RAID cache and VDIs

In a virtual environment with ESXi is there a significant issue or concern with setting up some unique RAID/cache settings on the the hardware that the hypervisor runs on? Or can you accept the defaults (the out of the box settings for the lack of a better word) and concentrate on optimizing performance in other areas?

Or is it better to 'tweak' the cache settings on the VDI image or is there a mechanism to control this in VMWare View or VCenter Server?
Who is Participating?
The defaults will be awful.  The correct, most efficient way to configure the RAID is going to be a function of
 - RAID level
 - # of disks in the array
 - "native" I/O size that the O/S is asking for, such as 256KB.  
So if ESXi is asking for 256KB at a time (which may or may not be correct, it depends on how the VMDK is configured), then you want to make sure that a write of 256KB is evenly balanced.  I.e, if you have a RAID5 with 4 disks total, each set for a chunk size of 64KB, then it is going to take twice as much I/O because you can only write 64x3 worth of that 256KB of data at once.

(That other 64K represents the parity for the first 192KB worth of data from ESXi)

 That leaves another 64KB of data, and since it is a RAID5, you have to do extra parity, and the parity data for the next chunk of 64KB is on another drive, so this is a lot of wasted I/O.

But if you configured a 5-disk RAID5, then all the data can be done at once because you have 64KB x 4 for the data, and the 5th drive has the parity.
xmouserAuthor Commented:
Nice, thanks.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.