There's a routine need to support this scenario: a service runs that can fail 
and needs to be restarted.

But if it just keeps failing it doesn't make sense to keep restarting forever, 
it's just overhead and the system is stuck. So in that case 1) the service 
needs to be left stopped in "failed" state, and 2) whatever failure action was 
specified - that action needs to be run.

Ideally, need to be able to limit restarts to a number per period of time, such 
as 3 restarts in 10 minutes, etc.

Is there a way to do this?


systemd-devel mailing list

Reply via email to