SQL 2012 Failover Cluster Build (Step by Step) – Part 1: Installing SQL 2012


Recently, my engagements were related to SQL and 2 of which were to build SQL failover clusters on based on Windows 2008R2/SQL 2008R2 SP1 and Windows 2012/SQL 2012 SP1. In this post, I will share with you my SQL Server 2012 Failover clustering experience.

Part 2: Adding a New Node to Cluster

Part 3: Adding a New Instance to Cluster


As displayed in the figure below,

  • The demo cluster has 2-nodes and uses File Server (FS) (Windows 2012) as storage and
  • Each node has at least 2 NIC - one for internal (LAN) and for private (heartbeat) network
  • FS provide 4 disks that are
    • Q: used for Quorum or witness disk (1GB)
    • E: used for Microsoft Distributed Transaction Coordinator (MSDTC ) (2GB)
    • I: used for SharePoint 2013 Content databases (30GB), yes, it is low, but I can afford only this muchSmile in my local.
    • H: used for Custom .NET application's databases (30GB)


Figure1: Windows Server 2012 failover cluster used in the demo

Please note that this is for demo purposes and built in my local environment (Windows 2012, Hyper 2012). My assumptions are:

  • SQL database instances will need database engine feature only and no BLOB type of record expected.
  • The pre-requisites for Windows and SQL failover clustering (storages, network configurations, user accounts, etc.) are already conducted.
  • You logged in Node 1 with cluster admin account.


How-to Build SQL 2012 Failover Cluster Steps

Step #

Screen capture

1- Start your SQL 2012 Server installation media and Click 'Installation'


2. Click 'New SQL Server failover cluster installation'


3. OK


4. Provide product key for your media. Then Next

sql 2012 PK

5. Accept License Agreement (and the other check if you would like). Then Next


6. Click Next (No worry, my VM not connected to Internet. Or you could uncheck 'include SQL Server prod… checkbox )


7. Setup Support Rules

Wait for Rules check. Once done make sure no failed task. Evaluate warnings for your environment (remember the VM is not connected to Internet). Then click Next


8. Next


9. Select Features then Next


10. Make sure feature Rule check complete with no Failed status. Then Next


11. Provide Network and Instance names. The SQL cluster name is CSQLP and my database instance name is WSSDB. MS recommends use of named instances. Then Next


12. Next


13. Next


14. Select Disk needed for the resource group (RG). Then Next


15. Enable IP4 and provide IP address for this RG. Then Next


16. MS recommends to use unique service accounts per each data services. Provide Service Accounts (and Collation if different than default - SQL_Latin1_General_CP1_Cl_AS). Then Next


17. Click on 'Add Current User' button. Then click on Data Directories tab menu item. I use 'I' labeled clustered disk for WSSDB content. Please consider having different disks for data, log and temp data. Then Next



18. Check the sharing error reports with MS if you would like. Then Next


19. Make sure no clusters rules failed. Then Next


20. Confirm the selection (last chance to edit before installation). Then Install


21. Enjoy your waiting with a coffee. Success!

Click Close to complete the process.

You should see it is reflected on Cluster roles.





In this post, as part of post series of building SQL 2012 Server failover cluster on top of Windows 2012 Server, I have demonstrated installing SQL 2012 Server failover cluster on node 1. In my next posts, I will demonstrate how to add a second node and how to install a new named database instance to the SQL cluster. Hope you like it and stay tuned!

Comments (30)

  1. Anonymous says:

    Thanks for blog, lots of valuable information step by step.

  2. Mecit Atmaca says:

    @Sawan, this is beyond of this post. That depends on variety things(edition, size, DR plan, etc.); yet I can say these: check/align with version upgrade requirements (http://technet.microsoft.com/en-us/library/ms143393.aspx), create your SQL failover cluster
    and then move your existing databases. You may consider SQL 2012 HA AG feature as well (see here for more details – http://technet.microsoft.com/en-us/library/ms190202.aspx).

  3. Mecit Atmaca says:

    @rasada, assuming as standalone? if so, yes. Yet, you should have a compelling reason to go with standalone vs leveraging the cluster.

  4. Mecit Atmaca says:

    @Vasant, glad you find it helpful!

  5. Mecit Atmaca says:

    @Bala, Yes, you can. I don’t recall any difference on steps.

