Azure Stack TP3 Stability (Reboot the XRP VM)


If you have been deploying and using Azure Stack TP3, you may have noticed after a few days the portal starts behaving slower and in my experience closer to a week it stops working altogether.  This will vary depending on what you're doing and your hardware.  Looking at the VM guests, you will notice that the XRP VM is consuming all its memory.  While you could give the machine more memory, this does mess with the expected infrastructure sizing and eventually, it will consume whatever memory you give it. This will hopefully be addressed soon. However, a simple workaround is to reboot the XRP VM. Why do anything manually when you can script it?  This very simple script creates a scheduled task that will run on Sunday night at 1 am.  The task will stop and start the XRP VM and then trigger the existing ColdStartMachine task that makes sure all the Azure Stack services are running.

[powershell] #run on host server as the AzureStackAdmin user $AzureStackAdminPAssword = 'YOURPASSWORD'

$Action = New-ScheduledTaskAction -Execute 'Powershell.exe' -argument '-command "Get-VM MAS-Xrp01 | Stop-VM -force;Get-VM MAS-Xrp01 | start-vm;sleep 180;Stop-ScheduledTask ColdStartMachine ;start-scheduledtask ColdStartMachine"' $trigger = New-ScheduledTaskTrigger -Weekly -DaysOfWeek Sunday -at 1am Register-ScheduledTask -Action $Action -Trigger $trigger -taskname "XRPReboot" -Description "Restart XRP VM weekly" -RunLevel Highest -User "$env:USERDOMAIN\$env:USERNAME" -Password $AzureStackAdminPAssword [/powershell]