Hi all,

Yes, Tom is correct. NetApp will send you the shelf without disks and it's a matter of performing the following:

1) Halt system & power down FAS controller & shelves 2) Remove disks from the shelf and set aside (don't worry about the order) 3) Remove cables, taking care to remember where they go (labels are good) 4) Remove the shelf from the rack (get a person to help, heavy) 5) Swap LCR/ESH/AT-FC modules if the new shelf doesn't come with new ones 6) Install the shelf into the rack (get a person to help, heavy) 7) Set shelf-id and loop speed (if applicable) 8) Install cables, taking care to put them in the right ports 9) Install disks, don't worry about the order 10) Turn on all shelves, wait for 60 seconds for the drives to spin up 11) Turn on FAS controller 12) Get a console and make sure the FAS controller boots OK, check for errors 13) Repair affected aggregate(s) and bring them back online

The order of the disks does not matter, as ONTAP/WAFL writes labels on each disk. When the system boots, it will work out where the disks are. It's a pretty cool demo to give. :-)

It is possible to perform the steps above with the system left online, but its not best practise and could cause problems if you don't have multipath cabling to your shelves.

If you've had a shelf failure, chances are that you have an aggregate that is offline and not too happy (inconsistent). The last time I had to swap out a failed shelf, we tried to bring the affected aggregate back online and ONTAP reported that the aggregate was inconsistent, and that we needed to run a wafl_iron over the aggregate before we could bring it online. I would seriously NOT recommend running wafl_iron on a system outside of a maintenance window, as it will peg the CPU at 100% and cause the FAS controller to become unresponsive for short periods of time. Running in a maintenance window should be fine, but its always a good idea to double check with your NetApp support representative before running a wafl_iron.

I hope this helps.

Kind regards, Mike Nye

-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of De Wit Tom (Consultant) Sent: Wednesday, 23 January 2008 11:01 PM To: Robert Lazzurs; toasters@mathworks.com Subject: RE: Failed shelf

If I am correct, then they mean the shelf enclosure. So they will only ship you the enclosure and you need to bring your system down, swap the hard disks from the failing shelf to the new one and then place the shelf in your disk loop.

So you have downtime of a few minutes, but no data migrations is needed.

Grtz, Tom

-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Robert Lazzurs Sent: woensdag 23 januari 2008 13:19 To: toasters@mathworks.com Subject: Failed shelf

Hello,

We have just been informed by NetApp that we have a failed shelf and a new shelf is on the way. I forgot to ask before I got off the phone if we would be able to use our trial license of SnapMirror to allow us to do this shelf swap over without downtime?

This is a FC shelf and we only have enough space to mirror the volumes on to a SATA shelf.

Will this work or am I mad?

Take care.