Be careful which node you install a clustered instance from

While installing a clustered SQL Server instance recently I ran across a bit of an annoyance.  When I was going through the SQL Installation process I didn’t pay any attention to which node of the cluster currently owned all the disks that I was planning on using for SQL Server.  Because of this when I got to the Data Directories screen none of the drives would validate.  This was because all of the disks were owned by the other node of the cluster.

There’s a few different ways to solve this problem.

  1. Run the installer on the other node.
  2. Move the available storage to the node you want it on via the command line.
  3. Down the node which is hosting the storage.

Option 1 pretty much sucks as you are about 1/2 way through the installer by the time to figure out this is a problem, but it will work.

Option 2 is a great solution, but you can’t do it via the Failover Cluster Administrator in Windows 2008 R2.  You have to use a command line tool called cluster.exe.  The syntax is as follows:

cluster group “Available Storage” /move:{New Node Name}

Option 3 is the fastest.  Just right click on the node in the Failover Cluster Administrator and select more options, then “Stop Cluster Service”.  Everything will failover to another node.  If you have a two node cluster this works fine, if you have a cluster with 3 or more nodes this probably won’t work very well.

Hopefully if you run into this annoying little issue, this will help you solve it.

Denny

Share

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Trust DCAC with your data

Your data systems may be treading water today, but are they prepared for the next phase of your business growth?