This is a long overdue topic that I have wanted to cover for a couple of years and have finally got a chance to setup for a lab environment. I wanted to expand upon Brent Ozar’s article: How to Set Up SQL Server 2012 AlwaysOn Availability Groups. I also used Brent’s SQL Server 2012 Availability Groups Checklist.
This will be the first article of a 4 part series that covers the entire process.
I am going to start with setting up a lab environment that covers all the features you would add to prep for a SQL Server 2012 AlwaysOn Availability group with a share nothing storage cluster configuration. I will share a few tips and scripts with you as well.
I begin this process by installing the following VM’s (all with the GUI and updated with latest patches via Windows Update) on my laptop:
- 1 x Domain Controller with Windows Server 2012R2
- 3 x SQL 2012 Servers (Primary, Secondary and Readable Replica) with Windows Server 2012R2
- 1 x Windows 8.1 Client with SMS (Central Management Server)
To prepare for this cluster lab, I created 3 Switches for my Hyper-V networks:
- Domain
- Cluster
- Public
To create these networks, you will need to open up Hyper V Manager and then click on the Virtual Switch Manager:
For the Domain and Cluster, I used internal network:
For the Public I created external network:
Next I am going to assign IP addresses for my network by using a series of netsh commands with PowerShell to run them all at once. I can run all of the commands at once, by listing each command then placing a semi colon (;) between each command.
Here is my plan for my LAN network:
- SMLW12R2DC01 172.16.0.10 255.255.0.0
- SMLW12SQLAG1 172.16.0.20 255.255.0.0
- SMLW12SQLAG2 172.16.0.21 255.255.0.0
- SMLW12SQLAG3 172.16.0.22 255.255.0.0
- SMLSQLCLUS 172.16.0.25 255.255.0.0 (Used Later)
- SMLW81VM04 172.16.0.30 255.255.0.0
Here is my plan for my Cluster network:
- SMLW12SQLAG1 192.168.0.20 255.255.0.0
- SMLW12SQLAG2 192.168.0.21 255.255.0.0
- SMLW12SQLAG3 192.168.0.22 255.255.0.0
Before I run the script I will need to verify each NIC is connected to the correct network and then I will need to rename them from the default Ethernet name in order for my script to work.
—————- Script to set LAN (Domain) NIC —————-
netsh interface ipv4 set address name=LAN source=static address=172.16.0.10 mask=255.255.0.0 gateway=172.16.0.1;netsh interface ip set dnsserver “LAN” static 172.16.0.10 primary;netsh interface ip add dnsserver name=LAN 172.16.0.1 index=2
—————- Script to set Cluster NIC —————-
netsh interface ipv4 set address name=Cluster source=static address=192.168.0.20 mask=255.255.0.0
—————- Script to clear NetBIOS/DNS cache on NIC (Plus enable Quick Edit on cmd.exe) —————-
ipconfig /flushdns; nbtstat -R;nbtstat -RR; ipconfig /registerdns; Set-ItemProperty –path “HKCU:\Console” –name QuickEdit –value 1
Update PowerShell help
Next I update my PowerShell help files, since I use PowerShell a good bit by typing Update-Help.
Now I run all of the commands listed above and my NICs are configured for each network:
Please see my earlier article on how you can create some cool short cuts like seen below:
Last but not least, I created 4 additional hard drives (vhd) to separate my database files:
- E- bin (10 GB)
- H- Data (10 GB)
- O- Logs (5 GB)
- T- Temp (5 GB)
Note: The reason we separate these files is that data files are written randomly, whereas logs are written sequentially. Having these files on the same drive will cause performance issues. TempDB is created every time SQL server is started and for security/sanity reasons, you should place backup files on another drive(s). For more information, please see this TechNet article.
I use the bin drive for configuration files, backups, scripts, etc. The data drive is for my SQL data files, logs is stored on my “O” drive and a tempdb drive. Now I am ready to setup the extra disks for my SQL drives.
I attempted to do it at first manually, but it got old fast and it was annoying to get all the pop ups:
So I decided to script it and for each drive you could do it like this. (It will check for all raw disks, initialize them and then create/formats (as NTFS with GPT) new partitions to use the entire disk space, assigns a drive letter and a label for the drives without prompting you for input) For me even that gets old and I was not happy with the code:
—————- Script to setup a single disk —————-
Initialize-Disk -Number 1
New-Partition -DiskNumber 1 -AssignDriveLetter -UseMaximumSize
Format-Volume -DriveLetter E -NewFileSystemLabel “bin” -Confirm:$false
Here is a cleaned up version that checks for all raw disks, initializes them, then creates/formats (as NTFS with GPT) new partitions to use the entire disk space, assigns a drive letter and a label for the drives without prompting you for input:
—————- Script to setup all disks —————-
Get-Disk | Where {$_.PartitionStyle -eq ‘raw’} | % {
Initialize-Disk -Number $_.number
if ($_.number -like ‘1’)
{
New-Partition -DiskNumber 1 -DriveLetter E -UseMaximumSize
Format-Volume -DriveLetter E -NewFileSystemLabel “bin” -Force -Confirm:$false
}
elseif ($_.number -like ‘2’)
{
New-Partition -DiskNumber 2 -DriveLetter H -UseMaximumSize
Format-Volume -DriveLetter H -NewFileSystemLabel “Data” -Force -Confirm:$false
}
elseif ($_.number -like ‘3’)
{
New-Partition -DiskNumber 3 -DriveLetter O -UseMaximumSize
Format-Volume -DriveLetter O -NewFileSystemLabel “Logs” -Force -Confirm:$false
}
elseif ($_.number -like ‘4’)
{
New-Partition -DiskNumber 4 -DriveLetter T -UseMaximumSize
Format-Volume -DriveLetter T -NewFileSystemLabel “Temp” -Force -Confirm:$false
}
}
—————- End of Script —————-
After running the script you can see here is the drives all configured:
At this point, I have all systems configured with machines names, IP addresses and the latest windows patches. In the next article, I will show you the steps on how to create the domain.