[Beowulf] Are disk MTBF ratings at all useful?

[Lux, Jim (337C)]

You set up 1000 drives, run them at high temperature (using a scaling factor developed by experience) and count how many fail after some length of time, then extrapolate to a failure rate which gets turned into a MTBF. 

It *is* fairly scientific and based on sound principles, although there are some "calibration factors".  
The calibration factors are adjusted on the basis of experience.  Ship a million drives and watch for the failure reports.

Jim Lux

-----Original Message-----
From: beowulf-bounces at beowulf.org [mailto:beowulf-bounces at beowulf.org] On Behalf Of mathog
Sent: Thursday, April 18, 2013 4:02 PM
To: beowulf at beowulf.org
Subject: [Beowulf] Are disk MTBF ratings at all useful?

High end SATA and SAS disks claim MTBF values that work out to over 100 years, and yet it is a common observation that certain models fail at rates entirely inconsistent with those values.  For instance, 75% of all drives of one model dead in < 6 years.  (Cited by one poster in this thread:

https://groups.google.com/forum/#!topic/comp.unix.solaris/zQjoyc8T01Y

).  Additionally, manufacturer warranties at best only go to 5 years, which suggests the manufacturers don't have a whole lot of faith in their MTBF values.

Some of you have huge amounts of storage, how many disk models lasted as long as their MTBF suggests they should?  (Personally we have only one set of disks that are still consistent with the claimed MTBF, a set of 6 Fibre Channel disks that came with a Sun server and are now
10 years old - with no failures.)

How do they come up with the MTBF values for disks anyway?  Clearly it is not based on watching a large sample of disks for countless years!

Thanks,

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech _______________________________________________
Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf