SQLServer 2008 R2 Failover Cluster on vSphere 5.0

2 Node Cluster on a Single Host


Part One – CREATING THE VIRTUAL MACHINE NODES
The purpose of this blog is to provide some insight into creating a SQL Server R2 cluster on Windows Server 2008 R2 in a virtual environment. It is a complex process but one that is well documented on the internet.

The basic steps are:
  1. Create the vm nodes
  2. Configure the Failover Cluster feature
  3. Install the Sql cluster
I found documents that addresses each step but I thought I would consolidate the information for my specific project. Below are some tips/tricks/insights that I felt were important.
I am testing with a 3 node (3 physical servers) vSphere 5.0 cluster with datastore’s that are Thin Provision (NFS) only and datastore’s (VMFS) that can create Thick Provision Eager Zeroed disks.

SERVER OS EDITION MUST BE ENTERPRISE!!

Only Enterprise or DataCenter editions have the Failover Clustering feature. Standard does not

CREATE VM’S THAT WILL BE CLUSTERED WITH THICK PROVISION EAGER ZEREOD DISKS!!

Thick Provisioned Eager Zeroed format (VMFS) means that all the space allocated to the disk size will be written and zeroed out at the time the vm is created or a new disk is added. This helps with performance in a clustered environment.

SHARED SCSI CONTROLLER TYPE MUST BE SET TO LSI LOGIC SAS!!

Windows Server 2008 R2

SHARED SCSI BUS SHARING MUST BE SET TO VIRTUAL!!

Disks can be shared between virtual machines on the same server.


Create the First Node for Clusters on One Physical Host

To create the first node, you create and configure a virtual machine with two virtual network adapters and install a guest operating system on the virtual machine. You configure the virtual network adapters to handle virtual machine traffic for the cluster: a private network connection for the private heartbeat and a public network connection.

Procedure

1.       Create VM

2.       Install OS

3.       Install VMWare Tools

4.       Configure NIC’s

VM_Cluster_#1

-SCSI Controller 0 – SCSI Controller Type = LSI Logic SAS, SCSI Bus Sharing = None(0:0)

- Hard Drive #1 – Thick Provision Eager Zeroed

- NIC #1 (your static network ip address) PUBLIC

- NIC #2 (static ip address on a separate subnet) PRIVATE



Create the Second Node for Clusters on One Physical Host

You can either create the second node from scratch using the first node instructions or you can clone Node #1 to template then deploy that template to a vm. Instructions are in the guide.

Add Hard Disks to the First Node for Clusters on One Physical Host

In a Failover cluster, storage disks are shared between nodes. You set up a quorum disk and an optional shared storage disk.

Procedure

1      In the vSphere Client inventory, select the newly created virtual machine and select Edit Settings.

The Virtual Machine Properties dialog box appears.

2      Click Add, select Hard Disk, and click Next.

3      Select Create a new virtual disk and click Next.

4      Select the disk size.

5      Under Disk Provisioning, select Thick Provision.

6      Click Next.

7      From the Virtual Device Node drop-down menu, select a new SCSI controller (for example, SCSI (1:0)).

NOTE  You must select a new virtual device node. You cannot use SCSI (0:0).

8      Click Next, and click Finish.

The wizard creates a new hard disk and a new SCSI controller.

9      In the Virtual Machine Properties dialog box, select the new SCSI controller and click Change Type.

The Change SCSI Controller Type dialog box appears.

10    Select the appropriate type of controller, depending on your operating system.

Operating System Type of Controller - Windows Server 2008 LSI Logic SAS (This should already be set if you are using Server 2008)

11    Click OK.

12    On the Virtual Machine Properties dialog box, set SCSI Bus Sharing to Virtual and click OK!

VM_Cluster_#1 with 2nd disk added

-SCSI Controller 0 – SCSI Controller Type = LSI Logic SAS, SCSI Bus Sharing = None (0:0)

- Hard Disk #1 – Thick Provision Eager Zeroed

- NIC #1 (your static network ip address) PUBLIC

- NIC #2 (static ip address on a separate subnet) PRIVATE

-SCSI Controller 1 – SCSI Controller Type = LSI Logic SAS, SCSI Bus Sharing = Virtual (1:0)

- Hard Disk #2 (Quorum Disk) – Thick Provision Eager Zeroed

Add Hard Disks to the Second Node for Clusters on One Physical Host

To allow shared access to clustered services and data, point the quorum disk of the second node to the same location as the first node’s quorum disk. Point shared storage disks to the same location as the first node’s shared storage disks.

Prerequisites

Before you begin, obtain the following information:

-          Which virtual device node is for the first virtual machine's shared storage disks (for example, SCSI (1:0)).

-          The location of the quorum disk specified for the first node.

Procedure

1      In the vSphere Client inventory, select the second virtual machine that you created and select Edit Settings.

The Virtual Machine Properties dialog box appears.

2      Click Add, select Hard Disk, and click Next.

3      Select Use an existing virtual disk and click Next.

4      Select the same virtual device node you chose for the first virtual machine’s shared storage disks (for example, SCSI (1:0)), and click Next.

NOTE -The location of the virtual device node for this virtual machine’s shared storage must match the corresponding virtual device node for the first virtual machine.

5      In Disk File Path, browse to the location of the quorum disk specified for the first node.

VM_Cluster_#2 with 2nd disk added

-SCSI Controller 0 – SCSI Controller Type = LSI Logic SAS, SCSI Bus Sharing = None (0:0)

- Hard Disk #1 – Thick Provision Eager Zeroed

- NIC #1 (your static network ip address) PUBLIC

- NIC #2 (static ip address on a separate subnet) PRIVATE

-SCSI Controller 1 – SCSI Controller Type = LSI Logic SAS, SCSI Bus Sharing = Virtual (1:0)

- Hard Disk #2 (Quorum Disk) – Thick Provision Eager Zeroed

So now you should have 2 virtual machines sharing a drive. Each vm should be identical to the other besides the name and ip addresses. Next step will be adding the Failover Clustering feature. THis will be in Part 2.

Comments

Popular posts from this blog

Exchange 2016 - Error adding mailbox databases copies to new DAG server

Configure Message Delivery Restrictions on a resource mailbox in Exchange 2016