Depending on amount of existing data it may be possible to copy existing volumes onto aggregate on another head before destroying aggregate. Or temporary take all spare and d-parity disks to build aggregate to copy data over.
He has two controllers with 8 and 4 disks each, so I doubt that there are any other aggregates then aggr0. In this case it is impossible without reboot.
But he can leave the controller with 4 disks running and only needs to alter the controller with the 8 disks and then he can reassign two of the spares to the still running controller. So it is just a 50% downtime
Consider of I have no any valuable data on both systems.
I have two aggregates on first node. eric.barlier: "(the one without boot volume of course)" Do you mean If I destroy all of them it would harm my netapp system (I do not use a VM system to boot anything)? What das boot "boot volume"?
So, I deleted one of aggregates, added to the rest aggregate 2 disks and now I have one aggregate with 6 disks and 2 disks in stage "spare" on first node and one aggregate with 3 disks and 1 disk marked "spare" on the second node. How could I reed of these 2 disks from first node to attach them to the second node.