====== Install SQL Server 2016 Always-On Cluster ======

I a big believer in SQL Server's Always-On technology. With storage being cheap, and the fast fail-over of an Always-On server, there is no reason to use the older clustering technology that required shared storage and could take up to 30 seconds in my experience to start the SQL Database Engine and complete the fail-over. With Always-On, there is also the ability to do "geo" based replication and fail-over. Geo-Replication enables you to configure readable secondary databases in the same or different data center locations. Secondary databases are available for querying and for fail-over in the case of a data center outage or the inability to connect to the primary database.

Here is how I set-up SQL Server 2016 Always-On Clusters. For these examples I'll use the following assumptions:
  * **Active Directory Domain:** contoso.com 
  * **SQL Server Node 1:** SQLAO1N1
  * **SQL Server Node 2:** SQLAO1N2
  * **SQL Cluster Name:** SQLAO1
  * **Group Managed Service Account:** svc-sqlao1

The first step is to install [[windows:install:win2012r2]] on both nodes. I also add two additional disks: one for data and one for logs. I also format them with 64K allocation units which I think gives SQL better performance:<code powershell>
$disks = @(
    $(Get-Disk -Number 1),
    $(Get-Disk -Number 2)
)

foreach ($disk in $disks) {
    Set-Disk -Number $disk.Number -IsOffline $false
    Set-Disk -Number $disk.Number -IsReadOnly $false

    Initialize-Disk -Number $disk.Number -PartitionStyle GPT
    New-Partition -DiskNumber $disk.Number -UseMaximumSize -GptType '{ebd0a0a2-b9e5-4433-87c0-68b6b72699c7}' -AssignDriveLetter

    $partition = (Get-Partition -DiskNumber $disk.Number | Sort-Object Size -Descending | Select-Object -First 1)

    Format-Volume -FileSystem NTFS -Partition $partition -AllocationUnitSize 65536 -Confirm:$false
}
</code>

The next step is to create an Active Directory Group for DBAs and the [[windows:ad:group_managed_service_account]] for the SQL Engine to use. On the Domain Controller or a system with a remote server administrative tools (RSAT) installed:<code powershell>
Import-Module ActiveDirectory
 
New-ADGroup -Name SQLAO1DBA -GroupCategory Security -GroupScope Global

New-ADServiceAccount -Name svc-sqlao1 -Enable $true -DNSHostName sqlao1.contoso.com
 
New-ADGroup -Name SQLAO1Nodes -GroupCategory Security -GroupScope Global
Add-ADGroupMember SQLAO1Nodes SQLAO1N1$,SQLAO1N2$ 
 
Set-ADServiceAccount -Identity svc-sqlao1 -PrincipalsAllowedToRetrieveManagedPassword SQLAO1Nodes
</code>

Then, reboot each node and then execute the following on each node:<code powershell>
Add-WindowsFeature RSAT-AD-PowerShell
Import-Module ActiveDirectory
 
Install-ADServiceAccount -Identity svc-sqlao1
</code>
Now the account is available on each node.
  - Mount the SQL 2016 Installation ISO.
  - Since Always-On does not use the traditional clustering technology or clustering services, select the SQL server stand-alone installation option.\\ {{:windows:install:sql2016:step1.png}}
  - On the license key page make sure you have a enterprise edition license key. Click Next.\\ {{:windows:install:sql2016:step2.png}}
  - On the next page accept the license key. Click Next.\\ {{:windows:install:sql2016:step3.png}}
  - One feature I like about the SQL installer is the ability to download and apply updates as part of the installation process. Click Next.\\ {{:windows:install:sql2016:step4.png}}{{:windows:install:sql2016:step4a.png}}
  - As part of the installation, the installer will conduct a series of system checks and will not let you proceed until all tests have passed. Click Next.\\ {{:windows:install:sql2016:step5.png}}
  - At this point, select SQL Server features to be installed. This will be very dependent on the environment but I typically do a minimum install for my Always-On nodes and then will setup things like Analysis or Reporting Services on a separate server.\\ {{:windows:install:sql2016:step6.png}}
  - For the SQL Instance configuration, Accept the the default instance. Click Next.\\ {{:windows:install:sql2016:step7.png}}
  - Select the group managed service account created above. The installer will blank out the password field for the gMSA account. Click Next.\\ {{:windows:install:sql2016:step8.png}}
  - Next, choose the authentication option. I choose Windows, since I set most application up to use integrated authentication but have on occasion set up Mixed Mode authentication if an application couldn't work with integrated authentication, I also add the DBA group created above to manage the node. I never add individual accounts to the SQL Server Administrators. Click on the Data Directories tab.\\  {{:windows:install:sql2016:step9.png}}
  - I Set the Data root Directory to ''D:\'' and remove the extra ''Program Files\Microsoft SQL Server'' from the path. I also switch the Log Directory to the ''E:\'' drive.\\  {{:windows:install:sql2016:step9a.png}}
  - On the TempDB tab, I change the log directory to the ''E:\'' drive as well. Click Next.\\ {{:windows:install:sql2016:step9b.png}}
  - Finally, review all of the configuration settings. Click Install and wait for it to complete.\\ {{:windows:install:sql2016:step10.png}}

