February 20, 2014 at 11:44 am
We will soon be migrating our SQL Cluster disk resources from a HP EVA SAN to a NETAPP SAN and I would like some help
Current Specs:
Servers: 2 nodes each with Windows 2003
SQL: SQL Server 2005
Cluster: Each node is running an Instance/group of SQL in active/active settings.
Resources: Several disks including the Quorum are added as resources to the two cluster group
Story:
We are running on an EVA SAN and just purchased and installed a NETAPP. The SAN Admin is moving LUNs from the EVA to NETAPP using the DTA Data Migration Appliance. I don’t know much but they tell me it’s a block level move and creates an exact replica of the source in the destination.
Anyway they asked me to come up with considerations of how this could affect our cluster and I was hoping that someone may have been through this before.
Honestly I am nervous about using this DTA thing and fear that my cluster resources / disks will not come back on line. The SAN Admin says we can always go back, but I do have concerns about that as well. My ideal is just to add the NETAPP disk/resources to the cluster and do a Detach, move and Attach of all the database; eventually removing the old EVA disks/resources.
Any information would be helpful.
Jeff
February 20, 2014 at 12:07 pm
Let me throw another concern out there for you.
Have they aligned the disks on the netapp prior to doing the DTA move?
I have done migrations using both (or at least similar block move methods) methods and it has been successful.
I am also a fan of having the database backed up prior to doing the full cut over for just in case.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
February 20, 2014 at 12:23 pm
Is disk alignment for everything or just for the clustered resources/disks?
I am not sure if the disks have been aligned or not. NETApp is helping out, so they should have that done.
Yes backup is key. I plan on taking the cluster off line, stopping the services and backing up the data and log files before doing anything.
So you have had success with DTA and that is good news. I know that the SAN Admin has been struggling with the migration and it took about 2 months before he got his first data set moved.
There seem to be a problem with the way the old EVA is presenting the disks and the NETApp; not sure. They were finding that the data would move, but when the servers were brought back on line it was not seeing the data or it was corrupt.
I think they got it working now, but very nervous.
Jeff
February 20, 2014 at 1:03 pm
jayoub (2/20/2014)
Is disk alignment for everything or just for the clustered resources/disks?I am not sure if the disks have been aligned or not. NETApp is helping out, so they should have that done.
Yes backup is key. I plan on taking the cluster off line, stopping the services and backing up the data and log files before doing anything.
So you have had success with DTA and that is good news. I know that the SAN Admin has been struggling with the migration and it took about 2 months before he got his first data set moved.
There seem to be a problem with the way the old EVA is presenting the disks and the NETApp; not sure. They were finding that the data would move, but when the servers were brought back on line it was not seeing the data or it was corrupt.
I think they got it working now, but very nervous.
It would have been far faster to just copy the data files 😀
Any LUN presented to windows should be aligned. I'd be concerned since they are doing block level copy from LUNs that were presented to Win2003 and that were more than likely not aligned.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
February 20, 2014 at 3:46 pm
Acctually they claim that using DTA is faster. Who am i to argue with them
Anyway, I do have another episode to this project and maybe you can help.
We are currently using FalconStore as the Storage Manager in Windows “SAN Disk Management” and we want to remove it.
My first question is: how will this affect the project since all the disks are presented to the two nodes via Falconstore.
My Second question is: how can I remove Falconstore and use either the EVA or the new NETAPP SAN as the Storage Manager in Windows “SAN Disk Management”
When i try to add a new disk it is greyed out.
Your help is appreciated.
Jeff
February 20, 2014 at 9:40 pm
Not sure on FalconStore. I have never used it. Maybe somebody else has and can chime in.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
February 25, 2014 at 3:42 pm
I figured out what Falconstore is. It is an MPIO that is installed on the server. To replace the MPIO on a server you must uninstall the old one and install the new one and you cannot have two MPIOs on a server at one time.
The SAN admin discussed the migration of the cluster resources/disks. He will use the DTA appliance to make an exact replica of the data from the old SAN to the New NETApp SAN. We will then remove the disk resources from the Cluster, disable (not uninstall) the current MPIO Falconstore from the server. Install the NETApp MPIO and I will then add the disk resources back to the cluster using the same drive letters and will find the data available. Spin up the SQL Services and pray.
If anything goes wrong we will remove the NetApp MPIO, enable the falcostore MPIO and add back the old disk resources.
Sounds a little risky.
Anyway let me know what you think
Jeff
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply