无论它们存在或崩溃的原因，如何重启python scrypt？

Question

Say I have this sh script to monitor my python script and restart it if it crashes:

until myserver; do
    echo "Server 'myserver' crashed with exit code $?.  Respawning.." >&2
    sleep 1
done

While it might work well for a python script which is supposed to do some work and exit , it won't for my case because I need my python script s (a few ones, not only one) to always work 24/7 in background. And if one of them ever exists that means it's crashed and should be restarted.

How should I handle my case?

Answer 1

This can be accomplished with crontab .
Run crontab -e , which will bring up a text editor.
Add a new line to that file for each script,

*/5 * * * * pgrep -f yourPythonScript1.py || nohup python /fullpathtoyourfile/yourPythonScript1.py

Save that file and exit the editor.

This will make a new crontab which runs every 5 minutes and launches each script unless it is already running. (Feel free to edit the frequency to what you need)

The command shown runs pgrep (basically, a find a process running with a given string name)

The -f searches the entire command used to run the process- you'll want this because you're running multiple python scripts)

|| means if the previous command fails, do what follows. So the pgrep will fail when your specific script isn't running and it will be relaunched.

Answer 2

If you have root on the system.

If you have a distro with systemd.

then...

You can use systemd to restart a process as soon as it ends using "Restart=always". For example:

[Unit]
Description=My cool service
After=network.target

[Service]
ExecStart=/usr/local/sbin/myservice
User=<myuser>
Group=<mygroup>
Restart=always

[Install]
WantedBy=multi-user.target

Answer 3

Use supervisord to manage your scripts. It provides you with restarting utilities, logging and remote monitoring. Is not hard to setup.

http://supervisord.org/