At 8/16/05 04:18 PM, Lex0r wrote: You said the old one had:
3 x 73gb - which is a total of: 219GB
And the new one has:
6 x 36gb - which is: 216GB
Actually what I said was:
3 x 73gb 15k drives in Raid 5 vs. 6 x 36gb 15k drives in Raid 10 plus one hot spare
RAID stands for Redunant Array of Inexpensive Disks or Redundant Array of Independent Disks (depending on who you ask). The RAID levels are important to note when computing disk space. In a RAID 5 configuration one disk is used for parity information therefore your effective space is N-1 or in this case, 146gb. In a RAID 10 (or 0+1) configuration, you are essentially striping a mirror (or mirroring a stripe) and your effective space is N/2 or in this case 108gb. For more details on RAID check out:
http://en.wikipedia...of_inexpensive_disks
It's a great read for anyone who's into technology or would like to be. You'll get a better understanding of why RAID is used and why having a hot spare is beneficial (especially when the datacenter is 30 minutes away).
Anyway, back to your initial question. While we will essentially have less space than before, storage space is pretty much our last concern when it comes to the database server. What we're most concerned about is input/output wait. I/O Wait is the time the cpu's spend waiting for data to be retrieved. For our database server, this is the largest contributor to slow query responses. By moving from a 3 disk raid5 config to a 6 disk raid 10 config we'll be speeding up the disk access quite a bit in hopes to keep those queries moving along at a nice clip.
Oh, just a quick note for the linux users out there. I had forgotten to mention the move from EXT3 to ReiserFS for the database storage and log arrays. This too will speed things up a bit.
I hope this helps to explain things a bit better, but if not, ask some more questions and I'll see what I can do ;)