SNAT pool member with IP conflict caused by F5?

As I understand the SNAT-pool is which ip the F5 will use as srcip towards the true destination.

 

 

If that is the case then of course your machines (true destination) shouldnt have any of the particular ip addresses binded to any of its interface (since the F5 already binded them).

 

 

In the persistence profile (for example a custom source_addr rule) I would suggest to alter the timeout for how long (of idle time) you want to bind a particular client to a particular destination server but also set on fail to reject. If one of your servers falls out of the pool the client(s) connected to this particular server will get a reject for next packet which will result in that the client will try to re-establish its connection to the vserver where the vserver takes a new loadbalancing decision (where the failing server is no longer part of the pool) and *poff* your client gets a fresh connection to one of the still functioning servers.