Keep in mind that a Virtual RAID is a logical structure presented to SAN/iQ.  This implementation parallels what you will find on real LeftHand SAN/iQ platforms that integrate physical system and RAID knowledge as a platform.  In a VMworld, the real physical is abstracted from what is presented to the virtual machine.  Although implementation of a software RAID configuration for physical devices or virtual disks represented as file could be implemented, it is not at this time.  A configuration allowing software RAID could inappropriately give the impression that there is redundancy when there is none.  (For example, 2 virtual disks on the same drive are RAID 1 mirrored...  however, the single drive failure will cause all data to be lost)

For redundancy, create a cluster of VSAs.  Ideally, they would be on unique VMware servers (ESX or otherwise).  After the VSA nodes are part of a cluster, when you create a volume select the replication level.  This provides for network RAID across the VSAs.

By using hardware based RAID, you present that data structure to VMware that just happens to be RAIDed.  This will provide you with the redundancy.  In order to manage that structure from within a virtualized machine (and have warnings on impending drive failures), you have to have an out-of-band path for notifications of RAID issues unless specific VMware aware drivers exist along with management (or the outer shell OS has this installed).  At this time, a virtualized environment has no concept of the allocated space behind it is a RAIDed structure.

Thanks for your patience Gene.

Christopher Viamonte