步骤 1：创建 Bash 脚本

搞了个定时脚本来解决 PVE 下 OpenWRT 意外卡死的问题。

这个 OpenWRT 容器作为网络中的唯一主路由，所以用最简单的方法就是直接在宿主机上检查网络联通性来推测 OpenWRT 是不是寄了。

定时任务使用了 systemed timer 来实现，感觉确实不错

-（Systemd 定时器教程 - 阮一峰的网络日志

• pct(1)

步骤 1：创建 Bash 脚本

1. 创建脚本文件 /usr/local/bin/restart_openwrt_when_error.sh：

sudo nano /usr/local/bin/restart_openwrt_when_error.sh

1. 将以下内容粘贴到脚本文件中：

#!/bin/bash

if [ -z "$1" ]; then
 echo -e "\e[31mUsage: $0 <LXC_ID>\e[0m" >&2
 exit 1
fi

LXC_ID=$1
RETRY_COUNT=3
RETRY_INTERVAL=5
fail_count=0

echo -e "\e[34mStarting network accessibility check for https://qq.com and https://baidu.com...\e[0m"

for ((i=1; i<=RETRY_COUNT; i++)); do
 if curl -s --head --connect-timeout 10 https://qq.com/ > /dev/null || curl -s --head --connect-timeout 10 https://baidu.com/ > /dev/null; then
 echo -e "\e[32mNetwork is accessible.\e[0m"
 fail_count=0
 break
 else展开全部 (39 行)
#!/bin/bash

if [ -z "$1" ]; then
 echo -e "\e[31mUsage: $0 <LXC_ID>\e[0m" >&2
 exit 1
fi

LXC_ID=$1
RETRY_COUNT=3
RETRY_INTERVAL=5
fail_count=0

echo -e "\e[34mStarting network accessibility check for https://qq.com and https://baidu.com...\e[0m"

for ((i=1; i<=RETRY_COUNT; i++)); do
 if curl -s --head --connect-timeout 10 https://qq.com/ > /dev/null || curl -s --head --connect-timeout 10 https://baidu.com/ > /dev/null; then
 echo -e "\e[32mNetwork is accessible.\e[0m"
 fail_count=0
 break
 else
 echo -e "\e[33mAttempt $i: Network is not accessible. Retrying in $RETRY_INTERVAL seconds...\e[0m"
 fail_count=$((fail_count + 1))
 fi
 sleep $RETRY_INTERVAL
done

if [ $fail_count -eq $RETRY_COUNT ]; then
 echo -e "\e[31mNetwork is not accessible after $RETRY_COUNT attempts. Restarting LXC container $LXC_ID...\e[0m" >&2
 if ! timeout 120 pct reboot $LXC_ID --timeout 30; then
 echo -e "\e[31mTimeout occurred during pct reboot. Forcing stop and start of LXC container $LXC_ID.\e[0m" >&2
 pct stop $LXC_ID
 sleep 5
 pct start $LXC_ID
 else
 echo -e "\e[32mLXC container $LXC_ID restarted successfully.\e[0m"
 fi
else
 echo -e "\e[34mNo need to restart LXC container $LXC_ID.\e[0m"
fi
收起

1. 保存文件并退出编辑器。

2. 赋予脚本执行权限：

sudo chmod +x /usr/local/bin/restart_openwrt_when_error.sh

步骤 2：创建 Systemd 服务文件

1. 创建服务文件 /etc/systemd/system/restart_openwrt_when_error.service：

sudo nano /etc/systemd/system/restart_openwrt_when_error.service

1. 添加以下内容：

[Unit]
Description=Check network access and restart OpenWRT LXC if unreachable

[Service]
Type=oneshot
ExecStart=/usr/local/bin/restart_openwrt_when_error.sh 305
TimeoutStartSec=5min

将 <LXC_ID> 替换为实际的 LXC 容器 ID。

1. 保存文件并退出编辑器。

步骤 3：创建 Systemd 定时器文件

1. 创建定时器文件 /etc/systemd/system/restart_openwrt_when_error.timer：

sudo nano /etc/systemd/system/restart_openwrt_when_error.timer

1. 添加以下内容：

[Unit]
Description=Run restart_openwrt_when_error.service every 10 minutes

[Timer]
OnBootSec=3min
OnUnitActiveSec=10min
AccuracySec=1s
Persistent=false

[Install]
WantedBy=timers.target

1. 保存文件并退出编辑器。

步骤 4：启用并启动服务和定时器

1. 重新加载 Systemd 配置：

sudo systemctl daemon-reload

1. 启用并启动定时器：

sudo systemctl enable restart_openwrt_when_error.timer
sudo systemctl start restart_openwrt_when_error.timer

1. 检查定时器状态，以确保其正常运行：

sudo systemctl status restart_openwrt_when_error.timer

步骤 5：查看日志

• 使用 journalctl 查看服务的日志输出，包括脚本的输出：

journalctl -u restart_openwrt_when_error.service