Download PsExec from the Sysinternals page.

Run psexec -i -s "c:\Program Files\Internet Explorer\iexplore.exe" from an elevated command prompt.

This command will open Internet Explorer.

Go to Tools > Internet options > Connections > LAN settings .

Check the proxy settings for the system account.

If no proxy is configured and proxy details are provided, remove the details.

If a proxy is configured and the proxy details are incorrect, ensure the Proxy IP and Port details are correct.

Close Internet Explorer.

Unable to download vault credential file

The Microsoft Azure Recovery Service Agent was unable to connect to Microsoft Azure Backup

The specified vault credential file cannot be used as it is not downloaded from the vault associated with this server

Backup jobs completed with warning

When the MARS agent iterates over files and folders during backup, it might encounter various conditions that can cause the backup to be marked as completed with warnings. During these conditions, a job shows as completed with warnings. That's fine, but it means that at least one file wasn't able to be backed up. So the job skipped that file, but backed up all other files in question on the data source.

Conditions that can cause the backups to skip files include:

The backup service will mark these files as failed in the log file, with the following naming convention: LastBackupFailedFilesxxxx.txt under the C:\Program Files\Microsoft Azure Recovery Service Agent\temp folder.

To resolve the issue, review the log file to understand the nature of the issue:

You can use Add Exclusion rules to existing policy to exclude unsupported, missing, or deleted files from your backup policy to ensure successful backups.

Avoid deleting and recreating protected folders with the same names in the top-level folder. Doing so could result in the backup completing with warnings with the error: A critical inconsistency was detected, therefore changes cannot be replicated. If you need to delete and recreate folders, then consider doing so in subfolders under the protected top-level folder.

Failed to set the encryption key for secure backups

The activation did not complete successfully

Encryption passphrase not correctly configured

Backups don't run according to schedule

If scheduled backups don't get triggered automatically but manual backups work correctly, try the following actions:

Ensure the Windows Server backup schedule doesn't conflict with the Azure files and folders backup schedule.

Ensure the online backup status is set to Enable . To verify the status, take these steps:

Ensure the user account selected for running the task is either SYSTEM or Local Administrators' group on the server. To verify the user account, go to the General tab and check the Security options.

Ensure PowerShell 3.0 or later is installed on the server. To check the PowerShell version, run this command and verify that the Major version number is 3 or later:

$PSVersionTable.PSVersion

Ensure this path is part of the PSMODULEPATH environment variable:

<MARS agent installation path>\Microsoft Azure Recovery Services Agent\bin\Modules\MSOnlineBackup

If the PowerShell execution policy for LocalMachine is set to restricted , the PowerShell cmdlet that triggers the backup task might fail. Run these commands in elevated mode to check and set the execution policy to either Unrestricted or RemoteSigned :

Get-ExecutionPolicy -List
Set-ExecutionPolicy Unrestricted
Ensure there are no missing or corrupt PowerShell module MSOnlineBackup files. If there are any missing or corrupt files, take these steps:
From any machine that has a MARS agent that's working properly, copy the MSOnlineBackup folder from C:\Program Files\Microsoft Azure Recovery Services Agent\bin\Modules.
On the problematic machine, paste the copied files at the same folder location (C:\Program Files\Microsoft Azure Recovery Services Agent\bin\Modules).
If there's already an MSOnlineBackup folder on the machine, paste the files into it or replace any existing files.
To ensure changes are applied consistently, restart the server after performing the preceding steps.
Resource not provisioned in service stamp
Error
Possible causes
Recommended actions
The current operation failed due to an internal service error "Resource not provisioned in service stamp". Please retry the operation after some time. (ID: 230006)
The protected server was renamed.
- Rename the server back to the original name as registered with the vault. 

