[Beowulf] Are disk MTBF ratings at all useful?
Lux, Jim (337C)
james.p.lux at jpl.nasa.gov
Thu Apr 18 17:33:42 PDT 2013
From: beowulf-bounces at beowulf.org [mailto:beowulf-bounces at beowulf.org] On Behalf Of mathog
Sent: Thursday, April 18, 2013 4:21 PM
To: Alex Chekholko
Cc: Beowulf List
Subject: Re: [Beowulf] Are disk MTBF ratings at all useful?
On 18-Apr-2013 16:03, Alex Chekholko wrote:
> On Thu, Apr 18, 2013 at 4:01 PM, mathog <mathog at caltech.edu> wrote:
>> 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
> How would you do it?
On a brand new design, I have no idea.
On an incremental design, which mechanically is what most disks are, I would take into account the failure rates of the components in the previous generation and extrapolate that to the new design (where possible). For instance, one could tear down and carefully examine the moving components to measure the wear they suffered over several years. They could also do "accelerated aging" studies (running the drives under less than ideal conditions). That method is good for testing paint, but I would not have a lot of faith in it for a device as complicated as a disk.
Accelerated life tests are a very common and validated way to do it. The challenge is in making sure you use the right "acceleration" factors. It's not (usually) a "twice as fast for every 10 degrees"
More information about the Beowulf