4 Replies Latest reply on Feb 23, 2016 6:35 AM by Toby Erkson

    Why does the Backgrounder fail though there seems no memory issue ?

    Oindrila B

      Hi All,

       

      I have Tableau Server version 9.2. Data sources are extracts and MySql data connection. I have 3 backgrounders. The system is of very high

      configuration (64bit, 32 Gb RAM, 7.5 TB memory). MySql Db is on different server than Tableau (which is also with good config)

       

      Sometimes there seems error as Lost Connection-

      com.tableausoftware.nativeapi.dll.TableauException: [MySQL][ODBC 5.3(w) Driver][mysqld-5.5.39]Lost connection to MySQL server during query

      in background status of extracts although the refresh extract query is not heavy and executes in seconds. When I look at show processlist there was no memory issue/lock or anything during the execution. Also the system memory on both the servers seem fine too during the error time.

       

      I have seen in some threads where it is mentioned that the backgrounder restart itself in some interval.

      Is that true?  I am getting this issue in this version only. Din't get in previous version. Also, not every time same extract gets failed,it's random everyday!

       

      Would really appreciate if anyone can share their thoughts on this.

      Thanks in advance,

      Oin

        • 1. Re: Why does the Backgrounder fail though there seems no memory issue ?
          sudhakar reddy

          Hi Ghosh,

           

          As per my understanding it seems that the extract is taking a lot of time to refresh.

           

          tableau has a mechanism of backgrounder getting restart after every 1800 sec so please increase the backgrounder timeout by using below command.

           

          See below link for the command.

           

           

          http://onlinehelp.tableau.com/v7.0/server/en-us/tabadmin_set.htm

           

           

          Regards,

          Sudhakar Reddy

          • 3. Re: Why does the Backgrounder fail though there seems no memory issue ?
            Oindrila B

            Thanks Toby and Sudhakar.

             

            I increased the backgrounder.extra_timeout_in_seconds to 3600 from 1800

            and gateway.timeout to 3600 from 1800 but the server

            got more slowed down and repetitively the tasks were getting failed.

             

            I raised the concern to support team and they asked me to check below points:-

             

            - A network issue.

            - The query has insufficient time to complete.

            - The query result exceeds the maximum packet setting.

            Then I checked,

             

            - A network issue-  Both the servers (Tableau and Mysql are on different

            system but on same network (LAN). Checked the network  connectivity ,no

            issue found.

             

            - The query has insufficient time to complete- There are extract queries

            which executes in seconds but falling under the issue category. I  guess

            probably this could not be the case but not sure.

             

            - The query result exceeds the maximum packet setting- I looked into

            increasing the following variables of our db:-

             

            *net_read_timeout to set 60 which was 30*

            *connect_timeout to set 900 which was 10*

            *max_allowed_packet to 16M which was 1048576*

             

            This setting in db actually reduced my failed jobs but not completely. Support team is analyzing the logs further. Will post any update I get.

            1 of 1 people found this helpful
            • 4. Re: Why does the Backgrounder fail though there seems no memory issue ?
              Toby Erkson

              O Ghosh wrote:

               

              Thanks Toby and Sudhakar.

               

              I increased the backgrounder.extra_timeout_in_seconds to 3600 from 1800 and gateway.timeout to 3600 from 1800 but the server got more slowed down and repetitively the tasks were getting failed.

              I raised the concern to support team and they asked me to check below points:-

               

              ...

              - The query result exceeds the maximum packet setting.

              Then I checked, ...

              - The query result exceeds the maximum packet setting- I looked into

              increasing the following variables of our db:-

                   *net_read_timeout to set 60 which was 30*

                   *connect_timeout to set 900 which was 10*

                   *max_allowed_packet to 16M which was 1048576*

              This setting in db actually reduced my failed jobs but not completely. Support team is analyzing the logs further. Will post any update I get.

              Thanks for the update, I find this interesting and informative   Please keep us updated!