir -p "$PULL_LOG_DIR"
retry_docker_pull "$REGISTRY_API" > "$PULL_LOG_DIR/api.log" 2>&1 &
PID_API=$!
retry_docker_pull "$REGISTRY_WORKER" > "$PULL_LOG_DIR/worker.log" 2>&1 &
PID_WORKER=$!
retry_docker_pull "$REGISTRY_WEB" > "$PULL_LOG_DIR/web.log" 2>&1 &
PID_WEB=$!
wait $PID_API $PID_WORKER $PID_WEB
echo ""
echo "Pull 结果:"
PULL_FAILED=0
for svc in api worker web; do
if tail -1 "$PULL_LOG_DIR/$svc.log" 2>/dev/null | grep -qE "Status:|Downloaded|already exists|is up to date"; then
echo " OK $svc"
elif grep -qE "Digest:|Status: Downloaded" "$PULL_LOG_DIR/$svc.log" 2>/dev/null; then
echo " OK $svc"
else
# 检查docker pull返回值不直接，用镜像是否存在来判断
img_var="REGISTRY_$(echo $svc | tr '[:lower:]' '[:upper:]')"
img_val=$(eval echo "\$$img_var")
if docker image inspect "$img_val" >/dev/null 2>&1; then
echo " OK $svc"
else
echo " FAIL $svc"
tail -5 "$PULL_LOG_DIR/$svc.log" 2>/dev/null || true
PULL_FAILED=$((PULL_FAILED + 1))
fi
fi
done
rm -rf "$PULL_LOG_DIR"
if [ "$PULL_FAILED" -gt 0 ]; then
echo ""
echo "ERROR: $PULL_FAILED 个镜像 pull 失败"
exit 1
fi
echo "All images pulled."
# ---- 备份 legacy assets ----
echo "Backing up legacy assets from current web container..."
if docker inspect xiaoxia-web-staging >/dev/null 2>&1; then
_tmpdir="/tmp/legacy-assets-$$"
rm -rf "$_tmpdir"
mkdir -p "$_tmpdir"
docker cp xiaoxia-web-staging:/usr/share/nginx/html/assets/. "$_tmpdir/" 2>/dev/null || true
# 只有目录非空才拷贝，避免覆盖有内容的 legacy assets
if [ -d "$_tmpdir" ] && [ "$(ls -A "$_tmpdir" 2>/dev/null)" ]; then
cp -an "$_tmpdir"/. "$LEGACY_ASSETS_DIR"/ 2>/dev/null || true
echo "Legacy assets backed up: $(ls "$_tmpdir" | wc -l) files"
@@ -264,13 +275,11 @@ else
echo "No existing web container, skipping legacy assets backup"
fi
# 清理 7 天前的 legacy assets
if [ -d "$LEGACY_ASSETS_DIR" ]; then
find "$LEGACY_ASSETS_DIR" -type f -mtime +7 -delete 2>/dev/null || true
echo "Legacy assets cleanup done (retain 7 days)"
fi
# ---- 检查基础设施容器 ----
echo "Checking infrastructure containers..."
for c in xiaoxia-postgres-staging xiaoxia-redis-staging; do
if ! docker inspect "$c" >/dev/null 2>&1; then
@@ -284,10 +293,8 @@ for c in xiaoxia-postgres-staging xiaoxia-redis-staging; do
fi
done
# ---- 创建网络（不存在则创建） ----
docker network create xiaoxia-net-staging 2>/dev/null || true
# ---- 数据库迁移 ----
if [ "$SKIP_MIGRATION" != "true" ]; then
echo "Running database migrations..."
docker run --rm \
@@ -296,8 +303,6 @@ if [ "$SKIP_MIGRATION" != "true" ]; then
-e APP_ENV=staging \
"$REGISTRY_API" sh -c "cd /app && alembic upgrade head" || {
echo "ERROR: Database migration failed"
echo "Note: Migration failures are NOT automatically rolled back (data safety)"
echo "Please manually check and fix the migration, then redeploy"
exit 1
}
echo "Migrations completed."
@@ -305,7 +310,6 @@ else
echo "Skipping migrations (SKIP_MIGRATION=true)"
fi
# ---- 停止旧容器 ----
echo "Stopping old containers..."
docker rm -f xiaoxia-api-staging 2>/dev/null || true
docker rm -f xiaoxia-worker-staging 2>/dev/null || true
@@ -313,8 +317,14 @@ docker rm -f xiaoxia-web-staging 2>/dev/null || true
LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
# ---- 启动 API ----
echo "Starting API container..."
# ---- 并行启动三个容器 ----
echo "Starting all containers (parallel)..."
LEGACY_VOLUME=""
if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
fi
docker run -d \
--name xiaoxia-api-staging \
--env-file "$ENV_FILE" \
@@ -333,10 +343,9 @@ docker run -d \
--health-retries 3 \
--health-start-period 40s \
$LOG_OPTS \
"$REGISTRY_API" || rollback
"$REGISTRY_API" &
PID_API_START=$!
