Re: MTBF interpretations (Re: ZFS performance)
On Sat, 12 Nov 2022 07:52:32 -0500 Dan Ritter <dsr@randomstring.org> wrote:
> No, my interpretation is that the average (mean) lifetime
> between failures should be the listed value. At 114 years, half
> of the population of drives should still be working.
>
> This is obviously not congruent with reality.
I'd say it's like with many things in life: If expectation doesn't match the definition, outcome is "not congruent with reality".
MTBF is, like Stefan already said, not a lifetime expectation.
> [...]
> Some drive models are lucky. Some are unlucky. Overall you
> should expect about 1.54% of disks to fail each year in a large
> mixed-age population, not quite double what you estimated. But
> the range of annualized failure rates is from 0 to 9% -- you
> could be lucky, or very unlucky.
That's what is to expect here. They are using desktop drives in enterprise environment. This naturally gives results not foreseen by the disk vendor. Even for some of the "sub par" disks the result is probably fine. And Backblaze itself do know this. They safe money with having some more failed cheap disks like the ST4000DM000 - a disk with a power-on-ratio of 9/5 instead of 24/7 (2400 hours per year) and no MTBF rating (AFAICS)?
I'd simply not mix this into some 1.000.000 hours MTBF rating expectation...
hede
Reply to: