none
After ASDK gracefull Shutdown and move, not all VM are mounted

    Question

  • ADSK Version: 1.0.171020.1

    Because I needed to move my ASDK server to another location, I proceded to a gracefull shutdown.

    I turned-it on later, even after a couple or hours or days, only the 13 "Insfrastructure VM are UP" : 
    All other VMs are invisible, not mounted in Hyper-V.
    Administration portal is very instable, with a lot of "cloud shining"

    I'm on a static IP configuration, behind a router. IP of AzS-BGPNAT01 are good.

    I did a second gracefull shutdown,...still with the same issue.

    Thank you for your help,

    Best regards,

    Daniel

    Thursday, December 07, 2017 4:46 PM

All replies

  • Hello,

    After restarting my Azure Stack host, some VMs may not automatically start.

      

    After rebooting your host, you may notice Azure Stack services are not immediately available. This is because Azure Stack infrastructure VMs and RPs take a little bit to check consistency, but will eventually start automatically.

     

    You may also notice that tenant VMs don't automatically start after a reboot of the Azure Stack development kit host. This is a known issue, and just requires a few manual steps to bring them online:

     

    1. On the Azure Stack development kit host, start Failover Cluster Manager from the Start Menu.
    2. Select the cluster S-Cluster.azurestack.local.
    3. Select Roles.
    4. Tenant VMs will appear in a saved state. Once all Infrastructure VMs are running, right-click the tenant VMs and select Start to resume the VM.

    from Known issues

     

    Let us know how it goes.

          

    We apologize for any inconvenience and appreciate your time and interest in Azure Stack.

    If you experience any issues with Azure Stack or the current ASDK release, please feel free to contact us.

           

     Thanks


    Gary Gallanes

    Thursday, December 07, 2017 8:37 PM
    Moderator
  • Hi Gary,

    Thank you for your answer.

    Following your emal, I started all the tenant VM who where in the Saved state, and let the VM in the stopped state as is. (Infrastructure VM were already started)

    Unfortunatly, the admin portal is still in an unhealthy state :(

    Thursday, December 07, 2017 10:54 PM
  • Hello,

    Are you getting an error message when attempting to access the Admin Portal?

    Can you post back the error you are getting?

    Also - Can try the following workaround to restart ARM app pool (or do an iisreset) on was01 VM and see if it restores your connectivity to the Admin Portal?

    1. Make sure the WAS & WASP VMs are running.

    2. On the VMs, make sure the sites are running. We first hit the shell site. Here is how you can do it from the console VM:

     

    Enter-PSSession -ComputerName AzS-WAS01

    Import-Module WebAdministration

    Get-ChildItem -PSPath IIS:\Sites | fl name,state | ft -AutoSize

     

    3. You can test connection to the site (name resolution, port connectivity):

     

    Test-NetConnection -Port 443 -ComputerName portal.local.azurestack.external -InformationLevel Detailed

    Test-NetConnection -Port 443 -ComputerName AzS-WAS01 -InformationLevel Detailed

     

    Repeat the tests for WASP VM also if the tenant portal is also failing. This will set you on a path to understand what is wrong.  We apologize for any inconvenience and appreciate your time and interest in Azure Stack.

             

    We apologize for any inconvenience and appreciate your time and interest in Azure Stack.

    If you experience any issues with Azure Stack or the current ASDK release, please feel free to contact us.

           

     Thanks



    Gary Gallanes



    Tuesday, December 12, 2017 12:30 AM
    Moderator
  • Hi Gary,

    Thanks you for your answers, but I'm afraid it doesn't help

    Here are the answers to your questions and screenshots of actions you requested: https://msdndaniel.blob.core.windows.net/azurestack/20171212Answers.pdf

    Best regards,

    Daniel

    Tuesday, December 12, 2017 10:48 AM
  • Hello Daniel,

    I have an ASDK in my lab having the exact same issues after a graceful shutdown = (Rebooting by running Stop-AzureStack from the Protected End Point.)

     

    Can you go to ‘All Resources’ in the Admin Portal and check the Health of the Storage Resource Provider?

    My is showing unhealthy. It looks like my S-Cluster instance didn’t come up clean on my reboot.

    I restarted the cluster, roles & VMs but it didn’t make a difference. Same sad Portal - crying clouds

     I’ve raised an internal incident and will post back the outcome.

    If the ASDK stamp is healthy, a graceful shutdown = (Rebooting by running Stop-AzureStack from the Protected End Point.) should allow you to safely shutdown ASDK and restart it at a Customer site for a demo. At this point, it looks like your best bet is to redeploy.

     

    We apologize for any inconvenience and appreciate your time and interest in Azure Stack.

    If you experience any issues with Azure Stack or the current ASDK release, please feel free to contact us.

         

     Thanks


    Gary Gallanes


    Thursday, December 14, 2017 12:29 AM
    Moderator
  • Hi Gary,

    I checked, and have the same symptom, the Storage Resource Provider has the same "red stop icon" as yours....my platform is dead !

    Please, warmly ask the team to test and re-test the "Gracefull shutdown"....and find a way to debug and repair the platform in case of error. I can't suffer to reinstall my platform every two weeks to both unrecoverable issues and new updates we can't install inPlace !

    Today, my concern is ASDK is not enough reliable, I can't trust it to plan demos to my customers...because it could broke at any time,...and I need 2 or 3 full days to rebuild it.
    For a product in GA, it's not possible to continue like that!

    We need to trust ASDK knowing that it share the same code as a real production AzureStack.
    I don't talk about performance, redondancy or RTO ou RPO. I talk about reliability, operations,maintenance tasks, trust of the cloud layer.

    If all the nodes of a real Azure Stack are down (datacenter reboot) we need to know that it is recoverable by Microsoft.

    What happen if a customer has a Electricity outage on a 4 nodes AzureStack ? the Azure Stack is dead with no possibility to recover ?
    Please Microsoft, Prove its working by leveraging the ASDK experience and support.

    Let me know if I need to reinstall it before Xmass or if there is a way to investigate on my platform or yours and recover it.
    tired to have for the third time no other choice than a full reinstall of the platform, with no garanties the next build fixed it.

    I stay tuned, and still happy to help you to leverage the platform with my feedbacks.

    Best regards,

    Daniel

    13 hours 5 minutes ago