Tableau Vizql, Backgrounder and Data server down

Question

Yesterday our extracts failed to refresh with the following message (image extract_error):

Failure: Failed 1 time. Sign in failed. Resolution Details: Check the Data Connection page for necessary updates to an access token or embedded credentials.

I verified that all our passwords were unchanged and test connections which were successful.

The tableau dashboards now give an error message saying:

HTTP 404: Unable to connect to the server "localhost". Check that the server is running and that you have access privileges to the requested database. (image tableau_error)

Further, when I opened the Server Status page, I saw that our one of our two Vizql, backgrounder and data servers were down. We have two of each and only one of them is active for all three of them. (image server_status)

So, I decided to remote desktop into the server and run the tabadmin status -v command and strangely it is showing that all processes are running. (image tabadmin_status)

Finally, I opened a case with Tableau Customer Portal and letting them know about this issue (they asked me send them the log.zip file) but the mean time I was trying to problem solve this issue. Any help would be really appreciated.

Answer 1

After trying a lot of things, one process seemed to work.

1. Stopped the tableau server
2. Configured it to run 1 Vizql server process instead of 2
3. Started the server again

Finally, it worked. The status page now shows all the processes are active. Hopefully, this helps someone who is facing a similar problem.

Answer 2

This may be caused by a firewall issue. Since tabadmin status -v returned all as "running" the cluster is healthy and this is a false alert. The firewall rules could be allowing just the first port and not the entire range (see https://onlinehelp.tableau.com/current/server/en-us/ports.htm) to respond to requests from the application server to build that fancy table with the green and red boxes.

The firewall can be reverted/altered behind the scenes for a number of reasons, usually windows updates or regular group policy synchronization.

Try disabling the windows firewall (https://www.faqforge.com/windows-server-2016/turn-off-firewall-windows-server-2016/), or add an inbound rule allowing access to all ports if your org policy doesn't allow you to actually turn it off. (Follow the steps here, except use "All Local Ports" instead of "Specific Local Ports" https://www.parallels.com/blogs/ras/configuring-windows-server-firewall-for-parallels-ras/)

Answer 3

I had a similar problem and followed these similar steps that Sravee mentioned above to bring the all processes back to active.

1. Stopped the server
2. Change the configuration for VizQL server from 2 to 1
3. Started the server
4. Enter the licence key (else the server status page will show unlicensed error)
5. Note: This does not bring the site back but this step is for 'tricking' VizQL server
6. Stopped the server again
7. Change the VizQL configuration from 1 to 2 now.
8. Start the server
9. Enter the license key

This steps did bring back the server back to active for us. Posting to see if this helps who faces the same problem. Thank you so much.