Install a clustered SQL instance Part 1

Part 1 Create a Windows Cluster

I’m about to start a new series of articles about the installation of a Failover Cluster SQL Instance. This is an environment I will be using primarily for testing in my lab hence with no intent to follow the strictest best practices when it comes down to hardening the infrastructure providing SQL Services. This wants to be a quick tutorial on how to simply proceed with the deployment. I would always recommend though to take a look at the official SQL documentation to adapt this steps to the personal requirements.

 For an easier reading (and to cut down on the number of screenshots!) I have organised this guide in 3 parts:

Part 1 – Create a Windows Cluster

Part 2 – Install the first SQL node of the clustered SQL instance

Part 3 – Add a node to the SQL clustered instance and test failover

So here we go with the first part about creating a SQL Cluster.

For my lab environment I decided to use a 2 nodes cluster based on Windows Server 2012 R2. These two nodes will also be the same serving the failover clustered SQL instance. Ideally I would like to have separate volumes hosting separate SQL components like for example:

It’s quite a big number of drives and for obvious reasons with the exclusion of P drive all the remaining ones are iSCSI drives connected to both nodes participating to the cluster. For more information on how to create and mount this drives this article comes handy.

So having a clustered SQL installation in mind this is the shopping list:

  • Create the shared drives

    • B disk for root directory, system databases and backups

    • T disk for TempDB (data and log)

    • Q disk for Quorum

    • S disk for User Databases data

    • L disk for User Databases logs

  • Create a DNS Host (A) Record for SQL Cluster name

    • My lab will use SQLCLU-2014 with IP 10.254.6.252

  • Create a Computer Object in Active Directory with SQL Cluster name

    • My lab will use SQLCLU-2014

  • Create a DNS Host (A) Record for Windows Cluster

    • My lab will use WINCLU-2012 with IP 10.254.6.253

  • Create a Computer Object in Active Directory with Windows Cluster name

    • Actually this will be automatically provisioned by the WSFC during the installation wizard

Install the Failover Cluster feature

First of all from the Windows Server Manager let’s add the role of Failover Clustering as per screenshots below. The same role has to be installed on all nodes participating to the cluster. For this article the installation in my lab will be executed on SQL2014A and SQL2014B

At this point before proceeding it is recommended to Validate the Configuration by opening the Cluster Manager snap-in.  A quicker way is to run with the command CLUAdmin.msc. This will run an extensive list of tests to check if all components are fit for the purpose. It might take some time depending on the components and also on the level of Software Drivers and Updates installed. It is recommended to start with a full scan with all tests the very first time to identify potential issues. Afterwards it is possible to select specific ones where issues (warning or errors) are found

Create the Windows Cluster

Now that all components have been scanned and cleared by clicking on Create Cluster link from the right side panel will start the wizard. Let’s specify the intended name for the Windows Cluster as this will be created as a user object in Active Directory (so make sure same name does not exist yet!). Also as per previous requirements a DNS Host (A) Record must exist in both forward and reverse zone. Even though it is not a requirement I suggest using a static IP address rather than DHCP

It’s now time to grant Full permissions to the Windows cluster Computer object to the SQL Cluster Computer object in Active Directory . To do this from the Advanced View  let’s select the SQLCLU-2014 properties and add Full permissions to the WINCLU-2012 as per screenshots below

 

This concludes the first part of the installation of a failover clustered SQL instance. Next step is to install the first SQL node of the cluster

Michele Domanico

Passionate about Virtualization, Storage, Data Availability and Software Defined Data Center technologies. The aim of Domalab.com is sharing with the Community the knowledge and experience gained with customers, industry leaders and like minded peers. Always open to constructive feedback and new challenges.

Leave a Reply