Learn more about SQL Server tools

 

Tutorials          DBA          Dev          BI          Career          Categories          Videos          Whitepapers          Today's Tip          Join

Tutorials      DBA      Dev      BI      Categories      Events

DBA    Dev    BI    Categories

 

Install SQL Server 2008 on a Windows Server 2008 Cluster Part 2


By:   |   Read Comments (18)   |   Related Tips: 1 | 2 | 3 | 4 | More > Clustering

Problem

In a previous tip on SQL Server 2008 Installation Process, we have seen how different SQL Server 2008 installation is from its previous versions. Now, we have another challenge to face: installing SQL Server 2008 on a Windows Server 2008 Cluster. Windows Server 2008 has a lot of differences from its previous versions and one of them is the clustering feature. How do I go about building a clustered SQL Server 2008 running on Windows Server 2008?

Solution

To continue this series on Installing SQL Server 2008 on a Windows Server 2008 Cluster, we will look at building our Windows Server 2008 cluster in preparation for SQL Server 2008. In Part 1, we have completed the installation of the Application Server role in both of the servers that we will be using as part of our cluster. This tip will walk you through the installation of the Failover Cluster Feature, validating the servers that will be a part of the cluster, and creating the cluster.

Adding the Failover Cluster Feature

Windows Server 2008 calls them features which are simply software programs that can support or augment the functionality of server roles. Since we've already installed the Application Server role in our server, let's define a feature from this perspective: failover clustering simply augments the role as an application server by making it highly available. It is disabled by default, unlike in Windows Server 2003 so we need to add it on both of the servers that will form a part of our cluster.

To add the Failover Clustering feature:

  1. Open the Server Manager console and select Features.
  2. Click the Add Features link. This will run the Add Features Wizard

  3. In the Select Features dialog box, select the Failover Clustering checkbox and click Next.

  4. In the Confirm Installation Selections dialog box, click Install to confirm the selection and proceed to do the installation of the Failover Clustering feature.

  5. In the Installation Results dialog box, click Close. This completes the installation of the Failover Clustering feature on the first node.

That's how simple and easy it is to add the Failover Clustering feature in Windows Server 2008. You will have to do this on both nodes to complete the process. Once you have managed to install the Failover Cluster Feature on both nodes, we can proceed to validate our servers if they are ready for clustering.

Running the Windows Server 2008 Validate Cluster Configuration

Unlike in previous versions of Windows where Microsoft had some sort of a hardware compatibility list (HCL) from which we had to find and select components tested to be clustering-supported, this wizard is like the "seal" that tells you whether or not the hardware you are using is supported. In fact, Microsoft has partnered with hardware vendors to create the Failover Cluster Configuration Program to make the acquisition of hardware for Windows Server 2008 Failover Clustering very easy and simple. Basically, your hardware will be supported for clustering if it meets these two requirements: the server has a "Certified for Windows Server 2008" logo and it passes this wizard.

One word of caution: do not skip any error message that this wizard generates in the final report. Doing so would simply mean that your configuration going forward will be unsupported. You only need to run this wizard on either of the nodes.

To run the Validate Cluster Configuration Wizard:

  1. Open the Failover Cluster Management console
  2. Under the Management section, click the Validate a Configuration link. This will run the Validate a Configuration Wizard

  3. In the Select Servers or a Cluster dialog box, enter the hostnames of the nodes that you want to add as members of your cluster and click Next.

  4. In the Testing Options dialog box, click Next to run all the necessary tests to validate whether or not the nodes are OK for clustering. If this is the first time to run the wizard, you must run all the tests for validation. For succeeding runs, especially when adding hardware like disk subsystems to network cards on your cluster nodes, you can selectively choose which tests to run as long as you have initially validated your hardware by running all tests.

