Data Backup and Recovery Discussions

rebalance over-loaded luns between aggregates

netappmagic

I have a few aggrs, one with only one volume and a number of luns is extremely over load, and the other aggrs are not so much loaded. I wanted to move some of luns from the busy aggr to the other. What is the best steps/commands to do that? Does that require luns to be disruppted? I don't have two many space left, so, don't have too much free room to play. Thanks for your advices!

12 REPLIES 12

Re: rebalance over-loaded luns between aggregates

aborzenkov

Yes, whatever method you chose it will be disruptive in your configuration. What is the best really depends on how much downtime you can afford, which licenses you have etc.

Re: rebalance over-loaded luns between aggregates

netappmagic

I have snampirror license, that is someting I think it might be related.

Would you please list a few methonds(steps/commands), no  matter how long the downtime would be if it could not avoided?

Re: rebalance over-loaded luns between aggregates

aborzenkov

The first obvious one (and requires no licenses) - stop access to LUN, used ndmpcopy to copy it to another volume, update lun mapping.

If LUNs are in qtrees, you can use QSM to migrate every qtree to different volume.

Re: rebalance over-loaded luns between aggregates

peter_lehmann

vol move

will do the trick...

Re: rebalance over-loaded luns between aggregates

aborzenkov

vol move

To quote original message:

only one volume and a number of luns

Re: rebalance over-loaded luns between aggregates

netappmagic

just wanted to make sure, does "vol move" also move luns? I only wanted to move some luns within a volume, not entire volume. could you show me the syntax to use "vol move" to move luns?

snapmirror will copy the entire volume not on luns level, right? If I use snapmirror, I would have to copy the whole volume first, then remove those un-needed luns at the destination?

Thank you all for your messages

Re: rebalance over-loaded luns between aggregates

resqme914

vol move will move an entire volume (that's why it's called vol move) and all the luns inside that volume will move along with it.  This is non-disruptive.  However, since you want to move only a few luns out of the volume, you can't use vol move.  If non-disruption is a priority, my recommendation is to create new matching luns on the destination/new volume and use host-based mirroring (if available on your OS) to get the data copied over.

Re: rebalance over-loaded luns between aggregates

vims

I've done it a few times, used snapmirror due to the lack of downtime. it was from 30 minutes to an hour, or sometimes even less 10-15 minutes. It depends on server (VM or physical) and snapdrive how fast they are

Re: rebalance over-loaded luns between aggregates

netappmagic

vims,

Here is my quesitons about using snapmirror. I only want to migrate some luns in a volume. As I understand, snapmirror only can copy entire volume not luns. Do you mean that I have to copy the entire volume first, then map only luns I need?

Also, as I understand, I still need to resync in last step while the lun is un-accessable, so, during this short time, there would be still a short downtime, is not it? even though

Re: rebalance over-loaded luns between aggregates

vims

netappmagic wrote:

vims,

Here is my quesitons about using snapmirror. I only want to migrate some luns in a volume. As I understand, snapmirror only can copy entire volume not luns. Do you mean that I have to copy the entire volume first, then map only luns I need?

Also, as I understand, I still need to resync in last step while the lun is un-accessable, so, during this short time, there would be still a short downtime, is not it? even though

1. Yes it is correct. You can delete luns you do not need later on and use luns of your choice. 

2. yes - resync is the vital step. I did it in MS SQL environment, stopped SQL server first - resync - reconnect

View solution in original post

Re: rebalance over-loaded luns between aggregates

netappmagic

Great, vims, Thanks a lot, we are in the same page.

Just wanted to confirm, and make sure I understand you. So, will there be a short downtime when you stop SQL server, in my case, I will have to somehow stop accessing the lun?

Re: rebalance over-loaded luns between aggregates

vims

netappmagic wrote:

Great, vims, Thanks a lot, we are in the same page.

Just wanted to confirm, and make sure I understand you. So, will there be a short downtime when you stop SQL server, in my case, I will have to somehow stop accessing the lun?

yes there will be a downtime, i coordinated it with DBAs. They stopped SQL Servers, then i did my job - run snapmirror update. Disconect drives (luns) and reconnect to the new location. However there are a few steps you have to complete with snapmirror relationship. Once update is completed - then break snapmirror and delete it.

Earn Rewards for Your Review!
GPI Review Banner
All Community Forums
Public