I have a question about ALUA on my CX3-80 (Running FLARE 26)
I guess I understand the basics of ALUA:
1) Set LUN to FAILOVER MODE 4
2) As long as the LUN gets accessed over the owning SP I have optimized access
3) When the LUN gets accessed over the non-owning SP I have non-optimized access
So, I can now reach my LUN via the active AND the inactive service processor. Any traffic that is directed to the non-active service processor will be routed internally to the active service processor.
What I am not sure about is the following:
What advantages does this give me?!?
I hope some people here have experience with the ALUA feature and can offer me some insight
Thanks in advance for any answers!
When using ALUA, the principal benefit is that it cuts down on trespassing. If you lose access to LUN on main path, IO is accessed over the internal pass thru. If the path comes back, you save the time you would have spend in trespassing the LUN.
The non-optimized path is not normally used by PowerPath in normal operation.
I believe that only Windows 2008 is supported with ALUA.
ESX 4.x is supported with ALUA, but not earlier.
See Support Solution emc254967 for the Agent information and emc245445 for iSCSI
Simply wanted to also point you to a short whitepaper that describes the feature and its benefits which is entitled:
White Paper: EMC CLARiiON Asymmetric Active/Active Feature — A Detailed Review
It is available from PowerLink via the following breadcrumb trail:
Home > Support > Technical Documentation and Advisories > Hardware/Platforms Documentation > CLARiiON CX3 UltraScale Series Systems > CLARiiON CX3-80 > White Papers
Furthermore, there is a matrix which lists the supported configuration using either native MPIO where for Windows your choices are limited to Windows 2008 as mentioned already or, as you'll notice, your support broadens when using a supported version of PowerPath.
One other reason to enable it might be to support certain features such as vSphere VAAI (vStorage API for Array Integration) which provides hardware acceleration (offload to the array) for various primitives and was introduced in vSphere 4.1. Please note, however, this is not relevant for your environment as it also requires array support, as you would expect, which is currently only available in FLARE 30 in addition to having ALUA enabled for your registered ESX hosts.
Actually, while ALUA FLARE support did begin in 26, VMware supports ALUA only with SCSI-3 reservations, which is not available to the CX3 (SCSI-2 reservation in flare code 26).