1 of 1 people found this helpful
Extract refreshes are done by the Backgrounder processes. Each Backgrounder process will have it's own log file. If you zip up the logs and then extract them out somewhere where they can be searched they'll show up such as "backgrounder-0.log", "backgrounder-1.log", "backgrounder-2.log", etc. Of course this depends if your system is multi-node and how the backgrounders are spread throughout the tableau cluster. If it's really just one or two dashboards failing then you could do a seach on "ERROR: canceled" from the error in the tooltip but if it's been a number of extracts failing it might be easier to search on the name of the extract itself.
Another option if you think you'll be doing this a lot would be to use the LogShark tool. The installation of LogShark comes with the MongoDB already included. Though you do have to install an instance of a Postgres database but the link to download is included in the installation instructions. LogShark will parse through a set of zipped logs and build workbooks for different components. One of the workbooks will cover all Backgrounder activity and the tooltips are set to be really large so would think it could show the issue without needing to go into the log file itself.
Oh geez, I feel silly now, I was searching on my Primary but all of the Backgrounders are on the Worker! Thanks Mark for indirectly pointing that out Redoing my search now...
While I'm now able to find and match the error with the workbook I still cannot see why there's a failure. The error logging is not verbose enough, too cryptic
For a deeper dive the backgrounder logs could be set in "debug" mode via the tabadmin command (tabadmin backgrounder.log.level debug). A restart will be required for the change to take effect. After the logs are captured can set back to "info".
Yup, doing that I have a batch script to set that up for me:
REM Turn on DEBUG error logging level for Tableau Server. tabadmin stop tabadmin set server.log.level debug tabadmin set vizqlserver.log.level debug tabadmin set vizportal.log.level debug tabadmin set wgserver.log.level debug tabadmin start
To return the server to regular logging:
REM Turn on default (INFO) error logging level for Tableau Server. tabadmin stop tabadmin set server.log.level -d tabadmin set vizqlserver.log.level -d tabadmin set vizportal.log.level -d tabadmin set wgserver.log.level -d tabadmin start
The additional logging info provided nothing helpful that I could see. The lack of robust error handling/reporting is a gripe of mine for pretty much all software.