Someone had posted the following info on this list in mid 2006. I have used this technique to migrate a few trad vol to flexvols and retain the snapshots. Hope this helps
<Quote>
If your data is inside qtrees on the tradvol, there is a clean way to migrate
everything from tradvol to flexvol. Here's the basic idea:
- stop all scheduled snapshotting on source & dest
- do a QSM initialize from your tradvol qtree(s) to your qtrees on the newly
created flexvol, but - and this is important - use the oldest snapshot on the
source as the baseline snapshot for the "snapmirror initialize" operation(s)
- afterthe transfer(s) has/have finished, take a snapshot on the destination
(same name as on the source)
- next, do a QSM update operation for the relations from your tradvol to your
flexvol, but use the "next" snapshot as a baseline for that snapshot
- after those transfers have finished, take a snapshot on the destination again.
In the end, you will have all your snapshots on the destination. The dates will
be all messed up, as those snapshots will all be taken on the same day,
even if they cover months of snapshot differences in data.
Now, if your tradvol data is not in qtrees, there is always a possibility to
do vol-to-qtree snapmirror, by using "-" as the source qtree. (Read up on it,
as this won't transfer data that isn't already in a real qtree - that
data will need
to be transferred separately). You will end up with a different path to your
data though. There's no way around that.
Also, I don't know how this migration procedure will work with your existing
snapmirror relations. Don't know if you'll be able to just move them over to
flexvol. Might work, but you need to test it out...
<end quote>
On 3/21/07, Chris Thompson <cet1@cus.cam.ac.uk
> wrote:chris.holloway@DNSArrow.co.uk
writes:
[...]
> Since you have space for both the traditional volumes and the
> aggregates, could you just not migrate the volumes contents to the new
> volumes and start snapshotting there, and at the same time retain the
> old volumes for a period of time until you don't require the old
> snapshots anymore.
That's what we did, and we didn't _really_ have enough space. I had
to downgrade the old traditional volume to raid4, _and_ run without
hot spares. Oh, and turn off autosupport to NetApp until it was all
over, so that it wouldn't keep automatically opening new support
cases ... :-(
--
Chris Thompson
Email:
cet1@cam.ac.uk