  5. In the Confirmation dialog box, click Next. This will run all the necessary validation tests.

  6. In the Summary dialog box, verify that all the report returns successful.

If you have reached this part of the process, the wizard will tell you whether or not you can proceed to the next step of creating your cluster. As I've mentioned earlier, do not attempt to go any further if this report returned any error messages. I have seen some installations where the shared disk is displaying an error in the validation report prompting me to reconfigure the shared disk. This could mean removing the disk subsystem from both nodes, creating new disks and presenting them on the nodes as mentioned in Part 1 of this series. It would be best to work with your storage engineers or your system administrators when in doubt as different vendors may have different implementations of their disk subsystems.

I've also seen issues pertaining to IPv6. This is a fairly common issue which can easily be resolved. The error message in the cluster validation report looks something similar to the one displayed below

Verifying that there are no duplicate IP addresses between any pair
of nodes. Found duplicate IP address fe80::100:7f:fffe%13 on node 
node1.domain.local adapter Local Area Connection* X and node 
node2.domain.local adapter Local Area Connection* X. 

This blog post outlines the step in resolving this issue. In a few cases, however, I needed to disable the Teredo Tunneling Pseudo-Interface adapter from Device Manager before I got a successful summary report generated by the Validate Cluster Configuration wizard. The bottom line is simply to make sure that the report returns a successful validation before creating the cluster.

Creating the Windows Server 2008 Cluster

You've finally reached this step in the entire process. This means you are now ready to create your Windows Server 2008 cluster. It's as easy as running the Create Cluster Wizard on either of the nodes. Make sure that you have your virtual hostname and IP address ready before proceeding

To run the Create a Cluster Wizard:

  1. Open the Failover Cluster Management console
  2. Under the Management section, click the Create a Cluster link. This will run the Create Cluster Wizard

  3. In the Select Servers dialog box, enter the hostnames of the nodes that you want to add as members of your cluster and click Next.

  4. In the Access Point for Administering the Cluster dialog box, enter the virtual hostname and IP address that you will use to administer the cluster. Click Next

  5. In the Confirmation dialog box, click Next. This will configure Failover Clustering on both nodes of the cluster, add DNS and Active Directory entries for the cluster hostname.

  6. In the Summary dialog box, verify that all the report returns successful.

Congratulations! You now have a working Windows Server 2008 cluster. Notice how easy it was to do all of these with fewer steps and less configuration compared to its predecessors. You can now validate whether your cluster is working or not. A simple test would be to do a continuous PING on the virtual hostname or IP address that you have assigned to your cluster. Reboot one of the nodes and see how your PING test responds. At this point, you are now ready to install SQL Server 2008.

OPTIONAL: Configuring your cluster quorum

This section is sometimes necessary especially when Windows Server 2008 decides to take a different disk subsystem as a quorum other than the one you've originally intended it to. Notice that in the Create a Cluster wizard, there was no option to select the disk subsystem that we can use as a quorum disk (now called the "witness" disk). By default, the Create a Cluster wizard will use the first available disk as the witness disk. I have seen cases where the originally planned witness disk is sized 1GB while the other shared disks are sized 200 GB. The wizard then selects one of the 200GB-sized disks as a witness disk, requiring you to move the witness disk to the original allocation. To validate, check the Storage node under the Failover Cluster Management console

To configure the quorum in a failover cluster:

  1. Open the Failover Cluster Management console
  2. Select the name of the cluster you have just created. Right-click on the cluster, select More Actions, and click Configure Cluster Quorum Settings... This will open up the Configure Cluster Quorum Wizard

  3. In the Select Quorum Configuration dialog box, select the Node and Disk Majority (recommended for your current number of nodes) option. The options presented will depend on how you want your cluster configured. The current selection is for a 2-node cluster configuration

  4. In the Configure Storage Witness dialog box, validate and select the disk that you want your cluster to use as the quorum/witness disk. Click Next

  5. In the Confirmation dialog box, verify that the disk configuration for the quorum/witness disk is correct. Click Next

  6. In the Summary dialog box, verify that all the configurations are successful.

For more details on Configuring the Quorum in a Failover Cluster in Windows Server 2008, check out this Microsoft TechNet article.

