feat(deploy): #1978 GPU节点自动部署配置文件入库
CI/CD Pipeline / Dedup Check - skip PR tests when covered by push pipeline (pull_request) Successful in 2s
CI/CD Pipeline / Check push changed paths (pull_request) Has been skipped
CI/CD Pipeline / Check if frontend-only change (pull_request) Successful in 2s
PR Automation / Auto Approve on CI Green (pull_request) Successful in 2m53s
CI/CD Pipeline / Frontend Lint (pull_request) Has been skipped
CI/CD Pipeline / Frontend Unit Tests (pull_request) Has been skipped
CI/CD Pipeline / PR Build API Image (pull_request) Successful in 1m17s
CI/CD Pipeline / PR Build Web Image (pull_request) Has been skipped
CI/CD Pipeline / PR Build Worker Image (pull_request) Successful in 1m19s
CI/CD Pipeline / Build Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Build Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Build Staging Worker Image (pull_request) Has been skipped
AI Code Review / AI Code Review (pull_request) Successful in 7m47s
CI/CD Pipeline / Retag skipped Staging API Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Web Image (pull_request) Has been skipped
CI/CD Pipeline / Retag skipped Staging Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Deploy Staging (Watchtower auto-deploy) (pull_request) Has been skipped
CI/CD Pipeline / Staging E2E Tests (pull_request) Has been skipped
CI/CD Pipeline / Staging API Integration Tests (pull_request) Has been skipped
CI/CD Pipeline / ACR Image Cleanup (pull_request) Has been skipped
Preview Deploy / Deploy Preview Environment (pull_request) Successful in 8m44s
PR Automation / Auto Merge on CI Green + Approved (pull_request) Successful in 10m37s
CI/CD Pipeline / Validate - Style (pull_request) Successful in 29m54s
CI/CD Pipeline / Unit Tests (pull_request) Failing after 28m18s
CI/CD Pipeline / Integration Tests (pull_request) Successful in 27m55s
ACR Cleanup / ACR Image Cleanup (pull_request_target) Successful in 24s
Preview Cleanup / Cleanup Preview Environment (pull_request) Successful in 7m43s
CI/CD Pipeline / Validate - Python (mypy + alembic) (pull_request) Successful in 40m20s
CI/CD Pipeline / Validate - Security (pull_request) Successful in 1h36m46s
CI/CD Pipeline / Build Production API Image (pull_request) Has been skipped
CI/CD Pipeline / Build Production Web Image (pull_request) Has been skipped
CI/CD Pipeline / Build Production Worker Image (pull_request) Has been skipped
CI/CD Pipeline / Deploy Production (pull_request) Has been skipped
CI/CD Pipeline / Canary Release to Production (pull_request) Has been skipped
CI/CD Pipeline / CI Gate (pull_request) Failing after 4s
CI/CD Pipeline / Production Browser E2E (pull_request) Has been skipped

新增 deploy/gpu_worker/ 下 3 个 systemd 单元 + 3 个运维脚本:
- musetalk-worker.service: MuseTalk 推理 API systemd 服务
- gpu-poll.service + gpu-poll.timer: 每30秒轮询Gitea最新commit
- scripts/update-gpu-worker.sh: 备份→拉代码→重启→健康检查→失败回滚
- scripts/poll_and_update.sh: SHA比对触发update
- scripts/setup-gpu-node.sh: 新节点一键初始化(apt依赖+目录+systemd+sudo免密+首次启动)
- README.md 追加第七章「自动部署」说明服务架构/部署步骤/更新机制/日志/注意事项

