NimbostratusHave a virtual server doing ssl offload with persistence of none. To a pool with two servers. Testing multiple connections from same source IP, statistics show all traffic connecting to the same pool member. With persistence none is this still expected behavior?
NoctilucentWhich lb method you have applied in pool? Have you attached oneconnect profile to vip? Try to hit vip with different browser and see..
NimbostratusDoing round robin, there's not a oneconnect profile. Tried from two browsers, all connections still seem to be hitting same node...
NoctilucentPlease refer details explanation of @Hannes
NimbostratusThis is normal behavior. By default, BigIP LTM balances TCP connections, not individual HTTP requests. Take note of 5 minute default TCP idle timeout. If you see that all connections go to same pool member even after 5 minutes since last request was made, then you have a problem that needs further investigation.
With OneConnect profile you can change this behavior so that BigIP will start to balance individual GET/POST/etc. requests within the same session but this is a double-edged sword - you will lose visibility of true client IP address in network stack and must rely on HTTP headers such as X-Forwarded-For to get client IP information in end-servers.
Using least-sessions or least-connections LB algorithm, you can achieve very good balancing results without OneConnect. You just need more concurrent clients to see those numbers even out nicely.