SCSI reliability?

Santilli · Oct 7, 2003

the Cheetah 15K.3's rated Mean Time Between Failures is 1,200,000 hours, which works out to a staggering 143 years

http://techreport.com/reviews/2003q4/cheetah-15k/index.x?pg=1

:mrgrn:

WOW
:eekers:

gs

blakerwry · Oct 7, 2003

I was thinking maxtor was claiming 1,000,000 hours on some of its ATA disks....

not sure exactly how they figure that mean time before failure, but it certainly isn't a measure of the average life of one of these disks.

Santilli · Oct 7, 2003

Quaxtor=drugs
gs

Pradeep · Oct 7, 2003

I believe it's more accurate to say that if were running 100 15k.3s (we wish...), you would get a failure in 1-3 of the drives each year.

Clocker · Oct 8, 2003

You're right on Pradeep. Those numbers are based on a population of drives that are running.

C

Fushigi · Oct 8, 2003

Pradeep said:
I believe it's more accurate to say that if were running 100 15k.3s (we wish...), you would get a failure in 1-3 of the drives each year.

I've got 93 IBM 15K drives across my 2 AS/400s. I lose about 3 drives a year. RAID is a wonderful thing.

Santilli · Oct 8, 2003

Fushigi:
Don't blame Seagate for IBM's high number of failing drives. :lol:

:wink:

gs

Mickey · Oct 8, 2003

blakerwry said:
I was thinking maxtor was claiming 1,000,000 hours on some of its ATA disks....

not sure exactly how they figure that mean time before failure, but it certainly isn't a measure of the average life of one of these disks.

1. Take a hundred* drives, run them through accelerated life testing (elevated temperature/humidity).
2. See how many accumulated hours of "run time" you get between failed drives.
3. Multiply by an acceleration factor (if applicable).

* Guess on my part. I know it's a large number, but I suppose it depends on how quickly you want to reach a certain MTBF rating (or how much test chamber space you have).