KNOWLEDGE BASE

Multi Node Tableau Server Backup with External File Store Fails With Error 'TTransportException'


Published: 13 Apr 2021
Last Modified Date: 19 Apr 2021

Issue

A Tableau Server multi-node environment with External File Store enabled, fails to complete the 'tsm maintenance snapshot-backup prepare' command when copying the repository backup to the network share location.

You may see the following error presented with the job failing at 86% "Error copying repository backup to the network share" and the following output from the CMD:

This job failed due to unexpected error: 'TTransportException'

Snapshot Prepare Error

The following error will be presented in the Tabadmincontroller log:

"ERROR com.tableausoftware.tabadmin.webapp.asyncjobs.JobStepRunner - Running step CopyRepositoryBackup failed
org.apache.thrift.TApplicationException: Internal error processing copyBackupToNetworkShare"

 

Environment

  • Tableau Server 2020.1+
  • Windows Server

Resolution

Adjust the Tableau Server topology so that the Active Repository is on the same node as where the Tabadmincontroller service is running. 

For example, when the Tabadmincontroller service is on node1 and the active Repository is on node2, perform a repository failover to make the node1 repository the active service.
  1.  tsm stop
  2.  tsm topology failover-repository -t node1
  3.  tsm pending-changes apply
  4.  tsm start
Once the active repository is running on the same node as the Tabadmincontroller service, run the snapshot-backup prepare again.

If the above does not apply or issues are encountered, please contact Tableau Support for further assistance.

Cause

The Tabadmincontroller process will copy the backup files to the External File Store location but if the leading postgres service is not colocated on the same node, Tabadmincontroller will not be able to copy those files.  

Tableau Development is actively implementing additional logic to this process which will be adjusted in a later release, currently under investigation.
Did this article resolve the issue?