I was looking for some kind of explanation as to why 6+2 or 14+2 is the preferred disk count for RAID6 pool? Or why 4+1 or 8+1 is the preferred disk count for RAID5? It does not mean though that I cannot use 4+2 for RAID6, or 6+1 for RAID5. Does anyone know how to explain this? Why EMC has this preferred disk counts, and what is going to happen if I use nonpreferred disk counts in my VNX configuragtions?
The short answer would be that that is what EMC has found to be the best size for the RAID group for the most I/O workloads and applications. Ultimately the goal is to achieve full stripe writes to your disks and these sizes seem to bring in the greatest amount of full stripe writes.
I've attached a document to my post. It's another best practice guide. If you skip straight to page 77 if discusses the RAID level performance characteristics and discusses the preferred RG sizes. I'd recommend starting a little higher on Page 74 and reading for a few pages to get some great background information on RAID leading up to the performance characteristics.
Was your question answered correctly? If so, please remember to mark your question Answered when you get the correct answer and award points to the person providing the answer. This helps others searching for a similar issue.
I'm curious why every "Best Practices" PDF I see if for Unified systems. What difference does it make if it's unified or block only? We use the latter.
Jyothi_P_Bharat
317 Posts
1020
1
Posted April 6th, 2015 00:00
Hi,
Please check the below mentioned link:
http://www.storagenetworks.com/documents/emc/emc-vnx-best-practices-wp.pdf
(from page num13)
Thanks
Jyothi