We’re tracking down something that might be an issue – we’re not sure. We don’t normally run frr with all debug logging turned on, and don’t know if this behavior is normal or not (we don’t have much to compare it to).
Every 30 seconds, with logging on, we see this in the logs (for each interface with route):
2024 Jul 5 11:31:41 hostredacted [daemon.debug] ospfd[960266]: [JZFS1-TNQFD] Zebra: interface services address delete 10.10.205.4/32
2024 Jul 5 11:31:41 hostredacted [daemon.debug] ospfd[960266]: [PH9HB-NQMHW] Zebra: interface services address add 10.10.205.4/32 vrf default id 0
It seems like this is flapping the interface to us, but maybe it’s a normal part of frr internals to do so? Is this normal for frr to delete and add the interface every 30 seconds?
Functionally, OSPF and our anycast DNS seems to be running ok. But our monitoring has missed some pings (which could be explained with the interface delete/add).
(this is on rhel9, frr-8.5.3-4, although I also get the same behavior when I downgrade to frr-8.3.1-11)
Thank you for any input!