Beekeeper Patch Automation: Powershell Validation Mneumonics and Recovery Actions

Written by John Hann on Friday, January 3rd 2020 — Categories: Beekeeper Software, Documentation

beekeeper patch automationBeekeeper, our proprietary patching automation tool, can be integrated with Powershell scripts for customized patching validation and recovery options. Read on to learn how to insert mneuomics into your script to accomplish this as well as how to recover from failed patching when detected.
 

Beekeeper Mnuemonics or Replacement Strings for PowerShell Validation Tasks

Beekeeper provides three mnemonics or replacement strings that are used to dynamically insert the cluster/node name into the script at execution time.

When a validation task is assigned to these phases:

You can use the #!nodename!#  replacement string in a PowerShell script to get the name of the node:

 

$OutTime = Get-Date

$TargetNode = ”#!nodename!#”

$WarningText = “Rebooting: ” + $TargetNode + ” at ” + $OutTime

Write-Warning $WarningText

Restart-Computer $TargetNode -Force

 

Make sure that the mnemonic is enclosed within quotes.

If you need to run a validation task during cluster or group phases such as:

You can use the #!clustername!# replacement string in a PowerShell script to get the name of the group/cluster.

SCCM – Check SCCM Agent at Cluster Start.ps1 in our github repository serves as an example on how to use the cluster/group name for Application Groups, Windows Failover Clusters, and Exchange DAGs.

The #!entityname!# would be used when the type (node/cluster) and is not important.
 

Patching failure recovery actions as PowerShell Validation Tasks

When patching execution jobs fail, you may need to recover services and such.  Beekeeper provides the “On Node Failure” and “On Group/Cluster Failure” validation task phases.

If a node belonging to an  Application Group fails to apply an update, it will be marked as failed and execute the “On Node Failure” phase if you have assigned a validation task to the phase.  An example would be to restart a service or even reboot that node. 

Windows Service example:

$Nodename = “#!nodename!#”

$SvcStatus = Get-Service “My Service” -ComputerName $NodeName 

$WarningText = $NodeName + ” My Service Status: ” + $SvcStatus.Status

Write-Warning $WarningText

If ($SvcStatus.Status -ne “Running”)

   {

    Get-Service “My Service” -ComputerName $NodeName | Restart-Service | Out-Null

    $WarningText = $NodeName + ” My Service Started”

    Write-Warning $WarningText

   }

This code uses the Beekeeper string replacement #!nodename!# to get the node that failed, and then uses “Write-Warning” cmdlet to get the output into the Beekeeper execution log.

Reboot-Node.ps1 in our github repository serves as an example on how to reboot and wait for the node to resume as a validation task. 

If you need to execute the same recovery action on all members of the Application Group, Windows Failover Cluster, or Exchange DAG, you can use the SCCM – Check SCCM Agent at Cluster Start.ps1 in our github repository serves as an example on how to run code against each member of the Application Group, Windows Failover Cluster, or Exchange DAG.