不改动现有 gpu_worker.py/musetalk_server.py/requirements.txt/xiaoxia-gpu-worker.service/.env.example
不改动CI/Docker/镜像构建。脚本路径写死/home/ying,后续多节点再参数化。
This commit is contained in:
xiaoxia
2026-09-20 10:28:24 +08:00
parent 60cacdf280
commit db621b4fcb
7 changed files with 337 additions and 0 deletions
@@ -0,0 +1,47 @@
#!/bin/bash
REPO_API="https://git.xiaoxiajianji.com/api/v1/repos/xiaoxia/xiaoxia-saas/commits?sha=develop&path=deploy/gpu_worker&limit=1"
STATE_FILE="/home/ying/projects/gpu-webhook/.last_commit"
UPDATE_SCRIPT="/home/ying/projects/update-gpu-worker.sh"
LOG_FILE="/tmp/gpu-poll.log"
log() {
echo "[$(date +"%Y-%m-%d %H:%M:%S")] $*" >> "$LOG_FILE"
}
LATEST_SHA=$(curl -sk --max-time 10 "$REPO_API" | python3 -c "
import sys, json
try:
data = json.load(sys.stdin)
if isinstance(data, list) and len(data) > 0:
print(data[0].get('sha', ''))
else:
print('')
except:
print('')
" 2>/dev/null)
if [ -z "$LATEST_SHA" ]; then
log "get latest commit failed, skip"
exit 0
fi
LAST_SHA=""
if [ -f "$STATE_FILE" ]; then
LAST_SHA=$(cat "$STATE_FILE")
fi
if [ "$LATEST_SHA" = "$LAST_SHA" ]; then
exit 0
fi
if [ -z "$LAST_SHA" ]; then
echo "$LATEST_SHA" > "$STATE_FILE"
log "first run, recording SHA: $LATEST_SHA"
exit 0
fi
log "new commit detected: $LAST_SHA -> $LATEST_SHA, triggering update"
echo "$LATEST_SHA" > "$STATE_FILE"
bash "$UPDATE_SCRIPT" >> "$LOG_FILE" 2>&1
log "update completed"
@@ -0,0 +1,62 @@
#!/bin/bash
# GPU节点一键初始化脚本 - 在全新GPU机器上执行
set -e
echo "=== 1. 安装系统依赖 ==="
sudo apt-get update -qq
sudo apt-get install -y -qq python3 python3-pip python3-venv ffmpeg wget curl git
echo "=== 2. 创建目录 ==="
mkdir -p ~/projects/MuseTalk ~/projects/gpu-webhook /opt/xiaoxia-gpu-worker
echo "=== 3. 安装nvidia-container-toolkit(如需要Docker)==="
# 可选,当前不使用Docker,跳过
# distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
# curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
# curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
# sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit
# sudo nvidia-ctk runtime configure --runtime=docker
# sudo systemctl restart docker
echo "=== 4. 拉取服务配置和脚本 ==="
REPO_URL="https://git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas/raw/branch/develop/deploy/gpu_worker"
wget -q -O /tmp/musetalk-worker.service "$REPO_URL/musetalk-worker.service"
wget -q -O /tmp/gpu-poll.service "$REPO_URL/gpu-poll.service"
wget -q -O /tmp/gpu-poll.timer "$REPO_URL/gpu-poll.timer"
wget -q -O ~/projects/update-gpu-worker.sh "$REPO_URL/scripts/update-gpu-worker.sh"
wget -q -O ~/projects/gpu-webhook/poll_and_update.sh "$REPO_URL/scripts/poll_and_update.sh"
chmod +x ~/projects/update-gpu-worker.sh ~/projects/gpu-webhook/poll_and_update.sh
echo "=== 5. 安装systemd服务 ==="
sudo cp /tmp/musetalk-worker.service /etc/systemd/system/
sudo cp /tmp/gpu-poll.service /etc/systemd/system/
sudo cp /tmp/gpu-poll.timer /etc/systemd/system/
echo "=== 6. 配置sudo免密 ==="
sudo bash -c 'cat > /etc/sudoers.d/ying-gpu-update << EOF
ying ALL=(ALL) NOPASSWD: /bin/systemctl restart musetalk-worker