Next Steps
  • Download and install an Evaluation copy of Windows Server 2008 for this tip
  • Review Part 1, Part3 and Part4 of this series on MSSQLTips
  • Start working on building your test environment in preparation for building a SQL Server 2008 cluster on Windows Server 2008


Last Update:






About the author
MSSQLTips author Edwin Sarmiento Edwin M Sarmiento is a Microsoft SQL Server MVP and Microsoft Certified Master from Ottawa, Canada specializing in high availability, disaster recovery and system infrastructures.

View all my tips


 









Post a comment or let the author know this tip helped.

All comments are reviewed, so stay on subject or we may delete your comment. Note: your email address is not published. Required fields are marked with an asterisk (*).

*Name    *Email    Notify for updates 


Get free SQL tips:

*Enter Code refresh code     



Wednesday, September 19, 2012 - 9:16:31 AM - bass_player Back To Top

After SQL Server has been installed on the very first node, it becomes the Active node. The SQL Server instance will live on this node. While installing the other nodes, it will ask for the SQL Server instance name. In a failover clustered instance, it is always the active node that runs the instance so by selecting the instance name, you have implicitly specified the active node.


Wednesday, September 19, 2012 - 2:35:35 AM - kamal Back To Top

 

As you mentioned in earlier comments that if ""You do not need to specify the passive node since every node you add after installing the SQL Server instance becomes the passive node, again, only if you use the Add Node option during the installation.""

 

But how to specify the active node after adding the new Nodes.. i.e. Active/Active or Active/Active/Passive

 


Monday, July 09, 2012 - 7:09:49 PM - bass_player Back To Top

Dave,

After installing SQL Server on Node1, you only need to install the SQL Server binaries on Node2 using the Add Node option in the installation media. This will allow the SQL Server Failover Cluster instance to failover between the two servers. You do not need to specify the passive node since every node you add after installing the SQL Server instance becomes the passive node, again, only if you use the Add Node option during the installation


Monday, July 09, 2012 - 12:15:29 PM - Dave Back To Top

 Hi, 

Sorry, a very simple question 

I looked at the steps and reading different articles but I'm not able to figure out this 

 

During the Install: Do I need to tell somewhere about the Passive Node ?

First we are installing SQL on Node1 and all services are installed on Node1 (This is called Active node) 

Now what do I need to do we Node2 

Do I install SQL on Node2, it will be A/A clustering ? or what are the steps for Passive ?

Please help understand, thanks 

Dave


Saturday, July 07, 2012 - 12:04:34 PM - bass_player Back To Top

dba_123,

I don't quite understand what you mean by having a SQL Server instance active on both nodes. A SQL Server Failover Cluster Instance cannot be active on more than one node. That is by design. A cluster resource, in this case the SQL Server instance, can only be owned by a node at a time.


Monday, June 11, 2012 - 3:46:26 PM - dba_123 Back To Top

Hi Edwin!

This is a great article!!!

How would I go about setting up a 2-node cluster where an instance of sql is active on both nodes and fail to the other node?

 

Thanks!


Thursday, May 24, 2012 - 7:23:43 PM - bass_player Back To Top

As I've previously mentioned, make sure that your shared storage supports SCSI-3 persistent reservation. The only ones who can help you here are your storage engineers


Thursday, May 24, 2012 - 12:15:42 AM - VAHID Back To Top

Althoug my system hard are  SCSI (HP SAN STORAGE)  and we work with VMW vspher 5.0, my problem not solve?


Wednesday, May 23, 2012 - 1:55:40 AM - VAHID Back To Top

Thank's Edwin

please sned me email.

 

 


Tuesday, May 22, 2012 - 8:58:31 AM - bass_player Back To Top

This is a storage-related issue. As the message implies, your shared storage must support SCSI-3 persistent reservation or it won't pass the validation. Talk to your storage engineers to assist you further on this issue


Tuesday, May 22, 2012 - 4:24:29 AM - VAHID Back To Top

please help to me

 

Validate SCSI device Vital Product Data (VPD)

Validate that storage supports necessary inquiry data (SCSI page 83h VPD descriptors) and that they are unique.

