[svlug] New server plans moving forward
jeff at frostconsultingllc.com
Sun Jan 21 18:41:56 PST 2007
On Sun, 21 Jan 2007, Jeff Frost wrote:
> On Sun, 21 Jan 2007, Daniel Gimpelevich wrote:
>>> As is typical for Linux servers, our Internet presence has so little
>>> need for CPU power that the RAID overhead really won't matter at all, by
>>> the way -- and wouldn't using RAID5, either (which was our original
>>> intent for the JBOD, to maximise usable space).
>> Huh? Doesn't RAID5 (parity) _reduce_ usable space, given the same
>> combination of drives?
> RAID 0 yields the disk space of n drives (6 36GB yield 216GB array)
> RAID 5 yields the disk space of n-1 drives (6 36GB yield 180GB array)
> RAID 6 yields the disk space of n-1 drives (6 36GB yield 144GB array)
Oops, that should read:
RAID 6 yields the disk space of n-2 drives (6 36GB yield 144GB array)
(the calculation was write, but not the formula)
> RAID 0+1 yields the disk space of n/2 drives (6 36GB yield 108GB array)
> RAID 10 yields the disk space of n/2 drives (6 36GB yield 108GB array)
> RAID 50 yields the disk space of n-2 drives (6 36GB yield 144GB array)
> Although, the RAID 50 calculation above assumes a stripe of two RAID 5
> volumes...the actual calculation for RAID 50 space is:
> (Size of Drives) * (Number of Drives In Each RAID 5 Set - 1) * (Number of RAID
> 5 Sets)
> RAID 10 is a big winner for write speed as well as reliability.
> Software raid on modern processors is of neglible performance impact. In
> fact, linux software raid is faster than many supposed hardware raid
> solutions. The only down side is the lack of hot swappability.
> Well performing hardware raid solutions that I've had experience with include:
> Areca, LSI and 3ware.
Jeff Frost, Owner <jeff at frostconsultingllc.com>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
More information about the svlug