ying ALL=(ALL) NOPASSWD: /bin/systemctl restart xiaoxia-gpu-worker
ying ALL=(ALL) NOPASSWD: /bin/systemctl status musetalk-worker
ying ALL=(ALL) NOPASSWD: /bin/systemctl status xiaoxia-gpu-worker
ying ALL=(ALL) NOPASSWD: /bin/systemctl daemon-reload
ying ALL=(ALL) NOPASSWD: /usr/bin/journalctl
ying ALL=(ALL) NOPASSWD: /bin/cp
ying ALL=(ALL) NOPASSWD: /bin/chmod
ying ALL=(ALL) NOPASSWD: /usr/bin/tee
EOF'
sudo chmod 440 /etc/sudoers.d/ying-gpu-update
echo "=== 7. 首次拉取代码并启动服务 ==="
bash ~/projects/update-gpu-worker.sh
sudo systemctl daemon-reload
sudo systemctl enable musetalk-worker xiaoxia-gpu-worker gpu-poll.timer
sudo systemctl start musetalk-worker xiaoxia-gpu-worker gpu-poll.timer
echo "=== 完成! ==="
echo "检查服务状态:"
echo " sudo systemctl status musetalk-worker"
echo " sudo systemctl status xiaoxia-gpu-worker"
echo " sudo systemctl status gpu-poll.timer"
echo "健康检查:curl http://127.0.0.1:7861/health"
echo "更新日志:tail -f /tmp/gpu-worker-update.log"
@@ -0,0 +1,62 @@
#!/bin/bash
set -e
REPO_URL="https://git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas/raw/branch/develop/deploy/gpu_worker"
MUSE_DIR="/home/ying/projects/MuseTalk"
WORKER_DIR="/opt/xiaoxia-gpu-worker"
LOG_FILE="/tmp/gpu-worker-update.log"
log() {
local NOW
NOW=$(date +"%Y-%m-%d %H:%M:%S")
echo "[$NOW] $*" | tee -a "$LOG_FILE"
}
log "========== start update =========="
BAK_SUFFIX=$(date +"%Y%m%d%H%M%S")
cp "$MUSE_DIR/musetalk_server.py" "$MUSE_DIR/musetalk_server.py.bak.$BAK_SUFFIX"
cp "$WORKER_DIR/gpu_worker.py" "$WORKER_DIR/gpu_worker.py.bak.$BAK_SUFFIX"
log "backup done ($BAK_SUFFIX)"
wget -q -O "$MUSE_DIR/musetalk_server.py" "$REPO_URL/musetalk_server.py"
log "musetalk_server.py updated"
wget -q -O "$WORKER_DIR/gpu_worker.py" "$REPO_URL/gpu_worker.py"
log "gpu_worker.py updated"
wget -q -O /tmp/gpu-requirements.txt "$REPO_URL/requirements.txt"
if [ -f "$WORKER_DIR/requirements.txt" ] && ! diff -q "$WORKER_DIR/requirements.txt" /tmp/gpu-requirements.txt > /dev/null 2>&1; then
log "requirements changed, updating..."
cp /tmp/gpu-requirements.txt "$WORKER_DIR/requirements.txt"
"$WORKER_DIR/venv/bin/pip" install -r "$WORKER_DIR/requirements.txt" -q
log "pip install done"
else
log "requirements no change, skip pip"
fi
sudo systemctl restart musetalk-worker
log "musetalk restarted"
sleep 5
sudo systemctl restart xiaoxia-gpu-worker
log "gpu-worker restarted"
sleep 8
HEALTH=$(curl -s http://127.0.0.1:7861/health 2>/dev/null)
if echo "$HEALTH" | grep -q "healthy\|ok"; then
log "health check OK"
log "========== update done =========="
exit 0
else
log "health check FAILED, rolling back..."
LATEST_MUSE_BAK=$(ls -t "$MUSE_DIR/musetalk_server.py.bak."* 2>/dev/null | head -1)
LATEST_WORKER_BAK=$(ls -t "$WORKER_DIR/gpu_worker.py.bak."* 2>/dev/null | head -1)
[ -n "$LATEST_MUSE_BAK" ] && cp "$LATEST_MUSE_BAK" "$MUSE_DIR/musetalk_server.py"
[ -n "$LATEST_WORKER_BAK" ] && cp "$LATEST_WORKER_BAK" "$WORKER_DIR/gpu_worker.py"
sudo systemctl restart musetalk-worker
sleep 5
sudo systemctl restart xiaoxia-gpu-worker
log "rolled back"
exit 1
fi