I am trying to set up MariaDB in a way that it automatically recovers -- basically restarts, from a crash.

Based on this article -- scroll down to Auto-starting Services with systemd, I opened /etc/systemd/system/multi-user.target.wants/mariadb.service file but there were nothing like this as the article explains:


So I added the [Service] section and the Restart=always underneath, restarted the service and killed the process to see if it will start automatically, but it didn't.

I was wondering if I'm editing the wrong file or if there are other ways of achieving the same thing. Server is a remote machine running CentOS 7.

  • What signal has been used to stop the service? Try SIGSEGV. Mar 21, 2016 at 11:48
  • @sam_pan_mariusz kill -9 oh is that because it wasn't a crash it won't restart?
    – Mahdi
    Mar 21, 2016 at 11:54
  • 1
    I've seen Upstart (the former init program in Ubuntu) not restarting certain services when they received SIGTERM, so I assumed a similar issue may exist with systemd. Other signals however worked as expected, SIGKILL (9) including. Mar 21, 2016 at 12:38
  • @sam_pan_mariusz thanks, I will try to see if there are similar solutions for Upstart that I can't apply on systemd.
    – Mahdi
    Mar 21, 2016 at 13:06

1 Answer 1


After making changes to systemd units, you need to run systemctl daemon-reload for those changes to have effect.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .