March 19, 2009 at 1:00 pm
I am trying to install SQL 2005 64 bit on a cluster that already has an instance on it. I have created a disk resource and a new instance group and have made sure that the resources are only able to be owned by the correct nodes.
When I go to install SQL it fails because it is trying to point to the original cluster's disk instead of it's own.
I do no see anywhere in the GUI under advanced settings where I would be able to tell the installer not to use the incorrect drive.
Anyone ever run into this issue?
March 19, 2009 at 1:02 pm
ok . is this for sql 2005? if its active / active cluster then install secondinstance from second node. make sur eyou can see all drived and info in second node.
:crazy: :alien:
Umar Iqbal
March 19, 2009 at 1:12 pm
This is SQL 2005. 1 Cluster, 2 Cluster Groups, 2 instances with failover. The issue is that the installer is looking for the wrong Disk Resource. The error log shows it trying to access the M: drive which both of the new nodes do not have access to.
March 19, 2009 at 2:03 pm
make sure you change path for binaries and data file when you setup parameter during installation. it might be coming from there. make sure u click advance and for all services change it to drive that is there.
:crazy: :alien:
Umar Iqbal
March 19, 2009 at 2:05 pm
And you put your disks inside the new group you created, and then selected that group during the install? (yes, probably, but just making sure).
March 19, 2009 at 2:40 pm
Yes, I went to advanced install and made sure all of the paths did not have the incorrect drive. It seems to be picking the M: drive up from the cluster somewhere. I will continue to research.
March 20, 2009 at 10:20 am
I think the issue is that we were trying to install two default instances on one cluster.
Here is a quote a co-worker sent me:
SQL Server depends on distinct registry keys and service names within the failover cluster to ensure that SQL Server functionality continues after a failover. Therefore, the name you provide for the instance of SQL Server, including the default instance, must be unique across all nodes in the failover cluster. Using unique instance names ensures that instances of SQL Server that are configured to fail over to a single server have distinct registry keys and service names.
Update: Using a named instance fixed the issue. Thanks everyone for your assistance as always.
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply