Hello Server admins.
First of all, i want to say thank you to Toby Erkson for published guides:
Please vote for the Idea here (also voted):
I want to share my little story.
We have Tableau Server 10.0.1.
4 nodes, separate node for 9 backgrounder processes.
We had spotted Tableau Server status page showed that one backgrounder was down (backgrounder 8:8258)
It was in runtime for 38K sec (10 hours). Don't know why (we have limit to 7200 sec). Based on Tableau Server admin view (Background Tasks for Extracts) that was last activity on that backgrounder.
Not sure about description error - that backgrounder was still hanging and didn't go live
On Tableau Server i found in Task Manager that one of backgrounder process had no activity:
- Columns Threads, IO Reads, IO Writes, IO Read Bytes IO Write Bytes were motionless.
Our assumption was after that long running task (extract update) it was frozen and stuck.
We were trying to identify which backgrounder.exe is for which TS process is. Main idea was just to kill it.
I explored your docs and did the same, but:
- Version 10 has another structure
- Task Manager cannot show full line for Command Line column
- We didn't receive any notification from Tableau Server that it is down (probably because it wasn't really down, it just stuck)
Logs from Splunk analysis had that error for vizportal:
2016-12-28 09:20:48.927 +0000 (Production,*username*,90ukkyOzIJUeHLR1b2Q7ybl0Jb5oPFXo,WGODlgohEQsAAC4kBcEAAAPL) catalina-exec-489 : WARN com.tableausoftware.domain.admin.BackgrounderWorkerStatusHelper - Connection error to url: http:/node1:80/admin/status_backgrounder/3/8258 with exception: Read timed out
2016-12-28 10:03:49.792 +0000 (Production,*username*,Idkxx2oJJEfSzUQVa0ejbB4GEzoJECks,WGON2wohEQsAAC4kXbUAAAOX) catalina-exec-475 : WARN com.tableausoftware.domain.admin.BackgrounderWorkerStatusHelper - Error response code '503' from status url: http://*backgrounder_node*:80/admin/status_backgrounder/3/8258
After manual killing - background process was automatically picked up and started refreshing extracts.
I hope that will help somebody to solve the same problem.
UPDATE: Sep 2018
Tried to kill long running extract - after kill, TS returned it to refresh... My guess is "backgrounder.reset_schedules_on_startup" option is re-activating canceled refresh after killing backgrouder.exe