Restart Limit in FRR service

[ChinmayaAgarwal]

Hi,

We have observed that FRR has a restart limit configured in its service file (i.e. 3 restarts within 3 minutes). What is the reason for configuring this limit?

Thanks and Regards,
Chinmaya Agarwal

[taspelund]

The point of a restart limit is to put an upper bound to the amount of impact an FRR failure can have on the system it's running on. By adding the restart limit you make sure the service doesn't get into a restart loop and you avoid the possibility of systemd getting stuck trying to manage that loop.

I have seen other services hosted by systemd crash + attempt to restart more than 100k times within a single minute when there wasn't a restart limit configured. The failure itself was not transient either, so it wasn't going to eventually resolve itself. For situations like that, it's better to leave the service dead so the system doesn't get consumed.