# ---- 启动 Worker ----
echo "Starting Worker container..."
docker run -d \
--name xiaoxia-worker-staging \
--env-file "$ENV_FILE" \
@@ -356,18 +365,9 @@ docker run -d \
--health-retries 3 \
--health-start-period 30s \
$LOG_OPTS \
"$REGISTRY_WORKER" || rollback
"$REGISTRY_WORKER" &
PID_WORKER_START=$!
# ---- 启动 Web ----
LEGACY_VOLUME=""
if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
echo "Web container: legacy assets mounted (fallback)"
else
echo "Web container: no legacy assets to mount"
fi
echo "Starting Web container..."
docker run -d \
--name xiaoxia-web-staging \
--network xiaoxia-net-staging \
@@ -379,53 +379,111 @@ docker run -d \
--health-timeout 5s \
--health-retries 3 \
$LOG_OPTS \
"$REGISTRY_WEB" || rollback
"$REGISTRY_WEB" &
PID_WEB_START=$!
# ---- 等待 API 健康 ----
echo "Waiting for API to become healthy..."
i=0
while [ "$i" -lt 40 ]; do
if curl -sf --max-time 5 http://127.0.0.1:8000/health >/dev/null 2>&1; then
echo "API is healthy!"
break
wait $PID_API_START $PID_WORKER_START $PID_WEB_START
START_FAILED=0
for c in xiaoxia-api-staging xiaoxia-worker-staging xiaoxia-web-staging; do
if ! docker inspect "$c" >/dev/null 2>&1; then
echo " FAIL $c: not created"
START_FAILED=$((START_FAILED + 1))
else
state=$(docker inspect -f '{{.State.Status}}' "$c")
if [ "$state" = "running" ] || [ "$state" = "starting" ]; then
echo " OK $c: $state"
else
echo " FAIL $c: $state"
docker logs --tail 20 "$c" 2>/dev/null || true
START_FAILED=$((START_FAILED + 1))
fi
fi
i=$((i + 1))
echo " Waiting... ($i/40)"
sleep 3
done
if [ "$i" -ge 40 ]; then
echo "ERROR: API did not become healthy within 120s"
if [ "$START_FAILED" -gt 0 ]; then
echo "ERROR: $START_FAILED 个容器启动失败"
rollback
fi
# ---- 并行等待 API 和 Web 健康 ----
echo ""
echo "Waiting for API + Web health (parallel)..."
HEALTH_LOG_DIR="/tmp/staging-health-$$"
mkdir -p "$HEALTH_LOG_DIR"
(
i=0
while [ "$i" -lt 40 ]; do
if curl -sf --max-time 5 http://127.0.0.1:8000/health >/dev/null 2>&1; then
echo "API healthy after $((i * 3))s"
exit 0
fi
i=$((i + 1))
sleep 3
done
echo "API FAILED after 120s"
exit 1
) > "$HEALTH_LOG_DIR/api.log" 2>&1 &
PID_API_HEALTH=$!
(
i=0
while [ "$i" -lt 15 ]; do
if curl -sf --max-time 5 http://127.0.0.1:3001/ >/dev/null 2>&1; then
echo "Web healthy after $((i * 2))s"
exit 0
fi
i=$((i + 1))
sleep 2
done
echo "Web FAILED after 30s"
exit 1
) > "$HEALTH_LOG_DIR/web.log" 2>&1 &
PID_WEB_HEALTH=$!
set +e
wait $PID_API_HEALTH
API_EXIT=$?
wait $PID_WEB_HEALTH
WEB_EXIT=$?
set -e
echo ""
echo "健康检查结果:"
API_OK=0
WEB_OK=0
if [ "$API_EXIT" -eq 0 ]; then
echo " OK API: $(cat "$HEALTH_LOG_DIR/api.log")"
API_OK=1
else
echo " FAIL API: 120s未就绪"
docker logs --tail 50 xiaoxia-api-staging
rollback
fi
# ---- 等待 Web 健康 ----
echo "Waiting for Web to become healthy..."
i=0
while [ "$i" -lt 15 ]; do
if curl -sf --max-time 5 http://127.0.0.1:3001/ >/dev/null 2>&1; then
echo "Web is healthy!"
break
fi
i=$((i + 1))
echo " Waiting... ($i/15)"
sleep 2
done
if [ "$i" -ge 15 ]; then
echo "ERROR: Web did not become healthy within 30s"
if [ "$WEB_EXIT" -eq 0 ]; then
echo " OK Web: $(cat "$HEALTH_LOG_DIR/web.log")"
WEB_OK=1
else
echo " FAIL Web: 30s未就绪"
docker logs --tail 30 xiaoxia-web-staging
fi
rm -rf "$HEALTH_LOG_DIR"
if [ "$API_OK" -eq 0 ] || [ "$WEB_OK" -eq 0 ]; then
echo ""
echo "ERROR: 健康检查失败"
rollback
fi
# ---- 清理旧镜像 ----
echo "Cleaning up old images..."
docker image prune -af --filter "until=168h" 2>/dev/null || true
docker builder prune -af --filter "until=168h" 2>/dev/null || true
echo ""
echo "=== Staging deployment complete ==="
echo "=== Staging deployment complete (并行优化版) ==="
echo "API: http://127.0.0.1:8000"
echo "Web: http://127.0.0.1:3001"
echo "Version: $IMAGE_TAG"