Moving a shelf, with disks, aggr and data from 2020 to 2040

[brettsnow]

I have a DS14mk2AT with 1TB SATA drives attached as aggr1 to a single controller FAS2020.  I want to move to that shelf, all of the disks the aggr and the data to a dual controller FAS2040.  Is there a procedure to do this?

[sharmadevender]

May be this will help you...

Physically Moving a volume

To physically move a volume, complete the following steps.

Step	Action
1	Enter the following command at the source filer to locate the disks that contain the volume: vol status vol_name -r  Result The locations of the data and parity disks in the volume appear under the volume name on the same line as the labels Data and Parity.
2	Place the volume offline and follow the instructions in the disk shelf hardware guide to remove the disks from the source filer.
3	Halt and turn off the destination filer.
4	Install the disks in a disk shelf connected to the destination filer.
5	Turn on and boot the destination filer. Result   When the destination filer boots, it places the foreign volume offline. If the foreign volume has the same name as an existing volume on the filer, the filer renames it volume_name(1), where volume_name is the original name of the volume.   Caution If the foreign volume is incomplete, repeat Steps 1 and 2 to add the missing disks. Do not try to add missing disks while online--doing so causes them to become hot spare disks.
6	If the filer renamed the foreign volume because of a name conflict, enter the following command to rename the volume: vol rename vol-name new-name    vol-name is the name of the volume you want to rename.   new-name is the new name of the volume. Example   The following command renames the users(1) volume as newusers: vol rename users(1) newusers
7	Enter the following command to bring the volume online in the new filer: vol online vol-name    vol-name is the name of the volume. Result:   The following events occur:   The volume is online in its new location.   Entries in the /etc/exports file and shares that refer to the volume become incorrect.   The in-memory information about active exports gets updated automatically, and clients continue to access the exports without problems.
8	Enter the following command to confirm that the added volume came online: vol status vol-name    vol-name is the name of the volume.
9	If you have...	Then...
	An NFS-only filer	Go to Step 11.
	A CIFS-only or multiprotocol filer	Create shares that refer to the new volume to enable clients to access it, then go to Step 10.
10	If you have...	Then...
	A CIFS-only filer	You are done.
	A multiprotocol filer	Go to Step 11.
11	Update the system /etc/exports file with the new volume.
12	Run exportfs.
13	Add the appropriate mount point information to the /etc/fstab or /etc/vfstab file on clients that mount volumes from the filer.

-DEV

[CCOLEMAN_]

cool!

is it ever possible for the physical disks to be associated with another volume as well?

[sharmadevender]

What kind of association ? Request you to elaborate

[CCOLEMAN_]

I'm just curious if data from more then 1 volume can exist on a single disk?

and if it can, what happens when  you start pulling out the physical disks? Maybe you're forced to move it as well?

[sharmadevender]

Yes that volume would also be forced to move with disks.For saving associated volumes, we need to have a copy of associated volume within the filler, to avoid such forceful physical migration situations.

[brettsnow]

I've been asked for more details;

FAS2020 is single controller

Data OnTap 7.2.4L1

12 internal 300GB drives - aggr0 is here

one expansion shelf - DS14mk2AT with 4 1TB drives - aggr1 is here

all data in aggr1 (DS14) is CIFS.

FAS2040 dual controller

Data OnTap 7.3.6

12 internal 600GB drives - aggr0 (cn1) and aggr 0 (cn2)

I want to move the DS14mk2AT with 4 1TB drives with the volumes, aggr1 and all data.  This aggr is NOT shared with the other aggr on the FAS2020, so all volumes and data in that aggr are on DS14 and 1TB drives.  So I want to move the shelf with the drives, not just the drives themselves.

I am hoping it is as easy as offline all volumes in the DS14 aggr.

Off Line aggr on DS14

Halt 2020

Physically move DS14 cables from FAS2020 to FAS2040.

Bring up new DS14 storage on FAS2040

Hopefully all data will be in place and all CIFS ACL info will be consistent.

Verify all CIFS data can be seen/shared with all existing hosts.

The 2020 and 2040 have a SnapMirror relationship today, and the CIFS data is on the 2040, but on aggr0 on the 600GB drives.  It needs to be on the 1TB drives.  So hopefully by moving the shelf and the 1TB drives, the data will be intact and moved to the 2040.

Does this make sense.  Does this sound workable and reasonable? 

[mrousseau]

2020 : halt

2040 : cf disable then halt both nodes

plug ds14 shelf into 2040

SERIAL or RLM : boot 2040 into maintenance mode (second CTRL-C at boot and choose option 5 ). At question "continue with boot" press Y

At maintenance mode prompt :

"disk show -v" (to check that all 1Tb disks are present)

"disk reassign" to change old disk serial by new one - serial number of your controller is available with "sysconfig -v"   Syntax:

disk reassign {-o old_name | -s old_sysid} [-n new_name] [-d new_sysid]

halt

at loader prompt : boot_ontap

once booted :

check everything is there "aggr status" "vol status" "sysconfig -r" then re-enable cluster : "cf enable"

done