 - Re-register the server to the vault with the new name.
Job could not be started as another job was in progress
If you notice a warning message in the MARS console > Job history, saying "Job could not be started as another job was in progress", then this could be because of a duplicate instance of the job triggered by the Task Scheduler.
To resolve this issue:
Launch the Task Scheduler snap-in by typing taskschd.msc in the Run window
In the left pane, navigate to Task Scheduler Library -> Microsoft -> OnlineBackup.
For each task in this library, double-click on the task to open properties and perform the following steps:
Switch to the Settings tab.
Change the option for If the task is already running, then the following rule applies. Choose Do not start a new instance.
Troubleshoot restore problems
Azure Backup might not successfully mount the recovery volume, even after several minutes. And you might receive error messages during the process. To begin recovering normally, take these steps:
Cancel the mount process if it's been running for several minutes.
Check if you have the latest version of the Backup agent. To check the version, on the Actions pane of the MARS console, select About Microsoft Azure Recovery Services Agent. Confirm that the Version number is equal to or higher than the version mentioned in this article. Select this link to download the latest version.
Go to Device Manager > Storage controllers and locate Microsoft iSCSI Initiator. If you locate it, go directly to step 7.
If you can't locate the Microsoft iSCSI Initiator service, try to find an entry under Device Manager > Storage controllers named Unknown Device with Hardware ID ROOT\ISCSIPRT.
Right-click Unknown Device and select Update Driver Software.
Update the driver by selecting the option to  Search automatically for updated driver software. This update should change Unknown Device to Microsoft iSCSI Initiator:
Go to Task Manager > Services (Local) > Microsoft iSCSI Initiator Service:
Restart the Microsoft iSCSI Initiator service. To do this, right-click the service and select Stop. Then right-click it again and select Start.
Retry recovery by using Instant Restore.
If the recovery still fails, restart your server or client. If you don't want to restart, or if the recovery still fails even after you restart the server, try recovering from another machine.
Troubleshoot Cache problems
Backup operation may fail if the cache folder (also referred as scratch folder) is incorrectly configured, missing prerequisites or has restricted access.
Prerequisites
For MARS agent operations to succeed the cache folder needs to adhere to the following requirements:
Ensure 5% to 10% free volume space is available in the scratch folder location
Ensure scratch folder location is valid and accessible
Ensure file attributes on the cache folder are supported
Ensure the allocated shadow copy storage space is sufficient for backup process
Ensure there are no other processes (ex. anti-virus software) restricting access to cache folder
Increase shadow copy storage
Backup operations could fail if there isn't sufficient shadow copy storage space that's required to protect the data source. To resolve this issue, increase the shadow copy storage space on the protected volume using vssadmin as shown below:
Check the current shadow storage space from the elevated command prompt:

vssadmin List ShadowStorage /For=[Volume letter]:
Increase the shadow storage space using the following command:

vssadmin Resize ShadowStorage /On=[Volume letter]: /For=[Volume letter]: /Maxsize=[size]
Another process or antivirus software blocking access to cache folder
If you've antivirus software installed on the server, add the exclusion rules to the antivirus scan for:
Every file and folder under the scratch and bin folder locations - <InstallPath>\Scratch\* and <InstallPath>\Bin\*.
cbengine.exe
Backup or restore job is displayed as in progress in Azure for many days but is not visible in the console
If a MARS Agent backup or restore job crashes during execution, it is marked as failed in the MARS console, but the status might not get propagated to Azure. Hence, the job might be displayed as "in progress" in the Azure Portal even when it is not running. This stale job entry will be removed from the Azure Portal automatically after 30 days.
Common issues
This section covers the common errors that you encounter while using MARS agent.
SalChecksumStoreInitializationFailed
Error message
Recommended action
Microsoft Azure Recovery Services Agent was unable to access backup checksum stored in scratch location
To resolve this issue, perform the following steps and restart the server 
 - Check if there is an antivirus or other processes locking the scratch location files
 - Check if the scratch location is valid and accessible to the MARS agent.
SalVhdInitializationError
Error message
Recommended action
Microsoft Azure Recovery Services Agent was unable to access the scratch location to initialize VHD
To resolve this issue, perform the following steps and restart the server 
 - Check if antivirus or other processes are locking the scratch location files
 - Check if the scratch location is valid and accessible to the MARS agent.
SalLowDiskSpace
Error message
Recommended action
Unable to find changes in a file. This could be due to various reasons. Please retry the operation
To resolve this issue, verify the following steps and retry the operation:
 - Ensure the MARS agent is latest 
 - Verify and resolve storage issues that impact backup scratch space
MARS offline seeding using customer-owned disks (Import/Export) is not working
Azure Import/Export now uses Azure Data Box APIs for offline seeding on customer-owned disks. The Azure portal also list the Import/Export jobs created using the new API under Azure Data Box jobs with the Model column as Import/Export.
MARS agent versions lower than 2.0.9250.0 used the old Azure Import/Export APIs, which will be discontinued after February 28, 2023 and the old MARS agents (version lower than 2.0.9250.0) can't do offline seeding using your own disks. So, we recommend you to use MARS agent 2.0.9250 or higher that uses the new Azure Data Box APIs for offline seeding on your own disks.
If you've ongoing Import/Export jobs created from older MARS agents, you can still monitor them in the Azure portal, under Import/Export jobs.
Next steps
Get more details on how to back up Windows Server with the Azure Backup agent.
If you need to restore a backup, see restore files to a Windows machine.