My predecessors created a /24 block - say 10.10.20.0/24 for iSCSI storage. The arrays are all on the 10.10.20.0/24. One option could be to add a new subnet. But the trick is that the default route on the ESX is out the management subnet - say 10.10.10.0/24. So traffic from a 10.10.20.0/24 to 1 10.10.22.0/23 would not go out/return via the storage vNIC. I could solve this by adding a route to every ESXi so that a an ESXi on 10.10.20.0/24 would have a route to 10.10.22.0/23 via its gateway 10.10.20.1 and a host on 10.10.22.0/23 would have a route to 10.10.20.0/24 via its 10.10.22.1 gateway.
The other option I was considering was to just change the subnet of each ESXi host from a 10.10.20.0/24 to a 10.10.0.20/23 and then as a final step change the storage array targets from 10.10.20.0/24 to 10.10.20.0/23. Since there's redundancy I should be able to shut one interface, make the change, unshut the interface. First I'd do this on all the ESX hosts and then on all the iSCSI targets.
Any notes of caution or thoughts on the best approach in terms of safety and ease? Thanks much!
