Re-assign shared disks FAS2520

[GLOYNSSNS]

Scenario FAS2520-HA, cDOT 8.3

System Setup with internal drives and 1xextra shelf - option chosen during setup to 'manually configure disks later'

Result: Storage balanced 50/50 between each node in HA pair, including the root 'partitions'.

Unable to reassign 'remainder' capacity of internal drives to a single controller as message states:  Cannot remove owner of disk "1.0.x" because it is
shared.

Please can anyone advise what is the best course of action to be able to maximise the usage of the remaining disk space on the internal drives?

Is there somewhere advanced that I can dive into to reassign all partitions to a single node (I take it that this would include root partitions too?)

Thanks, Garth

[Racoon]

Do you want the disks all handled by one controller? And the shelf handled by the other one? If so:

Login into your clustershell

Switch to your nodeshell (system node run)

priv set advanced

aggr status -r to check which disks are handled by your second controller

switch to the nodeshell of which controller currently handles the disks you want to unown

again, priv set advanced

disk show

disk remove_ownership <disk_id>

disk show -n to list all unowned disks

switch to your nodeshell of your prefered controller

disk assign <disk_id> -o unowned -s <sysid>

check again with aggr status -r that only one controller handles all disks, except the root partitions

[eduard_abrahamyan]

It is helped my issue.

Disk was assigned to node where shelf was not connected.

[aborzenkov]

You can reassign data partitions separately if necessary. You still need root partitions for second node. Note that it is recommended to have at least one disk with full ownership (root + data partition) for the purpose of writing core.

[GLOYNSSNS · Accepted solution]

Hi,

Thank you both for your feedback - in the end the answer was in the 8.3 Storage Guide from NetApp.

In advanced mode

storage disk assign -force -data true -owner NEWOWNERNODE -disk 1.0.x -> for each drive/partition

Then create aggregate leaving one free partition for spare space.

Thank You,

Garth