1 of 1 people found this helpful
1) -v(verbose) doesn't change anything other than showing the status of the individual services.
More detail information from the Online help,
"Tells you whether or not Tableau Server is running and, if you use the
--verboseoption, gives you details on individual server process status, including whether a process is running and its process ID. The
tabadmin statuscommand obtains its information by connecting to the Windows Service tabsvc.exe, which in turn queries the tabspawn executables for each process. Because of this, it can sometimes display different information for the server processes than the status table on the Maintenance page, which queries the processes directly."
2) Every time you perform a tabadmin commands, it's written to the log file location at the following location,
Server location\Program Data\tableau\tableau server\logs. This should give you more detail on what might be happening.
Couple of thoughts,
1) Make sure to run the cmd as Administrator
2) When you restore from a backup copy from another machine, make sure to use tabadmin restore --no-config. This restores only the data and not the server information.
We can help if you can provide more details from the log files.
3) Tableau support would be a great option to consider.
Hey Vikram, thanks for your response.
I know that the verbose flag only affects output verbosity and not the server itself. I'm concerned that verbosity is altered after a reboot. IE if I run with the -v flag right after rebooting, it will give me detailed breakdown as I posted in my gist above.
but if I run the same command later, it gives me nothing, regardless of whether tabsvc is running or not. This is broken.
I am running as an administrator.
And I am looking at the log. There are only two messages of usefulness:
1. failure to connect to postgres
2. 'unable to determine if all components of the service started'.
(1) makes it sound like postgres isn't running, but further above I see that Tableau Server Repository is correctly started and running. I dont see any failure messages elsewhere.
(2) is more or less useless.
We do have a support contract with Tableau, and I've submitted a case there as well.
Thanks for pointing out that tabsvc is the middleman in tabadmin status, Vikram. that answered one of my questions.
I also now understand that tabspawn is responsible for boxing all of services that tableau uses. Which means I now just have to figure out why the API Server is reporting as stopped, even though it has a PID, which means tabspawn got far enough into starting it.
So I think I figured it out.
'tabadmin stop' shuts everything down, including tabsvc.
'tabadmin start' turns most things on, but not tabsvc.
tabsvc not coming on results in wgserver, repository, and file store not coming on correctly.
this is why 'tabadmin restart' nor 'tabadmin stop' followed by 'tabadmin start' were working correctly for me. this is also why doing 'tabadmin status -v' was working perfectly after a reboot: tabsvc was up perfectly (again: thanks Vikram for the hint!)
I dunno if that's a known issue, but for whoever comes in the future, making sure to manually start tabsvc before I do 'tabadmin start' is all I have to do to make everything start correctly every time.
Makes since why tabadmin start didn't work after the stop. tabsvc manages every single process and if one of the services doesn't start, we may see Degraded status. The services don't start due to various reason( ports changed, permissions etc).
That being said, i don't think it's a bug and i would say it's expected behavior. From the error message about 'unable to connect to postgres" this could be because of the ports. Postgres runs on 8060. Make sure this port is not assigned to some other applications. I think when you reboot the ports are available for tableau to use and during the restart the port is could be reassigned to a different application( my assumption).
I was able to restart the application without manually starting tabsvc. I would suggest to open a support email and they can confirm if it's a bug.
I sent a reply to your support case regarding the root issue. It is likely the repository and file system cannot start due to an authentication issue. Regenerating internal tokens should resolve the issue.
In regards to this post, I wanted to mention that the recommended and supported methods of starting and stopping Tableau Server are the tabadmin commands or from Tableau Server Monitor in the System Tray.
The Tableau Server Application Manager service (tabsvc) will be stopped and started automatically by these methods, and the service should never be stopped from the Windows services console. This could result in potential data corruption because this method does not gracefully shut down Tableau Server and it's individual components.
Let us know if you have any additional questions or concerns.