drive selection for disk arrays
bob at proulx.com
Thu Mar 26 18:57:58 UTC 2020
David Christensen wrote:
> Have anyone seen a failure involving multiple similar drives all failing in
> the same mode at the same time?
For a corporate setting with an SLA and so forth the usual solution is
enough drives as hot spares and a fast enough SLA response time to
replace drives quickly before too many fail. But even so I have twice
seen large corporate arrays with multiple drives failed. They weren't
running ZFS though and so didn't detect it until too late. Then they
had to restore from backup.
Twice now I have had two sibling disks that started out brand new
together with relatively close serial numbers in a RAID1 configuration
fail within a week of each other. Both times they were in a rack
environment in a controlled access room. One was mine and I caught it
soon enough with a replacement. One was a client site where they did
not and it became a restore from backup task for me.
For myself I always buy dissimilar drives to decouple failure modes.
If that is not possible then I remix older drives into a set to
decouple failure modes. For myself I would rather have one brand new
drive with one older drive than two brand new drives. Regardless I
always ensure that backup is operating properly.
More information about the freebsd-questions