3PAR StoreServ Storage
cancel
Showing results for 
Search instead for 
Did you mean: 

What is the Maximum number of Simultaneous disk failures that can be handled by 3PAR

SOLVED
Go to solution
binu_mathew
Frequent Advisor

What is the Maximum number of Simultaneous disk failures that can be handled by 3PAR

Hi

 

I would like to know the Maximum number of Simultaneous  disk failures that can be handled by 7400c with 4 node storage without any data loss.

 

I have the below configuraton

 

56x SAS 1.8TB in RAID 5(7+1)

 

16x 480 GB SSD in RAID 1 ( adaptive cache)

 

1) I believe FOR RAID 1 I can tolerate upto 8 drive failure as its mirrored . Please confirm .

 

2) But what about the sas disks if I configure a single CPG with 56 drives ?

 

 

 

 

 

 

3 REPLIES
Sheldon Smith
Honored Contributor

Re: What is the Maximum number of Simultaneous disk failures that can be handled by 3PAR

RAID5, one disk.

RAID6, two disks.

 

Period.

 

Anything beyond that involves luck.


Note: While I work for Hewlett Packard Enterprise, all of my comments (whether noted or not), are my own and are not any official representation of the company.
----------
If my post was useful, click on my KUDOS! thumb below!
binu_mathew
Frequent Advisor

Re: What is the Maximum number of Simultaneous disk failures that can be handled by 3PAR

Hi,

 

Then what does RAID 5(7+1) means.. Is it for every 7 disk one parity disk or for every 7 data one parity block ?

 

 

 

Is there any option in 3par to add more hot spare disk in RAID 5

 

Or Is there any option to increase the disk failure protection level other than RAID 6

 

Kindly suggest the best ..

Sheldon Smith
Honored Contributor
Solution

Re: What is the Maximum number of Simultaneous disk failures that can be handled by 3PAR

RAID 5(7+1) means for every 7 data chunklets, there is one parity chunklet.
The 3PAR does not have hot spare disks. There are spare chunklets reserved on each physical disk; if a physical disk (starts to) go bad, the RAID 1/RAID 5/RAID 6 chunklets in use on that physical disk are either moved or regenerated on spare chunklets elsewhere in the array.

 

So with your earlier example, logical disks spread over 16 physical disks using RAID1, 8 disks could fail and if they are the "right" 8, it would still be operational, although degraded performance. If the two disks providing mirrored chunklets both go bad, access stops.

Likewise 56 RAID5(7+1): 7 disks could fail and if they are the "right" 7, it would still be operational, although degraded performance. If two disks in a single RAID5 both go bad, access stops.

 

So what is supported? Like I said before:
RAID5, one disk,
RAID6, two disks.
Period.

Anything beyond that involves luck. If you can predict which disks are going to fail, why are you wasting time in IT? Head to Vegas!


Note: While I work for Hewlett Packard Enterprise, all of my comments (whether noted or not), are my own and are not any official representation of the company.
----------
If my post was useful, click on my KUDOS! thumb below!