While SQL is being installed on the first node, repeat this on the other node to get it ready.

==== Fail-Over Clustering ====

Once both nodes are up and running and have SQL Server installed, it is time to install Fail-Over Clustering. Open an elevated PowerShell session on each node:<code powershell>
Install-WindowsFeature -Name Failover-Clustering -IncludeManagementTools 

Test-Cluster -Node SQLAO1N1,SQLAO1N2
</code>
You may get warnings regarding one network interface or no shared disk but these are ok as the Fail-over is only managing a shared IP address for the listener. On one of the nodes, execute the following in an elevated PowerShell session:<code powershell>
New-Cluster -Name SQLAO1 -Node SQLAO1N1,SQLAO1N2 -StaticAddress 172.16.101.81 -NoStorage
</code>

I usually set up Cluster-Aware Updating at this point so that I can apply updates easily, but I'll put the directions in another page since it is pretty much the same on all types of Windows clusters.

==== Availability Group Preparation ====
Always-On needs a place to store the database during the initial replication. I typically create a folder ''D:\Backups'' drive for backups and the initial replication. Share this folder and give the SQL gMSA account permission to Modify, Read & execute. Then:
  - On each node, open SQL Server Configuration Manager and locate the SQL Server service. Open the properties and click on the AlwaysOn High Availability. Check the ''Enable AlwaysOn Availability Groups'' and make sure that the Windows fail-over cluster name is correct.\\ {{:windows:install:sql2016:ao_enable.png?nolink|}} 
  - Restart the SQL Service on both nodes.

==== Availability Group Configuration ====
  - Open SQL Server Management Studio. In the Connect To Server window, ensure that SQLAO1N1 is listed, and use Windows Authentication to log in. Click Connect.
  - Create a Database named AOSeed. This database will be used to set-up the availability group.
  - Preform a full backup prior to proceeding.
  - In the left pane, expand Always-On High Availability.
  - Right-click Availability Groups, and select New Availability Group Wizard.
  - On the Introduction page, click Next.\\ {{:windows:install:sql2016:ao1.png}}
  - On the Specify Availability Group Name page, enter SQLAO1.\\ {{:windows:install:sql2016:ao2.png}}
  - On the Select Databases page, select the check box next to AOSeed. Note that this database is marked as meeting the prerequisites. Click Next.\\ {{:windows:install:sql2016:ao3.png}}
  - On the Specify Replicas page, click Add Replica. In the Connect To Server window, enter SQLAO1N2 next to Server Name, and click Connect.
  - For SQLAO1N1 and the newly added SQLAO1N2 server instance, select the box for Automatic Failover. This will also automatically select the box for synchronous replication for each SQL Server instance.\\ {{:windows:install:sql2016:ao4.png}}
  - On the Select Initial Data Synchronization page, select Full. The wizard subsequently triggers a full backup of the database, logs to a location you specify, and restores this backup to the secondary SQL Server instance, in this case, SQLAO1N2. From there, replication begins synchronously between SQLAO1N1 and SQLAO1N2.
  - In the Specify A Shared Network Location Accessible By All Replicas text box, enter \\SQLAO1N1\Backup, and click Next to begin validation.\\ {{:windows:install:sql2016:ao5.png}}
  - After validation is complete, review the results, and click Next.\\ {{:windows:install:sql2016:ao6.png}}
  - On the Summary page, review all of your selections, and click Finish to begin the process. This will take a few moments to complete. When the process is complete, close SQL Server Management Studio.\\ {{:windows:install:sql2016:ao7.png}}
  
At this point you should be able to use the cluster name as the database server name just like any other SQL server to access the databases. Unless you have a need to have separate listeners to segregate access to databases, there is no need to configure a listener. Skipping a separate listener also consume one less IP address on the network.