RAID-Z in a disk-failure.

[Bob Friesenhahn]

On Sat, 26 Mar 2011, Anders Andersson wrote:
>
> Have I missed something in this scenario?

I think that your summary is pretty accurate.  However, it is worth 
noting that zfs stores redundant copies of metadata blocks, and may 
optionally store redundant copies of user data blocks.

Zfs offers raidz2 for the same reasons that storage array vendors 
offer RAID6 (vs RAID5).  With today's large disks, a secondary data 
failure is not unlikely.

If you use zfs mirroring or raidz1, then a periodic zfs scrub is 
recommended in order to decrease the possibility of secondary data 
failure.  I would definitely do an initial zfs scrub after initially 
populating a pool with data.

Bob
-- 
Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/