如何确保IPFS daemon持续存活？单节点IPFS应用运维求助

Got it, let's tackle this problem—keeping your IPFS daemon up and running reliably, even when it crashes, especially since you're currently using tmux. Here are a few practical, battle-tested approaches you can implement right away:

1. Use Systemd (Recommended for Linux) #

Systemd is built into most modern Linux distros and is designed to manage long-running services with automatic restart capabilities. This is the most robust solution for keeping your IPFS daemon alive.

• Create a systemd service file:

  sudo nano /etc/systemd/system/ipfs-daemon.service
  

• Paste the following content (replace your_username with your actual user account):

  [Unit]
  Description=IPFS Daemon
  After=network.target
  
  [Service]
  User=your_username
  ExecStart=/usr/local/bin/ipfs daemon
  Restart=always
  RestartSec=5
  StandardOutput=journal+console
  StandardError=journal+console
  
  [Install]
  WantedBy=multi-user.target
  

• Save and exit the editor, then reload systemd and start the service:

  sudo systemctl daemon-reload
  sudo systemctl enable --now ipfs-daemon.service
  

• Check the service status or view logs anytime:

  # Check if the daemon is running
  systemctl status ipfs-daemon
  
  # View real-time logs
  journalctl -u ipfs-daemon -f
  

Systemd will automatically restart the daemon if it crashes, hangs, or exits for any reason.

2. Shell Script + Tmux #

If you prefer sticking with tmux, a simple loop script can handle automatic restarts for you.

• Create a restart script (e.g., ipfs-auto-restart.sh):

  nano ipfs-auto-restart.sh
  

• Add this loop logic:

  #!/bin/bash
  while true; do
      echo "[$(date)] Starting IPFS daemon..."
      # Run the daemon with debug logging to help diagnose crashes later
      ipfs daemon --debug
      echo "[$(date)] IPFS daemon exited. Restarting in 5 seconds..."
      sleep 5
  done
  

• Make the script executable:

  chmod +x ipfs-auto-restart.sh
  

• Run it in a tmux session (you can detach and leave it running in the background):

  # Create a new detached tmux session named "ipfs" running the script
  tmux new-session -d -s ipfs './ipfs-auto-restart.sh'
  
  # Reattach to the session later to check logs
  tmux attach -t ipfs
  

Now, every time the daemon crashes, the script will wait 5 seconds and restart it automatically.

3. Use Supervisor (Cross-Platform Option) #

Supervisor is a process control system that works across Linux distros and macOS, making it a flexible alternative if systemd isn't an option.

• Install Supervisor:

  # Debian/Ubuntu
  sudo apt install supervisor
  
  # macOS (with Homebrew)
  brew install supervisor
  

• Create a Supervisor config file:

  # Linux:
  sudo nano /etc/supervisor/conf.d/ipfs.conf
  
  # macOS:
  nano /usr/local/etc/supervisor.d/ipfs.ini
  

• Add this configuration (replace your_username):

  [program:ipfs-daemon]
  command=ipfs daemon --debug
  user=your_username
  autostart=true
  autorestart=true
  startretries=3
  stdout_logfile=/var/log/ipfs-daemon.log
  stderr_logfile=/var/log/ipfs-daemon.err.log
  

• Reload Supervisor and start the service:

  # Linux
  sudo supervisorctl reread
  sudo supervisorctl update
  sudo supervisorctl start ipfs-daemon
  
  # macOS
  brew services restart supervisor
  supervisorctl start ipfs-daemon
  

• Check the daemon status:

  supervisorctl status ipfs-daemon
  

Bonus: Diagnose the Root Cause #

While automatic restarts will keep your service running, it's worth figuring out why the daemon is crashing in the first place. Enable debug logging (as shown in the script and config examples) and check the logs—this might reveal issues like memory limits, network problems, or corrupted IPFS repo data. If you spot errors, you can address them to prevent crashes entirely.

内容的提问来源于stack exchange，提问作者Vasia Kotov