Updating StarWind VSAN Windows application on a 2-Node Hyper-V Cluster using SCCM
This article is specifically recommended for multi-location deployments, providing a detailed walkthrough to successfully update StarWind VSAN on a 2-Node Hyper-V Cluster using the SCCM.
Description
In case you have multiple locations with Failover Cluster running StarWind VSAN as a shared storage, installing updates for StarWind VSAN might take a lot of time and effort, thus it makes sense to automate it. This article provides a guide on updating StarWind VSAN using the System Center Configuration Manager (SCCM) package in a 2-node Hyper-V cluster, recommended for cases with distributed multiple locations.
Disclaimer: StarWind Support does not write scripts and SCCM package on demand. Custom script troubleshooting is not supported. The script provided in this article is as an example and can be customized by the end-user according to his needs.
Prerequisites: Before proceeding with the update, ensure the following prerequisites are met:
- Backup: Conduct a comprehensive backup of Hyper-V virtual machines and critical data to mitigate potential data loss during the update.
- System Center Configuration Manager (SCCM): Confirm that SCCM is properly configured and operational in the multi-location environment.
- StarWind VSAN Update Package: Download the latest StarWind VSAN update package from the official StarWind website, compatible with the current version.
- Cluster Health Check: Verify the health and status of the 2-Node Hyper-V Cluster across all locations to address any issues before initiating the update.
To avoid production interruption, StarWind VSAN cannot be installed on both cluster nodes simultaneously and must be updated in a one-by-one manner. The update script is designed to minimize the probability of production interruption and is intended to execute the update only on the local node with additional verification steps.
Assuming the above, the package has to be deployed to the SCCM collection which consists only of the “first” nodes in the cluster, like SW-HCA-01. Once it is executed successfully ( the script execution time could be up to 5 hours) the same package can be deployed to the “second” node in the cluster, like SW-HCA-02.
It is recommended to deploy the package to smaller collections and update nodes by clusters to avoid cases when in the same cluster the first node is running on the latest StarWind VSAN build, while the second one is running on the old build for a long time. It’s better to keep the same build for the partner nodes to avoid possible differences that could appear in future builds.
NOTE: Please keep in mind that it’s mandatory to test the StarWind VSAN update via SCCM on a pilot group before running it for multiple locations. The post-deployment check must be done to confirm that the StarWind VSAN version is updated, StarWind HA devices are synchronized and cluster nodes and roles are up and running.
SCCM Package description
Before deployment to the selected collection, the SCCM package should be created. The package consists of two files:
* starwind-v8.exe – StarWind VSAN installer. The actual version can be downloaded using this link (file name unchanged): https://www.starwindsoftware.com/tmplink/starwind-v8.exe
* swupdate.ps1 – PowerShell script which performs preparation steps, StarWind VSAN update, and post-install actions.
Script description
While executing on the StarWind node, the script swupdate.ps1 is writing a log file swupdate.log which is located in the same folder. The log file location can be changed if required, by changing $LogFile variable.
While installing, StarWind VSAN writes its own log file vsan_update.log which is also located in the folder with the script.
The script swupdate.ps1 performs the following actions:
- check if there is a pending reboot state on the server.
There are a lot of registry keys to check it, but most of them were commented on because in most cases StarWind VSAN installation it’s not so important. There were only the most important ones uncommented. If a pending reboot is detected, the script will exit with code 3001.
- check nodes state in the cluster – both nodes must be in the online state. If one of the nodes is not in an online state, the script will exit with code 3002.
- check the ISCSI sessions state on the local node. If some session is not connected or not persistent, the script will exit with code 3003.
- kill StarWindManagementConsole process. It’s required to run the StarWind VSAN installer. If the process is left running, the script will exit with code 3004.
- update the StarWindX PowerShell module, using the installer file. It has to be installed separately to avoid possible issues later. If the installation is failed, the script will exit with code 3005. To investigate the failure, vsan_update.log can be read additionally.
- check StarWind devices synchronization state on both nodes. If some device is not synchronized, the script will exit with code 3006.
- Identify a partner cluster node and move cluster resources to it. If failed, the script will exit with code 3007.
- pause the local node in the cluster. If that operation is not successful, the script will exit with code 3008.
- update StarWind VSAN on the local node, using the installer file. If the installation is failed, the script will exit with code 3009. To investigate the failure, vsan_update.log can be read additionally.
- check StarWind devices synchronization state on both nodes after the synchronization. If some device is not synchronized, the script will check the synchronization 30 times every 10 minutes until all devices are synchronized. It was implemented to make sure that devices are synchronized after the update in case full synchronization occurs. If some devices are not synchronized after 5 hours, the script will exit with code 3010. The number of checks and interval can be changed if required, by changing $timeoutSeconds and $timeoutCount variables.
