We have somehow the same Problem (DSSV6 up90 b5845):

Our ESX Hosts receive all ip-adresses for the targets including
all physical interfaces
(three DSS pairs, each DSS 6 physical interfaces) from our DSSs,
although we restricted access to the targets on DSS to the virtual ip subnets.

But our ESX-hosts do not have ip adresses configured for storage
in the physical address range of the DSSs interfaces, so there's no way for
the ESX hosts to connect to a physical interface of the DSSs.

So far, so good,
but if we have to rescan our ESXs iSCSI vmhba, it takes AGES to complete the rescan.
We suppose, that this is because the ESX tries to reach the physical interfaces
of the DSSs somehow and after a time runs finally in a timeout.

So the question is:
Is there any way to limit the "sendtargets" of DSS to the virtual interfaces as mentioned by Gr-R ?

Quote Originally Posted by Gr-R
In future we will have ability to designate which IP is broadcasting the target, so MPIO can be controlled easier.
We know, that we could configure all targets on all ESX hosts static to avoid this,
but that's a lot of work for 6 ESX hosts with 15 DSS targets with each 3 paths....

Regards,
Ralph