We have a 2050 that has two physical shelves and we are physically removing the last one in the stack (shelf 2)
There are no aggregates on it, the drives have been zeroed out. Aggr0 and vol0 are on internal drives so that won't impact the controller. The next thing was to remove disk ownership. Disable the cluster shutdown the heads, remove the cabling and then the shelf.
What else am I missing?
Ontap 7.3.7RC1 The shelf is a DS4243
One thing to note we have one drive that is reporting "Not available" from a sysconfig -d also vol status -f does not see and broken disks.
It sounds like you have things in order for removing the shelf. I'd remove the "Not Available" drive before you remove the shelf though. Better to have a clean system before you attempt to remove hardware. With disks that are in a "Not Available" the best way to remove them is simply to pull the drive. The system will then run through it's failed drive process. Before pulling it confirm that it's not in an existing Aggregate and associated with a back end rebuild.
Remove "Not Available" disk
filer>priv set advanced
The blink_on option will light up the second drive light and make it easier to find. Also note the second disk number octet indicates the shelf number.