• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 356
  • Last Modified:

Raid configuration assistance.

I have a HP server with a two dual channel UW320 raid controllers. The controller is then connected to two disk-cabinets containing each 14 disks. Any tips on how to plan the logical raids? See below for my current config.

Current
Controller 1, Channel 1: 7x32Gb disks, all seven disks in a raid 5, no spare
Controller 1, Channel 2: 7x16Gb disks, all seven disks in a raid 5, no spare
Controller 2, Channel 1: 7x16Gb disks, all seven disks in a raid 5, no spare
Controller 2, Channel 2: 7x16Gb disks, all seven disks in a raid 5, no spare

The machine is to be used as a long term lab enviroment where I will be testing and deploying a large number of machines. The machines will not be moved to a production enviroment later on. This is stricly for lab usage.  My reason for doing raid 5 is that I wish to maximize the space while still have a simple redundancy.

Comments?
0
Eidron1980
Asked:
Eidron1980
  • 2
1 Solution
 
Andres PeralesCommented:
I would add hotspares to your raid containers.
0
 
Eidron1980Author Commented:
Well, I thought about that, but since its a lab its not THAT importent if a disk breaks, and actualy two would have to break (and get broken un-noticed) before failure of data...
0
 
BogdanSUACommented:
The above suggested configuration will yield the following volumes (approximate raw unformatted size)

192GB
96GB
96GB
96GB
for a total of 480GB.  Not too shaby for a lab env.

I can understand your reasoning for going RAID5.  But try not to exceed more than 7 disks in a RAID5 volume, otherwise writing that parity bit will cause some serious performance hits.  I know you abided by this guidline, but I'm giving you a heads up in case you ever consider bumping it up more.

If you really don't want to use a hot spare, at least consider going out on ebay or some place and purchasing a few spare 32GB disks (like 2 or 3).  That way you'll have "warm" spares and you won't have to go out and find them, get approval for them, order them, ship them, wait for them, etc.

In the event of a failure, you can nip it in the bud before it mushrooms into data loss.

0
 
Eidron1980Author Commented:
Yeah, warm spares are stacked high. =) SCSI disks aren't that expensive anymore.
Did actualy not know about the >7 disk performance hit, how come this will cause performance hit?
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now