I am using a System Center 2019 DPM as a backup server that is running on its own hardware. Besides other backups, it backs up files from a file server cluster.

The file servers use a deduplicated volume to store the files. The backup job backs up a folder on that volume and not the entire drive. It stores the backup data on local MBS (HDDs). During scheduled synchronization - sometimes (not always) - the following error occurs:

The DPM service was unable to communicate with the protection agent on <THE_BACKUP_SERVER_ITSELF>. (ID 65 Details: An existing connection was forcibly closed by the remote host (0x80072746))

The error resolves itself automatically with the next scheduled synchronization on the next day, or by another self initiated synchronization (by the job itself).

How can I prevent this error from happening at all? It is strange that the error occurs when the backup server is performing a backup on another machine, but is having problems to communicate to the protection agent that is running on the backup server itself.

Similar problem, but on DPM 2012 (without a solution): https://social.technet.microsoft.com/Forums/windows/en-US/37c97307-db55-4fca-84e8-115de70ba93e/reccuring-dpm-erorr-0x80072746

What I tried

• I enabled throttling for both file servers, but that did not help.
• I set the following registry values on the DPM and both file servers and restarted each DPMRA service (according to this and this):

  New-ItemProperty -Path "HKLM:\SOFTWARE\Microsoft\Microsoft Data Protection Manager\Agent" -Name "ConnectionNoActivityTimeout" -PropertyType DWord -Value 7200
  New-ItemProperty -Path "HKLM:\SOFTWARE\Microsoft\Microsoft Data Protection Manager\Agent" -Name "ConnectionNoActivityTimeoutForNonCCJobs" -PropertyType DWord -Value 7200
  Restart-Service DPMRA