start.sh
Rama main · réplica de lectura del repositorio original.
#!/bin/sh
echo "=========================================="
echo " NOWEN - Starting..."
echo "=========================================="
echo "Architecture: $(uname -m)"
echo ""
# ============================================================
# 🔑 防呆自动配置 — 自动检测 Docker 环境并设置关键环境变量
# ============================================================
# 绿联/飞牛等 NAS 的 Docker GUI 用户通常不会手动设置环境变量
# 所以我们在启动时自动检测并设置，确保即使什么都不配也能正常工作
# 自动设置 NODE_ENV=production（如果未设置且检测到在 Docker 容器内）
if [ -z "${NODE_ENV}" ]; then
if [ -f "/.dockerenv" ] || [ -f "/app/start.sh" ] || grep -qsE 'docker|containerd|kubepods' /proc/1/cgroup 2>/dev/null; then
export NODE_ENV=production
echo "🔧 Auto-set NODE_ENV=production (Docker detected)"
fi
fi
# 自动设置 SI_FILESYSTEM_DISK_PREFIX（如果未设置且 /host 挂载存在）
if [ -z "${SI_FILESYSTEM_DISK_PREFIX}" ] && [ -d "/host/proc" ]; then
export SI_FILESYSTEM_DISK_PREFIX=/host
echo "🔧 Auto-set SI_FILESYSTEM_DISK_PREFIX=/host"
fi
# 自动设置 PROC_PATH（如果未设置且 /host/proc 存在）
if [ -z "${PROC_PATH}" ] && [ -d "/host/proc" ]; then
export PROC_PATH=/host/proc
echo "🔧 Auto-set PROC_PATH=/host/proc"
fi
# 自动设置 SYS_PATH（如果未设置且 /host/sys 存在）
if [ -z "${SYS_PATH}" ] && [ -d "/host/sys" ]; then
export SYS_PATH=/host/sys
echo "🔧 Auto-set SYS_PATH=/host/sys"
fi
# 自动设置 FS_PATH（如果未设置且 /host 存在）
if [ -z "${FS_PATH}" ] && [ -d "/host" ]; then
export FS_PATH=/host
echo "🔧 Auto-set FS_PATH=/host"
fi
echo " NODE_ENV=${NODE_ENV:-not set}"
echo ""
# ============================================================
# 🔑 数据持久化 — 终极防呆设计（兼容绿联/飞牛等 NAS）
# ============================================================
#
# ⚠️ 问题根因：
# 绿联/飞牛等 NAS 的 Docker GUI 会为每个容器创建独立的存储目录，
# 路径格式为: 共享文件夹/docker/<容器名>/app_server_data
# 当更新镜像重建容器时，容器名从 xxx-1 变成 xxx-2，
# 新的存储目录是空的，导致数据丢失。
# 更关键的是：如果 Dockerfile 声明了 VOLUME，NAS 会为 VOLUME
# 路径也创建独立目录，安全备份卷也一起丢失。
#
# 解决策略：
# 1. Dockerfile 中不声明 VOLUME（避免 NAS 自动创建新目录）
# 2. 安全备份放在 /app/.nowen-safe（镜像层内路径，NAS 不挂载）
# 3. 启动时扫描宿主机旧容器目录，自动找回数据
# 4. 同时兼容用户手动配置的 bind mount / named volume
# ============================================================
DATA_DIR="/app/server/data"
DB_FILE="${DATA_DIR}/zen-garden.db"
BACKUP_DIR="${DATA_DIR}/backups"
# 🔑 安全备份使用镜像层内路径（不是 VOLUME，NAS 不会挂载它）
SAFE_BACKUP_DIR="/app/.nowen-safe"
SAFE_DB_FILE="${SAFE_BACKUP_DIR}/zen-garden.db.safe"
# 确保目录存在
mkdir -p "${BACKUP_DIR}"
mkdir -p "${SAFE_BACKUP_DIR}"
# ============================================================
# 第一步：挂载检测
# ============================================================
check_mount() {
MOUNT_MARKER="${DATA_DIR}/.mount-marker"
if mount | grep -q "/app/server/data"; then
MOUNT_TYPE=$(mount | grep "/app/server/data" | head -1)
echo "✅ Data volume detected: mounted"
echo " Mount info: ${MOUNT_TYPE}"
return 0
else
if [ -f "${MOUNT_MARKER}" ]; then
echo "✅ Data volume detected: persistent (marker found)"
return 0
else
echo "⚠️ WARNING: Data directory may not be properly mounted!"
