Active IQ Unified Manager Discussions

Creating new loop with existing disk shelf

praveen121b
5,571 Views

I am planning to remove a disk shelf from existing loop and create a new loop with the separted disk shelf without loosing any data on it.

I have FAS 3140 (only one controller) with 3 loops and the loop which I want to work on consist of two disk shelfs DS14mk4 (300GB FC) & DS14mk2 (1TA AT).

Both the disk shelf are having one aggregate each.

Data Ontap versoion is 7.3.3.

I know how to remove disk shelf from the loop but I need steps for attaching the disk shelf in the new loop with existing data on it.

Please help.

Regards,

Praveen.

1 ACCEPTED SOLUTION

enrjfitz1
5,571 Views

No, the disks are already assigned. It's just a case of powering off the filer and shelves, recabling the existing loop to move out the shelf. Cable the new shelf to the new loop and power back on.

Once booted run the 'storage show disk -p' command to check that all disks are available and accessible and an 'aggr status' command too.

View solution in original post

8 REPLIES 8

enrjfitz1
5,571 Views

Why do you want to do this?

I'd firstly check the RAID group size is not larger than the size of the disk shelf (14 disks). If it is 14 I'd be confortable just removing the shelf and recabling it on another loop.

Are you running anything like SyncMirror on this?? If so, careful consideration is needed.

praveen121b
5,571 Views

I am adding an HBA and NIC card to the storage system.

And I am not running SyncMirror on it.

What about the existing data on the disk shelf which I would remove?

enrjfitz1
5,571 Views

Log into the controller and ensure that the aggregate data does not span the disk shelves by using the command 'aggr status *aggrname* -d' this will show all the disks that are part of the aggregate.

Ensure that the disk shelf you are moving only contains data for one aggregate...Also, using the command 'aggr status -v *aggrname*' check the raid group size is not larger than the number of disks in the shelf, in this case 14.

In other words, ensure that when you move the shelf you are moving one entire aggregate.

If this is the case, relocating the shelf onto another loop should not have any effect as the aggr will be moved in its entirity.

praveen121b
5,571 Views

Aggregate and raid group size is ok, everything is in single shelf. So no problem regarding that.

My concern is about the steps and procedure to follow after separating the disk shelf and adding the shelf to the new loop.

Do I need to do any disk assign command again or anything else?

enrjfitz1
5,572 Views

No, the disks are already assigned. It's just a case of powering off the filer and shelves, recabling the existing loop to move out the shelf. Cable the new shelf to the new loop and power back on.

Once booted run the 'storage show disk -p' command to check that all disks are available and accessible and an 'aggr status' command too.

praveen121b
5,571 Views

Thanks for ur answers.

Anything else I need to do on HBA and NIC cards?

On FC HBA I need to configure the adapters as initiators. Anything else is required?

Shall I take all the shares and aggreates offline before turningoff?

On NIC apart from ifgrp anything else need to be configured?

If you could provide me any document or anylink that will be of more help.

enrjfitz1
5,571 Views

In this order...

Install HBA and NIC.

Configure both as you require.

No need to take aggr offline.

Down the filer and shelves

Do some recabling

Power back on

Check the disks can be seen using the 'storage show disk -p' command

Check all vol's and aggr online using 'aggr status'

praveen121b
5,571 Views

Thanks you very much Sir for all the replys, they are very helpful.

I will come back to you if I encounter any problem or if I have any doubt.

Thanks again.

Public