The test was canceled.


Back to Summary
Back to Top


 

Validate SCSI-3 Persistent Reservation

Validate that storage supports the SCSI-3 Persistent Reservation commands.


Tuesday, May 22, 2012 - 4:21:49 AM - VAHID Back To Top

Hi

I have this problem in validating.

 

List disks visible to two or more nodes that will be validated for cluster compatibility. Online clustered disks will be excluded.

Checking that multi-path I/O (MPIO) works for disk with identifier 701cda52 on node psql.bzhos.com

Checking that multi-path I/O (MPIO) works for disk with identifier 701cda52 on node sqlnode1

Writing to sector 11 on disk with identifier 701cda52 from node psql.bzhos.com

Reading sector 11 on disk with identifier 701cda52 from node psql.bzhos.com

Reading sector 11 on disk with identifier 701cda52 from node sqlnode1

There are multiple disks with identifier 701cda52

There are multiple disks with identifier 701cda52


Thanks


Thursday, November 11, 2010 - 6:45:13 AM - JQ Back To Top

I have a question that relates to cluster node failure. In the event that a node fails, and is un-recoverable, how would you forcefully evict the node from the cluster with regard to removing all of the SQL Server configuration etc. This is for a situation where a complete server rebuild is required and there is no system state backup present.

I ask, primarily becase we are now asking these questions where I work and I am struggling to find information on how to do this type of eviction.

 

Cheers

 

JQ


Friday, March 19, 2010 - 10:44:37 AM - bass_player Back To Top

Check the GUID value. This could be another SQL Server component or something else. I'm thinking it has something to do with Active Directory and DNS

What's your infrastructure like? OS (2008 RTM or R2), SQL Server version and edition (2008 RTM or R2), Service Pack levels, virtualization platform?


Thursday, March 18, 2010 - 6:18:10 PM - Ronak Back To Top

Hi My Application Server installed successfully by wizard but when I go to Server Manager I see this error:  

The server {GUID} did not register with DCOM within the required timeout.EVENT ID 10010

Also when I try to register the server for validation  It gives me this error.

The computer servername.domain name is unreachable. Please ensure that this server is running. Also, ensure that this server's firewall,if enabled, allows Remote procedure call request.

My both servers are up and running. I am able to ping each other.. Please help me out.

Thanks in advance

 


Wednesday, July 15, 2009 - 1:29:39 AM - GDST_Admin Back To Top

Thanks for your info. 

I just want to clear up whether this is the case with Active / Passive Node Or does it have to be Active / Active Node.

I am going to be running just Active Passive.  From what you are saying.  If the disk quorum failed as long as both the Active and Passive nodes are still running.  Cluster will run but no failover ability.  Therefore, time to fix the quorum disk.  Once thats fixed, failover ability will return.


Tuesday, July 14, 2009 - 2:14:38 PM - bass_player Back To Top

In this quorum model, a cluster remains active until half of the nodes and its witness disk is available. In case the witness/quorum disk failed, the cluster requires a majority of nodes to be up and running in order to successfully run the cluster or in this case, the SQL Server service. Simply put, for a 2-node cluster, both nodes have to be up and running when the witness/cluster disk fails to make sure that the SQL Server service continues to run. Here's a challenge, run this test on a virtual environment having the cluster disks on a shared storage, probably an iSCSI disk. To simulate a failed qitness/quorum disk, disable the network card that connects to your shared storage. While you're at it, connect to SQL Server and see if you can connect properly. That will provethe point. 


Tuesday, July 14, 2009 - 7:51:51 AM - GDST_Admin Back To Top

If I was to create a Node and Disk Majority for an Active Passive Cluster on SQL 2008 (Just 2 nodes)

Is it true that, if the disk quorum failed.  SQL Server 2008 will continue running, but no failover ability until disk quorum is fixed.  Or will SQL Server 2008 stop running altogether until quorum is fixed?

  

 


Learn more about SQL Server tools