I am currently in the process of researching how to setup out CX4-120 and VMware vSphere 4.1 environment to take advantage of ALUA and using round robin for our storage. I have read through the EMC documentation (EMC CLARiiON Integration with VMware ESX and Using EMC CLARiiON Storage with VMware vSphere) as well as Jason Boche's great blog post about setting it up. What I am questioning is this our production environement and am curious what (if any) could the impact be as I go server to server to change the failover mode via Navisphere, configure the SATP to use Round Robin, and switch the path selection policy on a LUN to servers that can access that same LUN but have yet to be configured?
In order to set ALUA on a CX/VNX system, you simply need to set the failovermode to 4
im attaching a document that shows how to do it and talks about ALUA in general.(FLARE 29 but it's similiar in FLARE 30)
once you set your array to ALUA (FLARE 29 and above), you probably want to use RR as well or even better than this, EMC PowerPath/VE as it's the only component that will allow a lun trespassing failback to it's original SP Owner, not to mention the superior Load Balancing PowerPath/VE will do to your enviornment.
also, you will need to reboot your ESX in order for it to properly work, there shouldnt be any negative effect on the ESX Servers you havent set ALUA on yet.