RAID based on performance or purpose

It is widely said RAID 5 is good for data reads and RAID 10 for data writes, while RAID 10 is more better for reliability of data and more costly, while the concern with RAID 5 is if it fails, we may lose data...

if for a reporting sql server, where it is 95% read, will RAID 5 be a better option, even though RAID 10 is affordable?

what are the pros and cons of it.
thanks
LVL 6
anushahannaAsked:
Who is Participating?
 
MrMintanetCommented:
If a drive costs $1000US (and most are far less expensive than that) then switching from a 4 pair RAID10 array to a 5 drive RAID5 array will save 3 drives or $3000US. What is the cost of overtime, wear and tear on the technicians, DBAs, managers, and customers of even a recovery scare? What is the cost of reduced performance and possibly reduced customer satisfaction? Finally what is the cost of lost business if data is unrecoverable? I maintain that the drives are FAR cheaper! Hence my mantra:
NO RAID5! NO RAID5! NO RAID5! NO RAID5! NO RAID5! NO RAID5! NO RAID5!
0
 
chapmandewCommented:
raid 5 will never be better than raid 10.  however, for reading only, raid 5 should be fine.
0
 
anushahannaAuthor Commented:
MrMintanet
thanks for explaining the contrast of the real cost of RAID5, in case of a failure. So there is no way to have a RAID5 spare disk to take over in event of a failure?

Tim, do you say "raid 5 will never be better than raid 10" because of the failure possibility? ignoring the failure issue of disks , how much slower would RAID10 be compared to RAID5 for reads?
0
Ultimate Tool Kit for Technology Solution Provider

Broken down into practical pointers and step-by-step instructions, the IT Service Excellence Tool Kit delivers expert advice for technology solution providers. Get your free copy now.

 
chapmandewCommented:
raid10 should be faster for reads and writes.  If you can afford raid10, definetly go with it!
0
 
MrMintanetCommented:
0
 
DavidPresidentCommented:
One can mis-configure any RAID level to be grossly incorrect so it is ultimately going to be slower than any other RAID level. There is more to configuring than just RAID level.

For ultimate performance, on MS-SQL server, you need NTFS set up to do 64KB aligned I/Os, because that is the size of I/O that SQL server requests from the storage stack.

As such, optimal performance on a RAID10 would be to assign a stripe size on raid controller of 32KB.  If RAID5, stripe size is going to be a function of the number of disk drives.  In a 4+1 it needs to be set to 16KB

If you have anything but a bottom-barrel RAID controller, then it will balance I/Os on RAID1 so that it sends request to the 2 disks that can more quickly process a read, so in many cases it only takes 2 disk I/Os to get you the data, and they run in parallel.  Writes will still take 4 I/Os, all run in parallel (depending on RAID architecture)

RAID5 generally reads from all the disks (except for parity) in a single cycle -- but since I/O size is smaller, you have less throughput.  RAID5 has the "write penalty" which you can look up

So bottom line, there is right and wrong way, and optimal way. It all has to be configured properly to get best performance.
0
 
anushahannaAuthor Commented:
by configuring the optimal way, you are talking about
*Stripe Unit size and
*Allocation Unit size, right?

also, can you tell what you mean by bottom-barrel RAID controller? i have never heard that term before..
0
 
MrMintanetCommented:
A bottom barrel RAID controller is simply a cheap and inexpensive RAID controller.  Bottom of the barrel simply means- Junk.

I think we're done here?  Did you need anything else?
0
 
anushahannaAuthor Commented:
Thanks for your helpful insights.
0
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.