- resume local node in the cluster. If that operation is not successful, the script will exit with code 3011.
- If completed successfully, the script exits with code 0.
Script body:
Set-ExecutionPolicy -Scope Process Bypass $ErrorActionPreference = "Stop" $timeoutSeconds = 600 $timeoutCount = 30 $LogFile = "$PSScriptRoot\swupdate.log" $TimeStamp = (Get-Date).ToString("yyyy-MM-dd HH:mm:ss") Write-Host "$TimeStamp [INFO] Starting update for node $env:COMPUTERNAME" -ForegroundColor Green "$TimeStamp [INFO] Starting update for node $env:COMPUTERNAME" | Out-File $LogFile function WriteLog{ [CmdletBinding()] param ( [Parameter(Mandatory=$false, position = 1)] [ValidateSet('OK', 'INFO', 'WARNING')] $level, [Parameter(Mandatory=$false, position = 2)] $message, [Parameter(Mandatory=$false, position = 3)] $exitCode ) try { if ($null -eq $level){ Write-Host "$TimeStamp [ERROR] [$exitCode] $($error[0].Exception.Message)" -ForegroundColor Red "$TimeStamp [ERROR] [$exitCode] $($error[0].Exception.Message)" | Out-File $LogFile -Append "$TimeStamp #### Exception begin ####" | Out-File $LogFile -Append ">>> Exception Message >>>" | Out-File $LogFile -Append $error[0].Exception.Message | Out-File $LogFile -Append ">>> Script StackTrace >>>" | Out-File $LogFile -Append $_.ScriptStackTrace | Out-File $LogFile -Append ">>> Invocation Info >>>" | Out-File $LogFile -Append $_.InvocationInfo.line.trim() | Out-File $LogFile -Append "$TimeStamp #### Exception end ####" | Out-File $LogFile -Append exit $exitCode } else { Write-Host "$TimeStamp [$level] $message" -ForegroundColor Green "$TimeStamp [$level] $message" | Out-File $LogFile -Append } } catch { $error[0].Exception.Message | Out-File $LogFile -Append } } ### Check pending reboot requirements function Test-RegistryKey { [OutputType('bool')] [CmdletBinding()] param ( [Parameter(Mandatory)] [ValidateNotNullOrEmpty()] [string]$Key ) $ErrorActionPreference = 'Stop' if (Get-Item -Path $Key -ErrorAction Ignore) { $true } } function Test-RegistryValue { [OutputType('bool')] [CmdletBinding()] param ( [Parameter(Mandatory)] [ValidateNotNullOrEmpty()] [string]$Key, [Parameter(Mandatory)] [ValidateNotNullOrEmpty()] [string]$Value ) $ErrorActionPreference = 'Stop' if (Get-ItemProperty -Path $Key -Name $Value -ErrorAction Ignore) { $true } } function Test-RegistryValueNotNull { [OutputType('bool')] [CmdletBinding()] param ( [Parameter(Mandatory)] [ValidateNotNullOrEmpty()] [string]$Key, [Parameter(Mandatory)] [ValidateNotNullOrEmpty()] [string]$Value ) $ErrorActionPreference = 'Stop' if (($regVal = Get-ItemProperty -Path $Key -Name $Value -ErrorAction Ignore) -and $regVal.($Value)) { $true } } $pendingReboot = @( <# { Test-RegistryKey -Key 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Component Based Servicing\RebootPending' } { Test-RegistryKey -Key 'HKLM:\Software\Microsoft\Windows\CurrentVersion\Component Based Servicing\RebootInProgress' } { Test-RegistryKey -Key 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\WindowsUpdate\Auto Update\RebootRequired' } { Test-RegistryKey -Key 'HKLM:\Software\Microsoft\Windows\CurrentVersion\Component Based Servicing\PackagesPending' } { Test-RegistryKey -Key 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\WindowsUpdate\Auto Update\PostRebootReporting' } #> { Test-RegistryValueNotNull -Key 'HKLM:\SYSTEM\CurrentControlSet\Control\Session Manager' -Value 'PendingFileRenameOperations' } { Test-RegistryValueNotNull -Key 'HKLM:\SYSTEM\CurrentControlSet\Control\Session Manager' -Value 'PendingFileRenameOperations2' } # { Test-RegistryKey -Key 'HKLM:\SOFTWARE\Microsoft\ServerManager\CurrentRebootAttemps' } # { Test-RegistryValue -Key 'HKLM:\SYSTEM\CurrentControlSet\Services\Netlogon' -Value 'JoinDomain' } { (Get-ItemProperty -Path 'HKLM:\SYSTEM\CurrentControlSet\Control\ComputerName\ActiveComputerName').ComputerName -ne (Get-ItemProperty -Path 'HKLM:\SYSTEM\CurrentControlSet\Control\ComputerName\ComputerName').ComputerName } { if (Get-ChildItem -Path 