We have a VIP on our BigIP that sees a lot of connections. The TCP profile applied to the VIP is TCP Progressive with no Nagle. What we are seeing is after a very small amount of data the Window size on the F5 Self IP to Server connection is going to zero. The back-end server is able to maintain its buffer levels. The intention of applying a progressive profile was to take care of varying conditions and connections. We saw over 600 TCP Zero Window instance from the F5 to the server. all being sent by the F5. I am not sure what tweaks to the TCP profile i need to do to solve this problem.
have successfully implemented HA configuration on numerous f5 LTM unit across our network without encountering any issue. Specifically, we are wondering why only one application sems to be impacted after the trunk configuration in f5 LTM.
As of now, the application is functioning smoothly with standard VIP, and there have been no reported related to download speed or performance and currnely physical interface direclty map to VLAN. However, we are keen to understand why this specific application is experiencing difficulties following trunk configuration.
