Re: Problems with SMR hard drives

[Stefan Monnier <monnier@iro.umontreal.ca>]

>> That was not the experience reported in the link I sent.  And I'd expect
>> a RAID rebuild/scrub should result in large sequential operations, which
>> should not trigger the usual SMR weakness.  Maybe it depend on specifics
>> of how the rebuild/scrub proceeds?
> We might be wandering rather off-topic here, but FWIW I have first-hand
> experience of SMR on a mirrored ZFS system: an initial silvering that
> normally took 90 minutes on a 2TB CMR drive took 36 hours on a 2TB SMR
> drive. (I immediately thereafter removed the SMR drive and replaced it with
> a CMR drive.)

Oh, indeed, the link I sent (and my earlier message) did mention that
ZFS mirroring on SMR was a special case where performance really suffers.
Apparently mdadm RAID doesn't have such problems with SMR.


        Stefan