echo ""
echo "┌───────────────────────────────────────────────────┐"
echo "│ ⚠️ 数据持久化警告 / Data Persistence Warning │"
echo "├───────────────────────────────────────────────────┤"
echo "│ │"
echo "│ 未检测到数据卷挂载。 │"
echo "│ NOWEN 会自动在镜像内保留备份，但强烈建议 │"
echo "│ 在 NAS Docker 设置中手动指定宿主机存储路径： │"
echo "│ │"
echo "│ 宿主机路径(固定) → /app/server/data │"
echo "│ │"
echo "└───────────────────────────────────────────────────┘"
echo ""
echo "$(date -Iseconds)" > "${MOUNT_MARKER}"
return 1
fi
fi
}
# ============================================================
# 第二步：数据恢复（优先级从高到低）
# ============================================================
recover_database() {
if [ -f "${DB_FILE}" ]; then
DB_SIZE=$(ls -lh "${DB_FILE}" | awk '{print $5}')
echo "✅ Database found: zen-garden.db (${DB_SIZE})"
# 启动前备份到 data/backups/
TIMESTAMP=$(date +%Y%m%d_%H%M%S)
cp "${DB_FILE}" "${BACKUP_DIR}/zen-garden_${TIMESTAMP}.db.bak"
echo " Auto backup: zen-garden_${TIMESTAMP}.db.bak"
# 同步到镜像层内的安全备份
cp "${DB_FILE}" "${SAFE_DB_FILE}"
echo " Safe backup synced to image layer"
# 清理旧备份，只保留最近 5 个
cd "${BACKUP_DIR}" && ls -1t *.db.bak 2>/dev/null | tail -n +6 | xargs rm -f 2>/dev/null
cd /app
return 0
fi
echo "⚠️ Database not found in data directory!"
echo ""
# ============================================================
# 恢复源 1: 镜像层内安全备份（/app/.nowen-safe/）
# 这个路径在镜像层内，NAS 不会为它创建新目录
# ============================================================
if [ -f "${SAFE_DB_FILE}" ]; then
SAFE_SIZE=$(ls -lh "${SAFE_DB_FILE}" | awk '{print $5}')
echo "🔄 Found safe backup in image layer (${SAFE_SIZE}), restoring..."
cp "${SAFE_DB_FILE}" "${DB_FILE}"
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from image-layer safe backup!"
return 0
fi
fi
# ============================================================
# 恢复源 1.5: 兼容旧版安全备份路径（/app/.data-backup/）
# ============================================================
OLD_SAFE_FILE="/app/.data-backup/zen-garden.db.safe"
if [ -f "${OLD_SAFE_FILE}" ]; then
OLD_SAFE_SIZE=$(ls -lh "${OLD_SAFE_FILE}" | awk '{print $5}')
echo "🔄 Found legacy safe backup (${OLD_SAFE_SIZE}), restoring..."
cp "${OLD_SAFE_FILE}" "${DB_FILE}"
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from legacy safe backup!"
return 0
fi
fi
# ============================================================
# 恢复源 2: data/backups/ 目录
# ============================================================
LATEST_BACKUP=$(ls -1t "${BACKUP_DIR}"/*.db.bak 2>/dev/null | head -1)
if [ -n "${LATEST_BACKUP}" ]; then
echo "🔄 Found backup: ${LATEST_BACKUP}"
echo " Restoring from backup..."
cp "${LATEST_BACKUP}" "${DB_FILE}"
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from backup!"
return 0
fi
fi
# ============================================================
# 恢复源 3: 扫描宿主机常见路径 + 绿联/飞牛旧容器目录
# 这是关键！绿联 NAS 的 Docker 数据存储在宿主机的
# /docker/<容器名>/ 目录下，通过 /host 只读挂载可以访问
# ============================================================
echo "🔍 Scanning host filesystem for existing NOWEN data..."
# 3a: 扫描绿联/飞牛等 NAS 的旧容器目录
# 绿联挂载格式: 共享文件夹/docker/<容器名>/app_server_data/zen-garden.db
# 在 /host 挂载下查找
if [ -d "/host" ]; then
echo " Scanning NAS old container directories..."
# 搜索所有可能的 docker 目录
for DOCKER_BASE in \
"/host/docker" \
"/host/DATA/docker" \
"/host/volume1/docker" \
"/host/volume2/docker" \
"/host/mnt/docker" \
"/host/opt/docker" \
"/host/home/docker" \
"/host/mnt/user1/docker" \
"/host/mnt/media_rw"; do
if [ -d "${DOCKER_BASE}" ]; then
# 搜索该 docker 目录下所有包含 nowen 的容器目录
for CONTAINER_DIR in $(find "${DOCKER_BASE}" -maxdepth 2 -type d -iname "*nowen*" 2>/dev/null); do
# 搜索该容器目录下的所有 zen-garden.db 文件
for DB_FOUND in $(find "${CONTAINER_DIR}" -name "zen-garden.db" -type f 2>/dev/null); do
FOUND_SIZE=$(ls -lh "${DB_FOUND}" 2>/dev/null | awk '{print $5}')
# 验证文件大小（大于 100 字节才是有效数据库）
FOUND_BYTES=$(stat -c%s "${DB_FOUND}" 2>/dev/null || stat -f%z "${DB_FOUND}" 2>/dev/null || echo "0")
if [ "${FOUND_BYTES}" -gt 100 ] 2>/dev/null; then
echo "🎉 Found database in old container: ${DB_FOUND} (${FOUND_SIZE})"
cp "${DB_FOUND}" "${DB_FILE}" 2>/dev/null
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from NAS old container directory!"
# 同步到安全备份
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
return 0
fi
fi
done
done
# 也搜索非 nowen 名称但包含 cropflre 的目录
for CONTAINER_DIR in $(find "${DOCKER_BASE}" -maxdepth 2 -type d -iname "*cropflre*" 2>/dev/null); do
for DB_FOUND in $(find "${CONTAINER_DIR}" -name "zen-garden.db" -type f 2>/dev/null); do
FOUND_SIZE=$(ls -lh "${DB_FOUND}" 2>/dev/null | awk '{print $5}')
FOUND_BYTES=$(stat -c%s "${DB_FOUND}" 2>/dev/null || stat -f%z "${DB_FOUND}" 2>/dev/null || echo "0")
if [ "${FOUND_BYTES}" -gt 100 ] 2>/dev/null; then
echo "🎉 Found database in old container: ${DB_FOUND} (${FOUND_SIZE})"
cp "${DB_FOUND}" "${DB_FILE}" 2>/dev/null
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from NAS old container directory!"
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
return 0
fi
fi
done
done
fi
done
# 3b: 扫描传统绑定挂载路径
for SCAN_PATH in \
"/host/volume1/docker/nowen/data/zen-garden.db" \
"/host/volume2/docker/nowen/data/zen-garden.db" \
"/host/DATA/docker/nowen/data/zen-garden.db" \
"/host/mnt/docker/nowen/data/zen-garden.db" \
"/host/opt/docker/nowen/data/zen-garden.db" \
"/host/home/docker/nowen/data/zen-garden.db" \
"/host/root/nowen-data/zen-garden.db" \
"/host/home/nowen-data/zen-garden.db"; do
if [ -f "${SCAN_PATH}" ]; then
FOUND_SIZE=$(ls -lh "${SCAN_PATH}" | awk '{print $5}')
echo "🎉 Found existing database at: ${SCAN_PATH} (${FOUND_SIZE})"
cp "${SCAN_PATH}" "${DB_FILE}"
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from host filesystem!"
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
return 0
fi
fi
done
# 3c: 最后的搜索 — 全局搜索 /host 下所有 zen-garden.db（限时 30 秒）
echo " Deep scanning for zen-garden.db on host (timeout 30s)..."