  6. Vasant Sakpal says:

    fare enough demonstration, very helpful, I will get to know the exact screen and options going to ask during installation process

  7. Rod 955i says:

    sorry to many screen shots not enough discussion:

    the ip addresse for the sql cluster or instance – is this the cluster ip address already assigned or a new sql cluster address

    the sql shared disk cluster resources – looks like these have to be added after the mscs creation – i.e you cannot use a disk already added to the mscs cluster as a resoure?

  8. Ponka says:

    Can I use same disk for both MSDTC and SharePoint content database? Or they must be on separate disks because SQL Server won't allow it?

  9. johnredd says:

    where can i find the link to install another node in a cluster

  10. Mecit Atmaca says:

    @Jr, please refer to Part 2, listed at top

  11. Mecit Atmaca says:

    @Ponka, technically speaking you can use same physical disk (but different logical one -LUN) for MSDTC and content databases, but not recommended at all for production environment. Use of MSDTC is required for specific case (required only when transactional data occurs cross machines).

  12. Mecit Atmaca says:

    @Rod 955i, sorry that I am responding to your query late and followed LIFO:). Yes, the cluster and instances need different IP addresses. In fact, you need separate IPs per cluster, node (2 IPs – one internal and one external), instance (database instance), MSDTC (if needed). Most of the clustering efforts takes place before building phase, you should have specify/prepare LUNs, IPs, user and service accounts, FWs stuff ahead of build time. Hope that answers you question.

  13. michal says:

    Does SQL Server Cluster support "Standard" version of Windows Server host OS ?

  14. Mecit Atmaca says:

    @Michal, Specific to Standard edition, Windows 2008 R2 or prior versions do not support clustering, but later versions (2012 and 2012R2). Please note this is stated for std. edition.

  15. Raadee says:

    Michal – Yes, Windows Server 2012.

    And there is no requirement for a secondary NIC, also called heartbeat NIC.

    That was the good old days.

  16. Mecit Atmaca says:

    @Raadee, correct. Thanks for addition.

  17. Sawan says:

    Thanks for the blog, its helpful.. I need a help.
    We have a Standalone SQL 2005 server. Now our current pla is to migrate and upgrade.
    We have 2 new physical boxes ( SQL Cluster) with Windows Server 2008 R2. So our plan is to restore the database from Standalone to this new box and upgrade it to SQL 2012and SQL cluster needs to be setup.
    Please advise what would be the right path

  18. bala says:

    Hi there,

    Quick question,

    Can I build the same SQL 2 node cluster on Windows Server 2008 R2 enterprise?. Just wondering, if there are any change in steps, if the OS is Windows server 2008 R2 and the SQL is SQL2012 standard edition.


  19. rasada says:

    can i install SQL 2012 Instance on windows 2008R2 OS where i already have multiple SQL 2008 Cluster Instances installed?

  20. Anonymous says:

    Pingback from Failover Cluster Hk.

  21. Sunil says:

    hi ,
    the post is very helpful but a small doubt,in my setup with win 2008 enterprise and Sql 2012 std, i want to setup sql failover,so during windows failover cluster installation do in need to configure and services in HA eg MDtc or any other pls guide.


  22. Kesava says:

    Is there an article on how to create the following:

    Q: used for Quorum or witness disk (1GB)
    ◦E: used for Microsoft Distributed Transaction Coordinator (MSDTC ) (2GB)
    ◦I: used for SharePoint 2013 Content databases (30GB), yes, it is low, but I can afford only this muchSmile in my local.
    ◦H: used for Custom .NET application’s databases (30GB)

    or are they just additional hard drives you can add to the VM using Hyper-V Settings?


  23. George says:

    Where is the DB located ? In this example is it on the file Server or locally on the SQL Server ?

  24. Mario says:

    in any of the steps mentioned above it is required internet access?
    ie. in the health check?

    It is possible to install the cluster with no internet access at all?

  25. Pingala says:

    Does SQL server 2012 Alwayson require Windows Failover Cluster?

  26. subbarao.maguluri says:

    @This is very use full blog to every one , thanks very much

  27. SonomaRik says:

    like so many have addressed, I do not see the value in always requesting a cluster name for each new, named instance UNLESS the desire is to manage each named instance as a single fail-over clustered event from other named instances on the same node: comment?
    otherwise it would require a LOT of overhead [disks, ips etc.] to just install each named instance on a multi-node ‘always-on’ cluster environment.

  28. aiwa says:

    what specific IP should i put in ipv4? is it the ip itself? or the ip of the other vm?

  29. Hellosql says:

    Why arent there any answers to question posted here? I need to know if the VIP Address is the same as the cluster VIP

Skip to main content