ALUA behavior with Clariion and VMWare

hI,

Wondering if someone could explain the behavior expected with ALUA - Failover mode 4 with VMWARE and Clariion.

Host VMWARE 4.0

Clariion cx4 Flare 29

We have the luxury at the moment of being able to conduct some very specific testing so I wanted to see what happens with this ALUA mode as I have seen recommended by several people .

Basically the ESX server has 1 HBA physically connected via a switch to see SPA2 and the other HBA physically connected to a different switch to see SPB2 . Ports 0, 1 and 3 on both SP's are currently not connected to these switches.

The ESX server has been presented with Luns from SPA and SPB using ALUA failover mode 4 in Navisphere  and set to fixed path in ESX .

I can see the HBA's showing as having the correct ACTIVE I/O paths in the ESX gui as per my zoning.

I am trying to simulate what would happen in an NDU .

1) Tresspass all SPA owned luns in Navishere to SPB.

The ESX GUI does not relect a change in the ACTIVE I/O status - the administrator can still access the disk

2) Remove the fibre cable from HBA zoned to SPA

The ESX GUI relects a change in the ACTIVE I/O status - the administrator can still access the disk - LUNs are automatically tresspassed to SPB

3) Reconnect the cable

The ESX GUI DOES NOT relect a change in the ACTIVE I/O status - the administrator can still access the disk . The disks are still tresspassed to SPB .

I thought the whole idea of this thing was to enable automatic tresspassing and therefore load balancing of HBA's at ESX in the event of an NDU .

Am I  missing something ?

thanks

Steve