DEEP_FOUND=$(timeout 30 find /host -name "zen-garden.db" -type f -size +100c 2>/dev/null | head -5)
for DB_FOUND in ${DEEP_FOUND}; do
FOUND_SIZE=$(ls -lh "${DB_FOUND}" 2>/dev/null | awk '{print $5}')
echo "🎉 Found database: ${DB_FOUND} (${FOUND_SIZE})"
cp "${DB_FOUND}" "${DB_FILE}" 2>/dev/null
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from deep host scan!"
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
return 0
fi
done
fi
# ============================================================
# 恢复源 4: Docker 卷扫描
# ============================================================
if command -v docker > /dev/null 2>&1; then
echo "🔍 Scanning Docker volumes for orphaned data..."
for VOL in $(docker volume ls -q 2>/dev/null); do
INSPECT_PATH=$(docker volume inspect "${VOL}" --format '{{ .Mountpoint }}' 2>/dev/null)
if [ -n "${INSPECT_PATH}" ] && [ -f "${INSPECT_PATH}/zen-garden.db" ]; then
FOUND_SIZE=$(ls -lh "${INSPECT_PATH}/zen-garden.db" 2>/dev/null | awk '{print $5}')
echo "🎉 Found orphaned data in volume '${VOL}' (${FOUND_SIZE})"
cp "${INSPECT_PATH}/zen-garden.db" "${DB_FILE}" 2>/dev/null
if [ -f "${DB_FILE}" ]; then
echo "✅ Database restored from orphaned Docker volume!"
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
return 0
fi
fi
done
fi
echo "📝 No existing data found anywhere — creating fresh database"
echo ""
return 1
}
# ============================================================
# 第三步：运行时持续备份（后台进程）
# ============================================================
start_safe_backup_daemon() {
# 每 3 分钟将 db 同步到镜像层内的安全备份
while true; do
sleep 180
if [ -f "${DB_FILE}" ]; then
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
# 也尝试备份到旧路径（兼容已挂载 /app/.data-backup 的用户）
if [ -d "/app/.data-backup" ]; then
cp "${DB_FILE}" "/app/.data-backup/zen-garden.db.safe" 2>/dev/null
fi
fi
done &
BACKUP_DAEMON_PID=$!
echo " Safe backup daemon: PID ${BACKUP_DAEMON_PID} (every 3min)"
}
# ============================================================
# 第四步：优雅停止时保存
# ============================================================
graceful_shutdown() {
echo ""
echo "🛑 Shutting down NOWEN..."
# 同步最新数据到安全备份（镜像层内）
if [ -f "${DB_FILE}" ]; then
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
echo "✅ Final safe backup saved to image layer"
# 也备份到旧路径
if [ -d "/app/.data-backup" ]; then
cp "${DB_FILE}" "/app/.data-backup/zen-garden.db.safe" 2>/dev/null
fi
fi
# 停止后台进程
kill "${BACKUP_DAEMON_PID}" 2>/dev/null
kill "${BACKEND_PID}" 2>/dev/null
# 停止 nginx
nginx -s quit 2>/dev/null
echo "👋 NOWEN stopped"
exit 0
}
# 注册信号处理
trap graceful_shutdown SIGTERM SIGINT
# ============================================================
# 执行启动流程
# ============================================================
echo "--- Data Persistence Check ---"
check_mount
echo ""
echo "--- Database Recovery ---"
recover_database
RECOVERY_RESULT=$?
echo ""
# 显示数据目录内容
echo "📁 Data directory contents:"
ls -la "${DATA_DIR}/" 2>/dev/null
echo ""
# 启动安全备份守护进程
start_safe_backup_daemon
# 启动后端
echo "Starting backend on port 3001..."
cd /app/server && tsx src/index.ts &
BACKEND_PID=$!
sleep 2
# 启动后再次同步（initDatabase 可能创建了新库）
if [ -f "${DB_FILE}" ]; then
cp "${DB_FILE}" "${SAFE_DB_FILE}" 2>/dev/null
fi
# 启动 nginx（前台运行）
echo "Starting nginx on port 3000..."
nginx -g "daemon off;" &
NGINX_PID=$!
# 等待任意子进程退出
wait ${NGINX_PID} ${BACKEND_PID}