LTM VS 1 to Many (Clone Pool)?

You can technically use the 'clone pool' command multiple times in an iRule, but with a few caveats:

• If all of the servers are on the same VLAN, it's likely that a single clone pool statement will spray traffic to all of them. You can effectively "load balance" clone pool traffic if the servers are separated across VLANs.

   

• Clone pools don't mirror the L2 headers of the original traffic, so you'd expect to the F5's Mac address at the service.

   

• And to be clear, a clone pool is not bidirectional. It sprays a copy of the packets to an interface, and does not attempt to establish a TCP connection to the host. You need a passive/tap-style device that processes traffic in this way.

   

• A clone pool requires a pool of IPs, which is not typically what you expect a packet copier to need. To make clone pools work for passive/tap devices, create a VLAN and assign the correct interface. Create a unique/unused self-IP on the VLAN, then create a pool using a random IP in that subnet. After that, create a static ARP that maps the pool IP to the Mac address of the target device. If the device is on its own isolated VLAN, then the Mac address doesn't really matter. However, whatever you assign with show up on the device as the destination Mac, with the F5 as the source Mac. In v13 you can now also point directly at a VLAN: 'clone pool ', which negates the need for the pool and static arp.