RE: Converting Mirroring to Raid

["Andy Nolen-Parkhouse" <aparkhouse@xxxxxxxxx>]

Art,

You're right about getting free space for RAID.  If you have four drives
holding your parity bits, then you're looking at 25% of the disk used
for parity.  If you have eight drives, then the figure would be 12.5%.

I would imagine that it takes just as long to calculate and write a
parity bit over empty disk as it does over data-filled disk (anyone know
for sure?).  So the only time you would gain would be the duration of
clearing some data off the parity drives.  In return you're taking the
time to reload a lot of user libraries.

I would see if you can locate a PTF for the ASPBAL command at V4R2.  I
think that it was called DSKBAL in its first iterations.  If you can
balance your data after you come up, I would forgo the mass deletions.
If you can't do the balancing, then I would do the delete and reload so
your data would be spread around a bit.

Regards,
Andy Nolen-Parkhouse

> Subject: RE: Converting Mirroring to Raid
>
> Kirk:
>
> Thanks for the replies.
>
> We are at 92% with mirroring.
>
> We are trying to minimize the time spent watching the system turn raid
on.
> Do you know how long we can expect that to take?
>
> I thought that turning on Raid means the system has to make at least
25%
> space on each drive for the striping, therefore, the more
> data, the longer the process.  Would love to hear that I'm wrong.