Skip to main content

Azure VM: Add/attach the availability set to the existing VM

 Once we've created an Azure VM without an Availability set, later if we want to add it to an availability set there is no obvious option on the Azure portal. We can add an availability set only while creating the VM. 

The problem comes when we try to add always-on on IaaS SQL instance one of the pre-requisite is IaaS VMs should be on an availability set. So to enable Always on existing IaaS VM either we've to re-create the VM or need to create a new one & install the required s/w from scratch. 

To resolve this restriction, we can follow the below steps sequentially and achieve this. 

  • Get the list of attached OS + Data disk of existing VM
  • Get the Network interface details. 
  • Remove the Azure Disk Encryption (ADE) if it's enabled. 
  • Remove the Azure VM (delete it)
  • Now create the new with the same name & with Availability set. 
  • After re-creating the VM change the OS disks & Data disk [ask noted in step-1]. 
  • Start the VM and try to access it. 
  • Once you're able to access it & verify it's working, you can start enabling ADE (if required)

The above-mentioned process is all manual and it will take some time. there is well know PowerShell script to automate this. 

Here is the PowerShell script to add/attach availability set to existing VM. 


 
 # Set variables
    $resourceGroup = "<<your-resource-group>>"
    $vmName = "<<your-vm-name>>"
    $newAvailSetName = "<<your-availability-set-name>>"

# Get the details of the VM to be moved to the Availability Set
    $originalVM = Get-AzVM `
	   -ResourceGroupName $resourceGroup `
	   -Name $vmName

# Create new availability set if it does not exist
    $availSet = Get-AzAvailabilitySet `
	   -ResourceGroupName $resourceGroup `
	   -Name $newAvailSetName `
	   -ErrorAction Ignore
    if (-Not $availSet) {
    $availSet = New-AzAvailabilitySet `
	   -Location $originalVM.Location `
	   -Name $newAvailSetName `
	   -ResourceGroupName $resourceGroup `
	   -PlatformFaultDomainCount 2 `
	   -PlatformUpdateDomainCount 2 `
	   -Sku Aligned
    }
    
# Remove the original VM
    Remove-AzVM -ResourceGroupName $resourceGroup -Name $vmName    

# Create the basic configuration for the replacement VM. 
    $newVM = New-AzVMConfig `
	   -VMName $originalVM.Name `
	   -VMSize $originalVM.HardwareProfile.VmSize `
	   -AvailabilitySetId $availSet.Id
 
# For a Linux VM, change the last parameter from -Windows to -Linux 
    Set-AzVMOSDisk `
	   -VM $newVM -CreateOption Attach `
	   -ManagedDiskId $originalVM.StorageProfile.OsDisk.ManagedDisk.Id `
	   -Name $originalVM.StorageProfile.OsDisk.Name `
	   -Windows

# Add Data Disks
    foreach ($disk in $originalVM.StorageProfile.DataDisks) { 
    Add-AzVMDataDisk -VM $newVM `
	   -Name $disk.Name `
	   -ManagedDiskId $disk.ManagedDisk.Id `
	   -Caching $disk.Caching `
	   -Lun $disk.Lun `
	   -DiskSizeInGB $disk.DiskSizeGB `
	   -CreateOption Attach
    }
    
# Add NIC(s) and keep the same NIC as primary
	foreach ($nic in $originalVM.NetworkProfile.NetworkInterfaces) {	
	if ($nic.Primary -eq "True")
		{
    		Add-AzVMNetworkInterface `
       		-VM $newVM `
       		-Id $nic.Id -Primary
       		}
       	else
       		{
       		  Add-AzVMNetworkInterface `
      		  -VM $newVM `
      	 	  -Id $nic.Id 
                }
  	}

# Recreate the VM
    New-AzVM `
	   -ResourceGroupName $resourceGroup `
	   -Location $originalVM.Location `
	   -VM $newVM `
	   -DisableBginfoExtension
       
 

With this, we're good at adding the availability set to the existing Azure VM. The above steps & PowerShell script was tested on Jun/July 2020. 


Reference:

1. https://docs.microsoft.com/en-us/azure/virtual-machines/windows/change-availability-set

Comments

Popular posts from this blog

EFCore - Collate function

Search in SQL server query is generally case insensitive (by default or based on database level collation). Suppose we have an employees table with a row having first-name column value as "My-First-Name", so if we want to do the case-sensitive search we have to explicitly use the related collate: In EF 5 (currently in Release Candidate version [RC.2.20475.6]) Collate function got introduced which helps us to use our specific collation based search.  C# with EF5 code sample: var employeeCaseSensitiveSearch = _dbContext.Employees .Where(x => EF.Functions.Collate(x.FirstName, "Latin1_General_CS_AS") == "my-first-name") .FirstOrDefault(); A related database query will be something like this: T-SQL: Case sensitive search (use specific collation e.g.: Latin1_General_CS_AS) SELECT * FROM dbo.Employees AS e WHERE e.FirstName Collate Latin1_General_CS_AS = 'my-first-name' Some of the useful CSharp function which g...

Git - Update submodule

Sometimes in the large project when we separate it in different repositories we need to take reference of one repository in another where some the sub-module concepts. Now if we move our submodule to a different repository or renaming its repo-path then we've to also update the repository where this sub-module is getting referenced. Let's see how to update the git submodule.  Step-1: First clone the repo Step-2: Open git UI to create a new local branch from master  Step-3: Open git bash -- Now run below command to (remove submodules & add them):  -- Remove submodule git submodule deinit YOUR_FIRST_REPO git rm YOUR_FIRST_REPO git commit -m "Removed submodule YOUR_FIRST_REPO" rm -rf .git/modules/YOUR_FIRST_REPO -- Add submodule git submodule add <<..YOUR_FIRST_REPO git URL>> Step-4: All good, let's push our changes to the master

EFCore - Update command is not updating the underlying data

Why my update command using EFCore is not updating the underlying data in the database though in middle-tier code it's successful, not throwing any error? Let's try to get the root cause and how to solve this.  When creating the model class of related database object if the primary key name is Id then no need to explicitly mention the Key in ef-core OnModelCreating.  But if we've key name different from Id then we have to explicitly mention it in our OnModelCreating. Something like this:  entity.HasKey(c => new {c.CustId }); Assume in the database we've Customers table like this: CREATE TABLE [dbo].[Customers]( [CustId] [INT] IDENTITY(1,1) NOT NULL Primary Key, [FirstName] [VARCHAR](50) NULL ) Point to ponder: In the database table, column CustId is already defined as Primary key.  In EFCore model setup, it's up to us to configure the Key as (custId) or configure it as HasNoKey() Assume we've set our entity with Ha...