[svlug] New server plans moving forward
jeff at frostconsultingllc.com
Sun Jan 21 20:55:46 PST 2007
On Sun, 21 Jan 2007, Rick Moen wrote:
> Now: The existing RAID 1+0 setup _indeed_ can survive loss of half the
> drives prior to replacement and rebuild, but only if you're
> unbelievably lucky about which drives fail: 4/8 * 3/7 *
> 2/6 * 1/5 = a 1.4% likelihood that exactly the right drives fail to save
> your ass. To quote Damon Runyon, "That's not the way to bet."
Actually, I foolishly mentioned the extreme case, but didn't mention the more
likely case which is when you lose 2 drives. In an 8 disk RAID 5, if you lose
2 drives in a short period of time, you're dead for sure even if you have a
hot spare. In an 8 disk RAID10, if you lose 2 drives, it's unlikely you'll
lose the entire array, but still possible (1 in 7?).
So it just depends what you're trying to balance and it sounds like you guys
need space more than that small extra bit of reliability.
> Anyway, you think we have less than a 1.4% chance of keeping a large
> RAID5 array happy? Man, you must think we're real slackers.
I don't think that at all about you guys and don't even know which server
we're speaking of. This was just an academic conversation for me, so again I
apologize for interrupting.
> If I were running the 7-drive RAID5 array of 36GB SCSI drives I spoke of,
> with one hot-spare in the eighth drive slot, and one failed, (1) mdadm would
> send me mail about the drive failure and remirror onto the spare. (Disk I/O
> and CPU load would suck a bit during that remirror -- the price of doing the
> more complex varieties of "md" software RAID.) (2) Within a day or two, I'd
> scrounge a replacement drive, cycle over to the colo, bring down the array,
> replace the failed drive, bring up the machine, and add the replacement as
> the new hot spare.
Indeed everything sounds reasonable and I think that you'll be happy with the
> You'd do differently because you're a performance freak? Bully for you,
> and best of luck with those 1.4% odds.
Actually, I'd do differently because I value reliability over performance over
space in that order.
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