3 Replies Latest reply on Jul 23, 2018 9:37 AM by Jeff Strauss

    Extract processing for large tables, Hyper is slower than TDE

    Chirag Shah

      Hi,

       

      We have Tableau Server 10.3 deployed on an AWS EC2 instance (4-vCPU, 16 GB RAM and 100 GB disk space).

      There are 2 data sources scheduled to refresh the extracts twice a day from a Oracle table/view.

      DataSource1 has 70 columns/fields and 2.5 million records. This extract takes about 2-3 minutes for a full refresh.

      DataSource2 has 275 columns/fields and 2.5 million records. This extract takes about 40-50 minutes for a full refresh.

       

      We are upgrading the server to version 2018.1.3, with the expectation that .hyper extracts would perform similar to or better than .tde extracts.

       

      Installed Tableau Server 2018.1.3 on an identical AWS EC2 instance (4-vCPU, 16 GB RAM and 100 GB disk space), published new data sources (same as DataSource1 and DataSource2 mentioned above). Both data sources are referring to the same Oracle table/view.

      Extract for DataSource1 takes about 2-3 minutes for a full refresh (similar to that of .tde extract).

      Extract for DataSource2 always fails. com.tableausoftware.nativeapi.dll.ResourceGovernorException: The query time resource limit (8100 seconds) was exceeded.

       

      Both the data sources are clean extracts from DB, no calculated fields are present.

      Is there a way to improve the performance of .hyper extracts? Do .hyper extracts need more hardware resources?

       

      Thanks,

      Chirag