Migrate SQL CLuster from HSG80 to EVA5000

[Lee Messenger_1]

Hi,

We are migrating data from an HSG80 to an EVA 5000.

The HSG80 currently holds the data for our SQL Cluster.

So, I basically need to move all this data from the HSG disks to vdisks on the EVA, and get the cluster to use these disks as opposed to the ones on the HSG80.

Has anyone done this or could point me to information on how to do this.

Thanks

Lee

[Steven Clementi]

Lee:


Yes, several times.


Basically, you would present the new vdisks to one system, copy the data over to them, remove access from the old disks, run the cluster recovery utility to tell the cluster about the new disks (or you can use dumpcfg to export the disk signatures and then import them onto the new drives) and all should be good.

I can go into more detail where necessary. I don't have a doc or a reference to a doc that I can post.



Steven

[Lee Messenger_1]

Steven,

Thanks for your reply. I'm glad someone out there has done this.

Ok, if you don't mind, could you check through the steps listed below and tell me if I'm going wrong anywhere.

Firstly, this is a 2-node active/passive cluster.

1. Do I need to shut down the passive node before I do anything ?

2. Stop SQL Server

3. Present Vdisks to active node

4. Copy data to new disks

5. Unpresent HSG80 disks from active node

6. Run cluster recovery util to point SQL cluster to new disks

7. Change new disks drive letters back to original

8. Bring passive node back up.

I had also read somewhere that something funky has to be done if running MSDTC (which we are, and its installed on the quorum disk).

How does that sound ?

Thanks

Lee

[Lars Hunold_1]

Sounds good. But I would change a few steps

1. Present vDisk on active Node
2. Create cluster disk ressource and bring them online
3. Present vDisk on passive Node
4. Move vDisk to passive and check it.
5. Stop SQL-Server
6. Copy data
7. stop all cluster disk ressource (only quorum should be online)
8. Run cluster recovery util to point SQL cluster to new disks
9. Start SQL-Server on active node
10. Move SQL-Server to passive node
11. Remove old HSG80 cluster disk resource
12. Unpresent HSG80 disks from both node