Skip to content

A service is down

When: sites, mail, DNS, databases or the panel stop answering, or an alert says a service is down.

  1. flypanel monitor status — find the service and its state.
  2. Restart it through FlyPanel, not by hand: flypanel monitor restart SERVICE (the ids monitor status shows). The watchdog does this itself after two failed checks; if it already gave up, the service keeps failing on start.
  3. Read why: journalctl -u <unit> -n 100 --no-pager (the unit names are in monitor status).
  4. Web (OpenLiteSpeed): flypanel web rebuild re-renders every site’s configuration and reloads. Never run lswsctrl restart from SSH: OLS then runs in the wrong SELinux domain and later reloads fail; if it was done, kill the stray OLS master and systemctl start lshttpd.
  5. Mail: flypanel mail apply re-renders Postfix, Dovecot and Rspamd.
  6. Firewall: flypanel firewall status; flypanel firewall apply loads it again.
  7. flypanel doctor — confirm 0 errors.

It worked when: monitor status shows it running and the thing users reported works (load the site, send a test mail). Escalate when: the service fails again within minutes, the journal shows a disk, memory or SELinux denial (ausearch -m AVC -ts recent), or several services are down at once.