2. Setting up the existing controller as a netboot server
3. Physically installing and connecting the new controller
4. Assign an IP, netboot, and install ONTAP on the new controller
5. Install the OS on the new controller and add licenses
6. Run setup on the existing controller to add partner IP
7. Reboot bother controllers and enable failover
8. Clone the config to the new controller
This all makes sense except:
1. The document says the system auto-assigns 2 disks to the new controller. I have 4 spare disks on one shelf which I plan on removing ownership of 3 (one spare) of them for ONTAP. Does it matter what shelf ONTAP is installed on? Will the system auto-assign the unowned disks?
2. Will all my existing aggrs, vols, and data be ok?
3. What about reassigning disks now that there is a second controller?
1. You can use any drives in any shelf. The document tells you to disable auto-assign on existing controller – otherwise it may happen that when you remove ownership controller will re-assign it back (I do not remember scan frequency, I think it is around 5 minutes).
2. As long as you won’t remove ownership from any disk that is part of aggregate your existing data should be OK
3. I am not sure I understand the question. To reassign disks you need to know sysid of new controller. If you have this information, you could reassign them.
Thanks for the reply. To elaborate on the last question, when you do a disk show on an active/active cluster you see some disks are owned by one, some are owned by the other. Correct me if I am wrong but when I add the second controller, won't all disks be owned by the original controller? Is there any drawback to this?
Yes I'm clear about unassigning the spare disks for the ONTAP install of the new controller. But once I complete the procedure and have the active/active cluster up, when I do a disk show, 3 disks will be owned by the new controller, and the doezns of other disks will still be owned by the original controller. Is there any problem with that setup or can I just leave it that way?
you can just leave it that way , no need to worry as it is just the way it supposed to be .. If you see that disk later login to each filer with sysconfig -r , it will clerally tell other disks are "partner" disks.
Better to keep atleat one spare disk in each filer ( each filer should OWN a spare disk )