如何在Linux CentOS上实现MongoDB进程终止后自动启动

Got it, let's break down how to solve this problem—you want MongoDB to automatically restart if it gets killed (like with Ctrl+C), and also understand how to monitor its shutdown events to trigger a restart. Here are a few solid, tested approaches:

方案1：利用Systemd实现自动重启（推荐） #

This is the most reliable method for Linux systems, since systemd is built to monitor and manage service processes. It'll automatically restart MongoDB no matter why it stops—whether it's a manual Ctrl+C, crash, or normal exit.

1. Locate MongoDB's systemd service file. Common paths are /etc/systemd/system/mongodb.service or /lib/systemd/system/mongodb.service.
2. Edit the file and find the [Service] section. Add or modify these parameters:

   Restart=always
   RestartSec=5  # Wait 5 seconds before restarting, adjust as needed
   

   • Restart=always means systemd will restart MongoDB regardless of how it terminates. If you only want restarts on crashes, use Restart=on-failure instead.
3. Save the file, reload systemd's configuration, and restart the service:

   sudo systemctl daemon-reload
   sudo systemctl restart mongodb
   

4. Test it out: Kill MongoDB with sudo systemctl stop mongodb or Ctrl+C, then check status with sudo systemctl status mongodb after a few seconds—it should be back up.

方案2：自定义监控脚本 #

If you're on a system without systemd (like older Linux distros, macOS, or Windows), a simple polling script will work. It checks if MongoDB is running at regular intervals and restarts it if not.

Linux/macOS Bash Script #

#!/bin/bash

# Replace these paths with your actual MongoDB command and config
MONGODB_START="/usr/bin/mongod --config /etc/mongod.conf"
PROCESS_NAME="mongod"

while true; do
    # Check if mongod process is active
    if ! pgrep -x "$PROCESS_NAME" > /dev/null; then
        echo "$(date): MongoDB process missing—restarting..."
        $MONGODB_START &
    fi
    sleep 10  # Check every 10 seconds, tweak this value
done

• Save as mongodb_monitor.sh, make it executable: chmod +x mongodb_monitor.sh
• Run it in the background: nohup ./mongodb_monitor.sh > mongodb_monitor.log 2>&1 &

Windows Batch Script #

@echo off
set MONGODB_EXE="C:\Program Files\MongoDB\Server\6.0\bin\mongod.exe"
set CONFIG_FILE="C:\Program Files\MongoDB\Server\6.0\bin\mongod.cfg"

:loop
:: Check if mongod is running
tasklist /FI "IMAGENAME eq mongod.exe" 2>NUL | find /I /N "mongod.exe">NUL
if "%ERRORLEVEL%"=="1" (
    echo %date% %time%: MongoDB not running—restarting...
    start "" %MONGODB_EXE% --config %CONFIG_FILE%
)
timeout /t 10 /nobreak > NUL
goto loop

监听MongoDB Shutdown Events #

If you want to trigger a restart exactly when MongoDB shuts down (instead of polling), you can monitor its log file for shutdown-related messages.

1. Find your MongoDB log path (check the systemLog.path setting in your mongod.conf—common path: /var/log/mongodb/mongod.log).
2. Use this script to listen for shutdown signals in real time:

#!/bin/bash

MONGODB_START="/usr/bin/mongod --config /etc/mongod.conf"
LOG_FILE="/var/log/mongodb/mongod.log"

# Tail the log and watch for shutdown keywords
tail -F "$LOG_FILE" | while read line; do
    # Match common shutdown messages (adjust for your MongoDB version if needed)
    if echo "$line" | grep -E "(Shutting down|Received signal 2|Terminating)"; then
        echo "$(date): Detected MongoDB shutdown—restarting..."
        $MONGODB_START &
    fi
done

• Received signal 2 is the log entry for a Ctrl+C (SIGINT) shutdown
• Shutting down appears during normal termination
• This script will immediately trigger a restart as soon as it sees these messages

内容的提问来源于stack exchange，提问作者Ralph Marvin Addo