Edit

Windows reboot loop on an Azure VM

  • Article

This article describes the reboot loop you may experience on a Windows Virtual Machine (VM) in Microsoft Azure.

Symptom

When you use Boot diagnostics to get the screenshots of a VM, you find the virtual machine is booting but the boot process is getting interrupted and the process is starting over.

Screenshot shows the virtual machine is booting.

Cause

The reboot loop occurs because of the following causes:

Cause 1

There is a third-party service that is flagged as critical and it cannot be started. This causes the operating system to reboot. To resolve this issue, see Solution for cause 1.

Cause 2

Some changes were made to the operating system. Usually, these are related to an update installation, application installation, or a new policy. You may have to check the following logs for additional details:

  • Event Logs
  • CBS.logWindows
  • Update.log

To resolve this issue, see Solution for cause 2.

Cause 3

A registry hive corruption could cause this issue. To resolve this issue, see Solution for cause 3.

Solution

Tip

If you have a recent backup of the VM, you may try restoring the VM from the backup to fix the boot problem.

To resolve this problem, back up the OS disk, and attach the OS disk to a rescue VM, and then follow the solution options accordingly, or try the solutions one by one.

Solution for cause 1

  1. Once the OS disk is attached to a working VM, make sure that the disk is flagged as Online in the Disk Management console and note the drive letter of the partition that holds the \Windows folder.

  2. If the disk is set to Offline, then set it to Online.

  3. Create a copy of the \Windows\System32\config folder in case a rollback on the changes is needed.

  4. On the rescue VM, open the Windows Registry Editor (regedit).

  5. Select the HKEY_LOCAL_MACHINE key and then select File > Load Hive from the menu.

  6. Browse to the SYSTEM file in the \Windows\System32\config folder.

  7. Select Open, type BROKENSYSTEM for the name, expand the HKEY_LOCAL_MACHINE key, and then you will see an additional key called BROKENSYSTEM.

  8. Check which ControlSet the computer is booting from. You will see its key number in the following registry key.

    HKEY_LOCAL_MACHINE\BROKENSYSTEM\Select\Current

  9. Check which is the criticality of the VM agent service through the following registry key.

    HKEY_LOCAL_MACHINE\BROKENSYSTEM\ControlSet00x\Services\RDAgent\ErrorControl

  10. If the value of the registry key is not set to 2, then go to the next mitigation.

  11. If the value of the registry key is set to 2, then change the value from 2 to 1.

  12. If any of the following keys exist and they have value 2 or 3, and then change these values to 1 accordingly:

    • HKEY_LOCAL_MACHINE\BROKENSYSTEM\ControlSet00x\Services\AzureWLBackupCoordinatorSvc\ErrorControl
    • HKEY_LOCAL_MACHINE\BROKENSYSTEM\ControlSet00x\Services\AzureWLBackupInquirySvc\ErrorControl
    • HKEY_LOCAL_MACHINE\BROKENSYSTEM\ControlSet00x\Services\AzureWLBackupPluginSvc\ErrorControl

  13. Select the BROKENSYSTEM key and then select File > Unload Hive from the menu.

  14. Detach the OS disk from the troubleshooting VM.

  15. Remove the disk from the troubleshooting VM and wait about 2 minutes for Azure to release this disk.

  16. Create a new VM from the OS disk.

  17. If the issue is fixed, then you may have to reinstall the RDAgent (WaAppAgent.exe).

Solution for cause 2

Restore the VM to the last known good configuration, follow the steps in How to start Azure Windows VM with Last Known Good Configuration.

Solution for cause 3

Note

The following procedure should only be used as last resource. While restoring from regback may restore access to the machine, the OS is not considered stable since there is data lost in the registry between the timestamp of the hive and the current day. You need to build a new VM and make plans to migrate data.

  1. Once the disk is attached to a troubleshooting VM, make sure that the disk is flagged as Online in the Disk Management console.

  2. Create a copy of the \Windows\System32\config folder in case a rollback on the changes is needed.

  3. Copy the files in the \Windows\System32\config\regback folder and replace the files in the \Windows\System32\config folder.

  4. Remove the disk from the troubleshooting VM and wait about 2 minutes for Azure to release this disk.

  5. Create a new VM from the OS disk.

Contact us for help

If you have questions or need help, create a support request, or ask Azure community support. You can also submit product feedback to Azure feedback community.