Optimal RAID setup for SQL server

Question

We have an SQL 2005 database backend for our website, currently about 10GB in size. There are a lot more reads than writes, though I don\'t have the exact statistics.

We

Answer 1

I prefer raid 10 which i've already configured on my dell r210-II linux server. Amazing performance...

Answer 2

Given the small size of the database I would use four 15krpm 2.5" SFF SAS disks, setup as two separate RAID 1 mirrors. I'd run them through something like an Adaptec 5805 PCI-e x 8 SAS controller. I'd put the data on one array and the logs on another for safety. Bar memory-mapping the whole database, a very large/expensive SAN or using SSDs I'm pretty convinced this will be just about the quickest setup for the money.

Hope this helps.

Answer 3

Two things,

The theory is RAID 1 gives you two drives for a read, but dont' think that equals two times the performance. As a matter of fact, it doesn't, usually the sequential read spee ends up being exactly the same as one drive. Surprising, but true....do real world tests, don't rely on theory.

With that said, I would go with two RAID 1...but not as you said, one for OS, and the other for data. I would have some small OS partition on one of them, but definately give both raids to sql server for data. Give sql server all the sets you can.

SQL Server can stripe across the pairs, you absolutely do not want to go with 0+1 despite what anyone here says. They are again looking at theoretical benchmarks, not understand that sql server can stripe all its data across the disks, and can do so in an optimized way.

Raid 1 is just for data redundancy...other than that, give as much as you can to sql server, and let it do its thing...its very good at what it does.

You have a 0+1 and break it into two raid 1....do real world testing, you'll be surprised at which is faster for sql server work.

Answer 4

I'd spend a little time making sure your database is efficiently indexed; monitor it for Full Table Scans, make sure the most frequent queries read as little data as possible; I'd take an efficiently designed database on a slow disk setup over a database with improper indexes on SSDs.

Answer 5

Given that there are a lot more Reads than Writes, RAID 5 will be faster for your data files.

If you possibly can, use a separate RAID 1+0 array for the transaction log files.

EDIT: you will need at least 3 Disks to create a RAID 5 array.

Answer 6

I would also recommend establishing a set of baseline values which allows you to both do some more founded predictions than just putting your finger in the air (or asking StackOverflow, seems vaguely and oddly similar those two concepts).

With some baseline values you can also measure the effectiveness of your changes and be a good foundation for future upgrades...

(If not known, pick up some literature (or wikipedia) on Performance Engineering. It's a whole branch of stuff that deals with just this type of problems).