vThunder - VRRP Issues

dpcarlton · June 2020

We have been seeing an issue with one part of vThunders running vCS & VRRP.

Running 4.1.1-P11, build 47 which is the same version we have used in other deployments (20+).

We constantly get missed heartbeat messages and then eventually, we get into a dual active situation and then lose BGP. It takes a few minutes for them to resort out who is master and to settle back down. This happens from once to 3 or 4 times per day.

The VRRP config has been verified in the past with Tech Support, and this is the only site we see this issue on.

Also, we get high latency and jitter between the pair. Pinging between other devices and VMs is fine, ruling out underlying switch issue (Nexus 9k).

For instance pining from one vA10 to another, over the dedicated VRRP interface. I should note these are on the same switch, 10G links, and the appliances are relatively quiet, maybe 1000 sessions total, mostly UDP.

vtlb02-Standby-vMaster[1/2]#ping 10.1.99.245
PING 10.1.99.245 (10.1.99.245) 56(84) bytes of data.
64 bytes from 10.1.99.245: icmp_seq=1 ttl=64 time=20.3 ms
64 bytes from 10.1.99.245: icmp_seq=2 ttl=64 time=55.5 ms
64 bytes from 10.1.99.245: icmp_seq=3 ttl=64 time=30.0 ms
64 bytes from 10.1.99.245: icmp_seq=4 ttl=64 time=290 ms
64 bytes from 10.1.99.245: icmp_seq=5 ttl=64 time=20.9 ms

Has anyone seen this before? I am leaning towards them being deployed incorrectly somehow and to just redeploy them.