Installing SCCM 2007 on a Remote Database Cluster (Part 2 – Installing SQL on the Cluster)
I was able to get this finished much faster than I thought. You gotta love production change freeze windows to free up a little extra time, eh? Anyway, picking up where we left off let’s finish up a few more items on the cluster resources and move on to the SQL installation. If you need to review the first part of this tutorial, click here.
Preparing the Cluster for SQL
Microsoft recommends clustering the MSDTC which is used by SQL for transactions and distributed queries. From what I read on the topic you can still install SQL on the cluster without clustering the MSDTC, but it may hamper performance of the application. I’ve always clustered the MSDTC resource if I was clustering SQL.
We will need to create a couple of additional resources that will be dependencies for the MSDTC resource first. One of the resources will be the network name of the cluster. This is the name that you would use when you wanted to access a share, or RDP, etc to the cluster and access the active node without having to verify which node was currently active. For this example, I will use SERVER as the network name for the cluster. This name will allow me to simply RDP to SERVER and not care whether or not SERVER01 or SERVER02 was currently active; it will automatically connect me to whichever node is active.
Note – do not confuse the Network Name resource with the cluster management name (SCCMCLUSTER in this example). That name is simply for connecting to the cluster mgmt console.
- Open Cluster Administrator; expand the resource group (not the Cluster Group). Select New -> Resource.
- In the New Resource screen, name the resource IP Address.
- Change the Resource type to IP Address.
- Leave the Group set to default. Do not use the Cluster Group for any of these resources.
- In the Possible Owners screen, verify both nodes are listed under Possible Owners and click Next.
- Click Next at the Dependencies screen without making any changes. The IP Address resource does not require any dependencies.
- Enter a valid IP address in the Address window as well as a valid subnet mask. Select Public for the Network dropdown and leave Enable NetBIOS for this address checked.
- Click Finish.
- When the wizard completes, right click on the IP Address resource and select Bring Online.
- Right click on the resource group again (not the Cluster Group) and select New -> Resource.
- Since this will be the network name of our cluster, let’s simply make the Name field SERVER.
- For the Resource Type, select Network Name in the dropdown.
- Leave the group as default and click Next.
- At the Possible Owners screen, verify that both server nodes are included under Possible Owners. Click Next.
- Since a host name needs an IP address to resolve against, select the IP Address resource from Available resources and move it over to the Resource dependencies column. Click Next.
- At the Network Name Parameters screen, enter in the name of the cluster (in this example, it’s SERVER). Enable the checkboxes next to DNS Registration Must Succeed and Enable Kerberos Authentication. Click Finish.
- Right click the new resource in the Cluster Administrator and select Bring Online.
- Note – This may take several seconds as the resource will create a DNS entry in Active Directory as well create the SPN for the cluster service.
- If the resource fails to come online, verify that a DNS record was created for the cluster name with the appropriate IP address from the dependency resource. Also verify that the server has the following SPN’s registered in Active Directory:
- If you need to manually create the SPN’s, you can use adsiedit or create them using the setspn utility with the following syntax:
Setspn –A MSClusterVirtualServer/SERVER SERVER
Setspn –A MSClusterVirtualServer/SERVER.DOMAIN.COM SERVER
(Replace SERVER with the appropriate server hostname and FQDN and yes, SERVER is listed twice in the above examples because you must specify the computer object target to modify the ServicePrincipalName attributes in Active Directory.)
MSClusterVirtualServer/SERVER
MSClusterVirtualServer/SERVER.DOMAIN.COM (Replace SERVER with the appropriate server hostname and FQDN)
- Right click on the resource group name one more time and select New -> Resource.
- In the Name field, enter MSDTC.
- Select Distributed Transaction Coordinator as the Resource Type.
- Leave the Group field unchanged as before (remember, NOT the Cluster Group). Click Next.
- Verify both nodes are included in the Possible owners column. Click Next.
- The MSDTC resource will require two dependencies; the Physical Disk (shared disk) and the Network Name resources as shown here.
- Add the two resources to the Resource dependencies column and click Finish.
- Right click on the MSDTC resource and select Bring Online.
- Take a breather; we are now ready to start installing SQL Server 2005.
Installing SQL Server 2005 on the Cluster
- Verify both nodes are online and SERVER01 is the active node. Run the install from SERVER01.
- The install must be run using a domain account with local administrator access on both nodes of the cluster.
- Accept the EULA and click Next.
- Install the required prerequisites by selecting the Install button.
- Select Next to begin the installation wizard.
- Address any issues that occur in the System Configuration Check and click Next.
- Enter the Registration Information and click Next.
- At the Component to Install screen, select the following and click the Advanced button.
a. SQL Server Database Services
b. Create a SQL Server failover cluster
c. Workstation components, Books Online, and development tools. (only required if you want a local copy of books online and dev related items; not required for SCCM).
- Click Next after making any adjustments to the install.
- At the Instance Name screen, select the Default instance (unless you need to specify a Named instance on the server) and click Next.
- At the Virtual Server Name screen, enter the name of the SQL virtual server. This is the server name that will be used to actually connect to the SQL instance, not the server OS itself. You will specify this name anytime you need to connect to this server for SQL services such as to install a database, or open the instance using SQL Management Studio, etc. For this example, we will use SERVERSQL. Click Next.
- In the Virtual Server Configuration screen, enter an IP address that will be used for the Virtual Server Name entered in the previous step. Be sure to select Public in the dropdown for the Network to use field. Click the Add button and then Next.
- Select the resource group (not the Cluster Group) in the Cluster Group Selection screen. Also, verify that the correct shared disk (not the Quorum disk) is listed in the Data files field. It will read something like E:\Microsoft SQL Server\. Click Next.
- here.
- Select Windows Authentication Mode at the Authentication Mode screen. Click Next.
- Accept the default settings at the Collation Settings screen and click Next.
- Select any checkboxes you wish on the Error and Usage Report Settings screen and click Next.
- Click the Install button to begin the installation process.
- Obviously when the process has completed, click the Finish button to complete.
- Launch the SQL Server Management Studio and verify you can connect to the new instance. Remember, you will need to connect to the network name of the SQL Server instance, not the server itself. In this example, we specify SERVERSQL in the connection window instead of SERVER.
- Update SQL to Service Pack 3
The next post in this walkthrough tutorial will cover installing System Center Configuration Manager 2007 on the clustered environment covered in these two tutorials as well as some of the pitfalls you may run into during the install process.
You can download the PDF version of the entire tutorial here.




