SQL Server 2008 R2 Cluster

  • Folks,

    I have the following requirements and was wondering if anyone had any advise as to how to get this done. I have not worked with Clusters since 2005 so I am a little rusty. Here is the scenario.

    Windows Server 2008 R2

    SQL Server 2008 R2

    Active Passive

    We are currently using local drives x, y, z for logs data and quorum. I need to switch to the SAN due to space constraints and would like to use the same drive letters, what is special that I need to do? System Databases are on Y. Any help would be appreciated!

    Thanks

  • Hi,

    If you plan to remove your local drives from the machine, I do not see any negative impact if you wish to use the same drive letter. But I would recommend you to name your new SAN drive other drive letter instead. It is recommended to host your database files in SAN drive instead your local drives in order to gain better database performance.

    I hope this will help. Just my 2 cents. Thanks.

    Regards,monkeyDBAWe dont grow when things are easy; We grow when we face challenges.

  • I would take SQL Offline leaving the disk resources online.

    I would mount the new drives in as different drive letters.

    I would add the new drives to the resource group.

    Then I would copy the data over to the new drives.

    Then I would rename all the drives into the letters I wanted.

    I would bring SQL back online and make sure it comes up, if it has any problems resolve them

    I would then remove the old drives from the resource group.

    I would fail it back and forth a couple times just to be paranoid.

    CEWII

  • We are currently using local drives x, y, z for logs data and quorum. I need to switch to the SAN due to space constraints and would like to use the same drive letters, what is special that I need to do? System Databases are on Y. Any help would be appreciated!

    backup your SQL Server database with SQL Server backup software [/url]and then save the backup copy to different location, enlarge your original partition with Partition software[/url].

  • Not a particular fan of the backup restore method in this case, it seems like a lot of wasted effort and risk.

    Using the process I outlined above you have BOTH copies at the same time so if you need to rollback you can, pretty easily. With a backup restore almost as soon as the restore starts you are pretty much committed for that database, you can't really rollback.

    While I agree it is a possible method, I don't agree with it as a recommended method.

    CEWII

  • Elliott Whitlow (10/27/2011)


    I would take all SQL resources Offline.

    I would mount the new drives in as different drive letters.

    I would add the new drives to the resource group.

    Then I would copy the data over to the new drives.

    Then i would delete the old drives and the new drives from the cluster group and swap the drive mappings over (do this in pairs as the sql resource will need at least one disk resource present each time)

    Then i would ensure all the new disks are added back to the group with the old drive letters retained. Ensure dependencies are set

    I would bring SQL disk resources back online first and make sure it comes up, if it has any problems resolve them. Fail back and forth a couple of times

    I would then bring SQL Server service online and if there are any problems resolve them

    I would fail it back and forth a couple times just to be paranoid.

    CEWII

    As Elliott says but with just a small addition in red

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • Perry Whittle (10/31/2011)


    Elliott Whitlow (10/27/2011)


    I would take all SQL resources Offline.

    I would mount the new drives in as different drive letters.

    I would add the new drives to the resource group.

    Then I would copy the data over to the new drives.

    Then i would delete the old drives and the new drives from the cluster group and swap the drive mappings over (do this in pairs as the sql resource will need at least one disk resource present each time)

    Then i would ensure all the new disks are added back to the group with the old drive letters retained. Ensure dependencies are set

    I would bring SQL disk resources back online first and make sure it comes up, if it has any problems resolve them. Fail back and forth a couple of times

    I would then bring SQL Server service online and if there are any problems resolve them

    I would fail it back and forth a couple times just to be paranoid.

    CEWII

    As Elliott says but with just a small addition in red

    You need to take SQL, Agent, and maybe SSAS offline, BUT you need to leave ALL the disk resources for that resource group online. This would leave that resource group in a "partially online" state and this is ok.

    At no point do the disk resources need to be taken offline, you can change the drive letters at will.

    Personally I would wait to remove the disks from the resouce group until after I was sure all was well. The would exist in the resource group with different letters, so no big deal.

    CEWII

  • take your point but dont agree.

    we'll agree to disagree on this one 😉

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • And I'm ok with that.

    CEWII

  • Thanks guys I greatly appreciate your feedback.

Viewing 10 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic. Login to reply