Do I understand this correctly?

[Klorese, Roger B.A.]

On Wednesday, June 11, 1997 12:14 AM, rich [SMTP:rich@odysseysys.com]
wrote:
> To all,
> 
> Do I understand this correctly?
> 
> How does the SPARCstorage Array Volume Manager know that the parity data is
> the bad data and the one in need of repair?

The problem described in this bug report is that the parity gets
regenerated incorrectly on the move operation, so it's the parity that
needs to be recreated.  The data doesn't need to be changed in the data
has never been reconstructed.

If you've already had a disk fail and reconstructed the data using bad
parity that was created by a reconstruction, it's too late for a fix --
your only option is to repload from a backup.

(Note, once again, that this won't happen on *all* moves, only moves
which are not stripe-unit-aligned, and that if you have not yet lost a
disk in the volume, parity can be recreated without data loss.)