SRENMAN
Contributor
Contributor

VMFS storage Partition Performance

Jump to solution

I am trying to determine If I am going to get better performance from 6 FC-300Gb 15K Disks in a RAID10 array or 6 FC-300GB 15K disks in a RAID5 array?

I don't have a lot of time to really sit down and experment (would love to but time= $$).

This is the array that is going to store between 5-9 WIN2k3 VM's.

I like the idea of RAID 10 allowing more that 1 disk to fail (as long as its not the same mirror set) and still survive with quick rebuild times.

THis question has been put out before, but not with the quantity of drives I'm suggesting

Any help is appreciated!

0 Kudos
1 Solution

Accepted Solutions
larstr
Champion
Champion

If you have a mixed workload (most do) then RAID10 will give you much better performance than RAID5. RAID5 has faster sequential reads than RAID10, but RAID10 is normally the one that gives you best performance in a VMware environment where multiple VMs are accessing the same storage.

Lars

View solution in original post

0 Kudos
2 Replies
larstr
Champion
Champion

If you have a mixed workload (most do) then RAID10 will give you much better performance than RAID5. RAID5 has faster sequential reads than RAID10, but RAID10 is normally the one that gives you best performance in a VMware environment where multiple VMs are accessing the same storage.

Lars

View solution in original post

0 Kudos
SRENMAN
Contributor
Contributor

And on a 500GB VMFS Storage Part, should I be looking at a 1mb or 2mb block size? it keeps defaulting to 1 for that size range but If VM's like 2mb or more I would like to start on the right track,

Thanks

0 Kudos