[BUG] Get-AksHciConfig and Install-AksHci cannot find cloudconfig in the remote host

[erjosito]

Describe the bug No cloudconfig file can be found

To Reproduce Not sure. I am deploying in Azure in a 2 node HCI cluster. I initialized the nodes and used Set-AksHciConfig, no difference

Expected behavior Install-AksHciConfig and Get-AksHciConfig working

Screenshots

PS C:\Users\labadmin> set-akshciconfig
[10/20/2020 02:12:42] Creating configuration
 - Removing old configuration...
 - New configuration has been saved

PS C:\Users\labadmin> get-akshciconfig
[10/20/2020 02:07:27] Checking for configuration
 - Merging Windows Admin Center configuration
 - Loading Windows Admin Center configuration from 'C:\Users\labadmin\Windows Admin Center\aks-hci-settings.json'...
 - Processing configuration...
[10/20/2020 02:07:27] Creating configuration
 - Removing old configuration...
 - New configuration has been saved
[10/20/2020 02:07:27] Reading configuration
[10/20/2020 02:07:28] Validating configuration
[10/20/2020 02:07:28] Confirming Configuration
[10/20/2020 02:07:28] Determining deployment type
 - This is a multi-node deployment using failover cluster: AZSHCI
[10/20/2020 02:07:28] Verifying cloudconfig access file
 - Retrieving access file from WIN-MUCC37Q1OIO
Copy-Item : Cannot find path '\\WIN-MUCC37Q1OIO\C$\Users\labadmin\.wssd\cloudconfig' because it does not exist.
At C:\Program Files\WindowsPowerShell\Modules\Moc\0.2.8\Common.psm1:946 char:5
+     Copy-Item -Path $remotePath -Destination $destination
+     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : ObjectNotFound: (\\WIN-MUCC37Q1O...ssd\cloudconfig:String) [Copy-Item], ItemNotFoundException
    + FullyQualifiedErrorId : PathNotFound,Microsoft.PowerShell.Commands.CopyItemCommand
Unable to locate a valid cloudconfig access file.
At C:\Program Files\WindowsPowerShell\Modules\Moc\0.2.8\Common.psm1:1437 char:5
+     throw "Unable to locate a valid cloudconfig access file."
+     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : OperationStopped: (Unable to locat...ig access file.:String) [], RuntimeException
    + FullyQualifiedErrorId : Unable to locate a valid cloudconfig access file.

Environment (please complete the following information):

• OS: [e.g. iOS]
• Browser [e.g. chrome, safari]
• Version [e.g. 22]
• AKS-HCI Version (i.e. Public Preview 1)
• Kubernetes Version (i.e. 1.18.6)

Additional context Add any other context about the problem here.

Collect log files

• From a PowerShell Admin window run Collect-AksHciLogs
• If you are running into issues with the deployment wizard in Windows Admin Center, run Get-SMEUILogs.ps1 from the machine hosting Windows Admin Center.”

PS C:\Users\labadmin> get-akshcilogs
[10/20/2020 02:15:02] Checking for configuration
 - Merging Windows Admin Center configuration
 - Loading Windows Admin Center configuration from 'C:\Users\labadmin\Windows Admin Center\aks-hci-settings.json'...
 - Processing configuration...
[10/20/2020 02:15:02] Creating configuration
 - Removing old configuration...
 - New configuration has been saved
[10/20/2020 02:15:02] Reading configuration
[10/20/2020 02:15:02] Validating configuration
[10/20/2020 02:15:02] Confirming Configuration
[10/20/2020 02:15:02] Determining deployment type
 - This is a multi-node deployment using failover cluster: AZSHCI
[10/20/2020 02:15:02] Verifying cloudconfig access file
 - Retrieving access file from WIN-PSIERIALC37
Copy-Item : Cannot find path '\\WIN-PSIERIALC37\C$\Users\labadmin\.wssd\cloudconfig' because it does not exist.
At C:\Program Files\WindowsPowerShell\Modules\AksHci\0.2.8\Common.psm1:946 char:5
+     Copy-Item -Path $remotePath -Destination $destination
+     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : ObjectNotFound: (\\WIN-PSIERIALC...ssd\cloudconfig:String) [Copy-Item], ItemNotFoundException
    + FullyQualifiedErrorId : PathNotFound,Microsoft.PowerShell.Commands.CopyItemCommand
Unable to locate a valid cloudconfig access file.
At C:\Program Files\WindowsPowerShell\Modules\AksHci\0.2.8\Common.psm1:1437 char:5
+     throw "Unable to locate a valid cloudconfig access file."
+     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : OperationStopped: (Unable to locat...ig access file.:String) [], RuntimeException
    + FullyQualifiedErrorId : Unable to locate a valid cloudconfig access file.

PS C:\Users\labadmin> get-module -name akshci
ModuleType Version    Name                                ExportedCommands
---------- -------    ----                                ----------------
Script     0.2.8      AksHci                              {Get-AksHciCluster, Get-AksHciCredential, Get-AksHciKubernetesVersion, Get-AksHciLogs...}

Not sure where to find Get-SMEUILogs.ps1, it is not in the preview package I downloaded

[abhilashaagarwala]

Are you deploying this on Azure VMs?

[erjosito]

@abhilashaagarwala Yes

[erjosito]

Any idea what can I troubleshoot or try? Thanks!

[nwoodmsft]

You mentioned that you are attempting to install AksHci, but the script is indicating that you still have Windows Admin Center configuration present (which is a sign that Windows Admin Center was used to deploy AksHci on this cluster/node previously and no uninstall has been performed yet):

- Loading Windows Admin Center configuration from 'C:\Users\labadmin\Windows Admin Center\aks-hci-settings.json'...

Did your previous Windows Admin Center deployment complete successfully? I am wondering if you had a failed deployment through WAC and some aspects of that deployment are still present and have not been cleaned up yet.

We are improving the powershell "Uninstall-AksHci" cmdlet to be more resilient to this type of issue/state moving forward. For now, you may need to perform some manual cleanup on the cluster/nodes if you are planning to switch to using the Powershell for Day0 deployment of AksHci.

[erjosito]

Hey @nwoodmsft thanks for your answer! Yes, first I tried WAC, but after 2h it didnt work. After retrying with WAC a couple of times I switched over to pwsh. Would you have a link on how to "perform some manual cleanup on the cluster/nodes" to switch to pwsh from WAC?

[erjosito]

I renamed C:\users\labadmin\Windows Admin Center\aks-hci-settings.json and issued set-akshciconfig in each node, and install-akshci in one of the nodes. There was an error when removing the cloudagent directories, not sure whether this is going to affect the process...

- Removing cloudagent directory on WIN-PSIERIALC37...
Access is denied
    + CategoryInfo          : NotSpecified: (:) [Remove-Item], Win32Exception
    + FullyQualifiedErrorId : System.ComponentModel.Win32Exception,Microsoft.PowerShell.Commands.RemoveItemCommand
    + PSComputerName        : WIN-PSIERIALC37

[erjosito]

I think we can close this one, since deleting/renaming C:\users\labadmin\Windows Admin Center\aks-hci-settings.json made the trick :-)