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\WindowsUpdate\Services\Pending') { $true } } ) try { WriteLog INFO "Looking for pending reboot requirements" $needUpdate = 0 foreach ($element in $pendingReboot) { if (& $element) { $needUpdate += 1 } } if ($needUpdate -ne 0) { Write-Error "There is pending reboot requirements detected" } else { WriteLog OK "There are no pending reboot requirements detected" } } catch { WriteLog -exitCode 3001 } ### Check ClusterNodes try { WriteLog INFO "Get cluster node state" foreach ($node in Get-ClusterNode) { if ($node.State -eq "Up"){ WriteLog OK "Cluster node $node ONLINE" } else { Write-Error -Message "Cluster node $node NOT ONLINE" } } } catch { WriteLog -exitCode 3002 } ### Check iSCSI sessions try{ WriteLog INFO "Get iSCSI sessions state" foreach ($session in Get-IscsiSession) { if ($session.IsConnected){ WriteLog OK "iSCSI session $($session.SessionIdentifier) is CONNECTED" } else { Write-Error -Message "iSCSI session $($session.SessionIdentifier) not CONNECTED" } if ($session.IsPersistent) { WriteLog OK "iSCSI session $($session.SessionIdentifier) is PERSISTENT" } else { Write-Error -Message "iSCSI session $($session.SessionIdentifier) not PERSISTENT" } } } catch { WriteLog -exitCode 3003 } ### Try to kill StarWindManagementConsole process if exist try { if (Get-Process | Where-Object {$_.Name -eq "StarWindManagementConsole"}) { WriteLog INFO "Try to kill StarWindManagementConsole process" Get-Process | Where-Object {$_.Name -eq "StarWindManagementConsole"} | Stop-Process WriteLog OK "StarWindManagementConsole process killed" } } catch { WriteLog -exitCode 3004 } ### Update StarWindX try { WriteLog INFO "Try to install StarWindX" $pinfo = New-Object System.Diagnostics.ProcessStartInfo $pinfo.FileName = "$PSScriptRoot\starwind-v8.exe" $pinfo.RedirectStandardError = $false $pinfo.RedirectStandardOutput = $false #$pinfo.CreateNoWindow = $true $pinfo.UseShellExecute = $false $pinfo.Arguments = "/NORESTART /LOG=vsan_update.log /VERYSILENT /SUPPRESSMSGBOXES /COMPONENTS=StarWindxDll,StarWindXDll\powerShellEx" $p = New-Object System.Diagnostics.Process $p.StartInfo = $pinfo $p.Start() | Out-Null $p.WaitForExit() WriteLog INFO "StarWindX EXECUTED" if ($p.ExitCode -ne 0) { $swexitcode=$p.ExitCode Write-Error "Installation of StarWindX failed. For details see vsan_update.log in $PSSCriptRoot directory. Error code $swexitcode" } else { WriteLog OK "StarWindX installed successfully" } } catch { WriteLog -exitCode 3005 } ### Check StarWind HA sync state on local and partner node try { WriteLog INFO "Get StarWind HA devices synchronization status" Import-Module StarWindX $server = New-SWServer 127.0.0.1 3261 root starwind $server.Connect() foreach($target in $server.Targets) { if ($target.Devices[0].DeviceType -eq "HA Image") { if ($target.Devices[0].SyncStatus -eq "1"){ WriteLog OK "HA device $($target.Alias) on server $($server.IP) synchronized" $partner = New-SWServer $target.Devices[0].Partners[0].HeartbeatChannels[0].Address 3261 root starwind $partner.Connect() foreach($partnerTarget in $partner.Targets) { if ($partnerTarget.Devices[0].DeviceType -eq "HA Image"){ if ($partnerTarget.Devices[0].SyncStatus -ne "1"){ Write-Error "Partner HA device $($partnerTarget.Alias) on server $($partner.IP) not synchronized" } else { WriteLog OK "HA device $($partnerTarget.Alias) on server $($partner.IP) synchronized" } } } } else { Write-Error "Partner HA device $($Target.Alias) on server $($server.IP) not synchronized" } } } Remove-Module StarWindX } catch { WriteLog -exitCode 3006 } ### Move cluster resourses to second node try { WriteLog INFO "Get free cluster node" $freeClusterNode = Get-ClusterNode | Where-Object {$_.Name -ne $env:COMPUTERNAME -and $_.State -eq "Up"}[0] WriteLog OK "Free cluster node is $freeClusterNode" WriteLog INFO "Move cluster resources to $freeClusterNode" Get-ClusterNode | Get-ClusterGroup | Move-ClusterGroup -Node $freeClusterNode | Out-Null WriteLog OK "Move cluster resources to $freeClusterNode done" WriteLog INFO "Move CSV to $freeClusterNode" Get-ClusterSharedVolume | Move-ClusterSharedVolume -Node $freeClusterNode | Out-Null WriteLog OK "Move CSV to $freeClusterNode done" } catch { WriteLog -exitCode 3007 } ## Suspend cluster node try { WriteLog INFO "Try to PAUSE $env:COMPUTERNAME in cluster" Suspend-ClusterNode -Name $env:COMPUTERNAME | Out-Null WriteLog OK "$env:COMPUTERNAME PAUSED" } catch { WriteLog -exitCode 3008 } ### Update starWind VSAN try { WriteLog INFO "Try to install StarWind VSAN" $pinfo = New-Object System.Diagnostics.ProcessStartInfo $pinfo.FileName = "$PSScriptRoot\starwind-v8.exe" $pinfo.RedirectStandardError = $false $pinfo.RedirectStandardOutput = $false #$pinfo.CreateNoWindow = $true $pinfo.UseShellExecute = $false $pinfo.Arguments = "/NORESTART /LOG=vsan_update.log /VERYSILENT /SUPPRESSMSGBOXES /COMPONENTS=Service,service\starflb,service\starportdriver" #/NOCLOSEAPPLICATIONS /NORESTARTAPPLICATIONS stop-service -name "starwindservice" -Force sleep 300 get-process | ?{$_.name -eq "StarWindService"} | stop-process -Force sleep 30 $p = New-Object System.Diagnostics.Process $p.StartInfo = $pinfo $p.Start() | Out-Null $p.WaitForExit() WriteLog INFO "StarWind VSAN EXECUTED" if ($p.ExitCode -ne 0) { $swexitcode=$p.ExitCode Write-Error "Installation of StarWind VSAN failed. For details see vsan_update.log in $PSSCriptRoot directory. Error code $swexitcode" } else { WriteLog OK "New StarWind VASN build installed successful" } } catch { WriteLog -exitCode 3009 } ### Check SW sync state try { WriteLog INFO "Get StarWind HA devices synchronization status" Import-Module StarWindX $syncState = 0 foreach($element in 1..$timeoutCount){ $server = New-SWServer 127.0.0.1 3261 root starwind if (!$server.Connected) { Start-Sleep -Seconds 10 $server.Connect() | Out-Null } foreach($target in $server.Targets) { if ($target.Devices[0].DeviceType -eq "HA Image") { if ($target.Devices[0].SyncStatus -ne "1"){ $syncState = 0 WriteLog WARNING "HA device $($target.Alias) on server $($server.IP) not synchronized" } else { WriteLog OK "HA device $($target.Alias) on server $($server.IP) synchronized" $syncState = 1 } } } if ($syncState -ne "1"){ WriteLog WARNING "NOT all HA device on server $($server.IP) synchronized. Retry $element of $timeoutCount" Start-Sleep -Seconds $timeoutSeconds } if ($element -eq $timeoutCount) { Write-Error "HA devices on the server $($server.IP) are not synchronized after $timeoutCount intervals of $timeoutSeconds seconds" } else { $partner = New-SWServer $target.Devices[0].Partners[0].HeartbeatChannels[0].Address 3261 root starwind $partner.Connect() foreach($partnerTarget in $partner.Targets) { if ($partnerTarget.Devices[0].DeviceType -eq "HA Image"){ if ($partnerTarget.Devices[0].SyncStatus -ne "1"){ Write-Error "HA device $($partnerTarget.Alias) on server $($partner.IP) not synchronized" } else { WriteLog OK "HA device $($partnerTarget.Alias) on server $($partner.IP) synchronized" } } } WriteLog OK "All HA devices on server $($server.IP) and server $($partner.IP) are synchronized" $server.Disconnect() break } } Remove-Module StarWindX } catch { WriteLog -exitCode 3010 } ## Resume cluster node try { WriteLog INFO "Try to RESUME $env:COMPUTERNAME in cluster" Resume-ClusterNode -Name $env:COMPUTERNAME -Failback NoFailback WriteLog OK "$env:COMPUTERNAME UP" } catch { WriteLog -exitCode 3011 } Write-Host "$TimeStamp [OK] update for node $env:COMPUTERNAME SUCCESSFUL" -ForegroundColor Green "$TimeStamp [OK] update for node $env:COMPUTERNAME SUCCESSFUL" | Out-File $LogFile -Append exit 0
Useful links
The recent StarWind Virtual SAN build can be downloaded here: https://www.starwindsoftware.com/starwind-virtual-san#download
The complete Release Notes can be viewed by following this link: https://www.starwindsoftware.com/release-notes-build
Request a Product Feature
To request a new product feature or to provide feedback on a StarWind product, please email to our support at support@starwind.com and put “Request a Product Feature” as the subject.