Symptoms
When there is a failover in a high availability (HA) setup with SIP monitors, the SIP backend servers start flapping on both units. The reason this occurs is that after the failover, the two BIG-IP systems send SIP monitoring messages to the pool members with the same branch parameter on their Via headers. The backend server internal logic gets confused by the request coming from LB2 because it uses the same branch parameters of the request coming from LB1.
Impact
This causes the backend server erroneously to send a response message to LB1 instead of LB2.
Conditions
SIP branch hash string length is small enough that when sufficient SIP monitor messages were inundated, possible branch collision.
Fix Information
Branch collisions no longer occur in this configuration.
