First off, I'm new to F5 and trying to get a better understanding of how things work.
We have 2 pairs of BIG-IPs (v16.1) at 2 locations currently in place and kind of working. Both pairs are setup as primary and standby. And the pairs are tied together via DNS/GTM/GSLB.
Is there any reasonable scenario where one would need more than one floating IP for each pair for external and internal connections in our setup?
Is it best practice to have the VIP use the floating IP as the destination address? Or should it be using a different IP?
@vimleyudruIt is not recommended to use the self or floating IP as a destination for a virtual server because you will limit yourself on what you can use those IPs for by doing so. You should always allocate a new IP if possible and then when a failover event occurs the devices will use the floating IP to move routing between the two units which will move all virtual servers as well because they are configured virtual servers. Self and floating IPs typically are only used for routing traffic and failover communication and additional allocated IPs are for other purposes such as virtual servers, DNS resolution, and so on.
The floating and none floating address on the external interface are purely for the BIG-IP.
Use a different IP for the virtual server that in the same range as the external interfaces. This address moves at the same time as the floating address. But think as one for the system and one for your service.
Other than that it'll depend on what services are running and how you are desiging them.
So one vs/ IP per service/app per site.
Or one vs/per per site and you can manage other services by url and via a policy and use something like vip targetting vip to only present 1 IP per site.
If you pay per IP at your DC's then this can really help scalability.
Then you can simply use GTM to manage which site/service to be using and where.