Scenario 1: You have a filer that was sent with a dedicated root aggregate and somebody has messed it up. I have seen it with customers a lot; someone puts something on the root aggregate, or extends it improperly—whatever. And now you are tasked with rebuilding the root aggregate.
Option 1: Non-disruptive (NDU, though technically without the U—ha!) move.
Scenario 2: Your root volume is consuming all your space and you want to rebuild it to make space for some rather insignificant stuff: boot from SAN volumes, ISOs, etc.—stuff without a lot of traffic that may only get touched every once in a while, so you feel okay with sticking it on the root aggregate.
Option 2: NDMP copy the volume to a new volume and reboot through Takeover/Giveback.
Scenario 1: Vol Move
Let's say you were moving a root aggregate to a new location; you had a root aggregate with 5 disks, and you want to use only 3, and you have some fresh disks to use. What is the full sequence of steps you would take?
One downside to this method is there is not really a way to check the progress; there is only a "setup" status that shows the volume is moving and awaiting cutover. Once you have confirmed everything is good, offline aggr0_old and then delete it.
Scenario 2: NDMPCOPY
Start off by examining your NDMP options:
And when it is done, there are but two things left to do: mark the new vol0 as the root volume, and then takeover the filer and giveback to implement the change.