Move volume?

[pbutler6163]

I have a netap 3020 with a mix of scsi and sata disks. My storage on my scsi is getting too full and I just purchsed an additional shelf of sata disks. I would like to move the data from the scsi volume to the sata and retain permissions as well as the old vlume name if possible. can this be done easily and how?

[adamfox]

You can. You can use either Volume SnapMirror (if you have a SnapMirror license) or vol copy if you do not to move the data. Then you can make the volume name the same with the vol rename command.

If you have NAS exports, you may still have to recreate those as vol rename will typically change your exports as well. Note: This is not transparent and data will need to be inaccessible (or at least read-only) during the data move (or only for the final data move in the case of SnapMirror). But both are block-level moves so permissions should stay the same.

NetApp Professional Services can also come in on a contract basis and do the moves for you if you choose. But bottom line, it is posible.

[BrendonHiggins]

Do you have any luns on the volumes?

If so you will need note the disk ID [serial_number] of the lun.

lun show -v

Power off the server, offline the lun, move the volume to the new aggregate then

lun serial lun_path [serial_number]

Online lun, add igroup, cifs power on the server

[BrendonHiggins]

Let me know if you need anymore help as I have just moved 30+ volumes and have a script to speed up the process.

[pbutler6163]

Actually Yes that would be great if you could email to me at patrick.butler@lonza.com ?

[BrendonHiggins]

Email sent

[navinita1]

Hi ,

I saw your update in one of the netapp community. I want move some of my volumes from one NAS to another to balance the load.

It will be really great if you can email me the procedure

Thank you so much

Navin

nasingh@nsf.gov

[AtherosComm]

can you send me the script to move volumes?  coyaba_99@yahoo.com.  Thank you.

[dboutorwincat]

There is a relatively easy way to do it (assuming no LUNs of course):

1. Add SATA shelf

2. Create new aggregate using the newly added SATA drives

3. Create new flexible volume on new aggregate called "original_volume_name_new"

4. NDMP copy the contents of the original volume to the new volume

5. Schedule an outage for the folks who access the volume you want to move

6. NDMP copy level 1 the contents of the original volume (level 1 only copies changes since baseline copy); should be very fast

7. Rename source volume to "original_volume_name_orig"

8. Rename destination volume to "original_volume_name

9. Re export your nfs filesystems

[pum]

I would point out that using ndmpcopy in this manner will probably not give you an identical volume copy.

Running the ndmpcopy level 1 will copy over any new files, but will not delete any files that have been deleted on the source volume in the time since the level 0 was run.

Net effect is that there will likely be some extra files after this process.

In my opinion, using SnapMirror within the same controller is a better way of doing this, as SM will give you that short cutover time as well as properly reflecting deleted files on the destination.

Temporary license keys for SnapMirror are easy enough to generate.

[adamfox]

If you want an exact copy and can be down for a single transfer (or at least read-only), then you could also use vol copy which does not require a SnapMirror license and is basically a level-0 only version of volume snapmirror. There is also an aggegate level version of this called aggr copy.

Just another option.

[dezineforce]

Would the vol copy also retain snapshots of the volume?

And if this answer to this is yes can I use it to move an existing snapmirror source volume to a new aggregate, retain its name (rename after move) and then just carry on snapmirroring it without having to reinitialise it?

[BrendonHiggins]

Yes ~ the snaps will still be there on the new volume.

[dezineforce]

Perfect answer, thanks

[pum]

'vol copy' will NOT copy snapshots from the source volume.

'ndmpcopy' will NOT copy snapshots from the source volume.

QSM (qtree SnapMirror) will NOT copy snapshots from the source volume.

Only VSM (volume SnapMirror) will copy snapshots from the source volume.

[adamfox]

Not true, my friend. Both VOLUME snapmirror and vol/aggr copy will copy snapshots. These are physical copies rather than logical copies. NDMPcopy and QTREE SnapMirror are logical and thus will not automatically bring over snapshots. It's possible to do, but it's really tedious.

If you are doing VOLUME SnapMirror, you can copy it over and you should not have to re-initialize as long as you have a common snapshot between the new copy and the other side which is usually the case. Migrating with Qtree SnapMirror is not as forgiving and would require an re-init.

Hope this helps.

[pum]

Hmmm. That doesn't match up with what I see.

When I tested 'vol copy' on the two systems I have immediate access to (R150 running 7.1.3 and FAS2050 running 7.3.0) - the resulting volume does not have the snapshots that the source volume does.

This was tested with flexvols - I'll try tradvols to see if there is a difference there, although I would not expect one.

mke2050> vol create src aggr1 100m
Creation of volume 'src' with size 100m on containing aggregate
'aggr1' has completed.

mke2050> snap create src snap1
mke2050> snap create src snap2

mke2050> snap list src
Volume src
working....

%/used %/total date name
---------- ---------- ------------ --------
29% (29%) 0% ( 0%) Oct 10 10:01 snap2
44% (27%) 0% ( 0%) Oct 10 10:01 snap1

mke2050> vol create dst aggr1 100m
Creation of volume 'dst' with size 100m on containing aggregate
'aggr1' has completed.

mke2050> vol restrict dst
Volume 'dst' is now restricted.

mke2050> vol copy start src dst
Copy Volume: src on machine 127.0.0.1 to Volume: dst
VOLCOPY: Starting on volume 1.
10:01:53 CDT : vol copy restore 0 : begun, 0 MB to be copied.
10:01:53 CDT : vol copy restore 0 : 100% done, 0 MB copied.

mke2050> vol online dst
Volume 'dst' is now online.

mke2050> snap list dst
Volume dst
working....

%/used %/total date name
---------- ---------- ------------ --------
27% (27%) 0% ( 0%) Oct 10 10:01 snapshot_for_volcopy.4

[pum]

Traditional volume behaved the same as the flexvol - with 'vol copy' not copying snapshots.

After trying the tradvol, I tried the 'aggr copy' - which, honestly, is not a command I had ever tried before.

'aggr copy' did copy the snapshots for any flexvols contained within.

'aggr copy' did not copy aggregate snapshots, however - not that anyone uses them. 🙂

-daniel

[scottgelb]

With aggr/volcopy use the "-S" option to copy all source snaps the same as VSM. "vol copy start -S srcvol dstvol" if you don't specify -S it takes one snap only. "-s snapname" lets you take one specific snapshot and with no -s it creates a snapshot for the copy. use uppercase -S to take all snaps on the source. It's the same as a VSM initialize except you can't do incremental updates.

[pum]

Ahhh.

Well color me stupid...

I'll check the man pages before posting again.

Thanks for the info.

[harperadams]

I have successfully copied a volume from FC to SATA drives on the same filer but am having difficulty copying between filers (BOTH 3020)

I have two filers and need to move a guest from the first to the second so I can free up some FC space on the first filer and use SATA on the second.

I have created a destination volume of the right size and restricted it.

I ran the following (edit to hides real names!)

san02> vol copy start -s nightly.0 san01:orgSERVER newSERVER
This can be a long-running operation. Use Control - C (^C) to interrupt.
Copy Volume: orgSERVER on machine san01 to Volume: newSERVER
san01: Connection refused
VOLCOPY: Could not connect to filer san01.
VOLCOPY: Cannot Init Input, aborting

Clearly it looks like san02 wants to authenticate with san01 but can't

No sure how to fix this.

I have root logins for both sans.

Help much appreciated

thanks

Howard