Teradata Connector TPT Bulk performance
Posted: Thu Sep 02, 2010 12:21 pm
I am working on tuning the performance of a job that processes 2-3 billion records. The job Design is read from sequential file [using file pattern, as there will be many files], transform data using transformer and load to Teradata target table using Teradata connection with Bulk/update and Parallel Sync (TPT). I am testing the job with different configurations (GRID/Cluster, partitions etc).
When I run the job on Grid using 2 node 4 partiion (or on Cluster with 8 node config file) I am getting the optimum performace. I am not seeing any difference in performance by varying max sessions/max partition sessions and min sessions.
I would like to know if some thing can be tuned further with 'requested sessions' and 'sessions per player'. But I am not sure how it works with Teradata connector in TPT and what are the environment variables for these? Any insight on this would be greatly appreciated. Thanks!
When I run the job on Grid using 2 node 4 partiion (or on Cluster with 8 node config file) I am getting the optimum performace. I am not seeing any difference in performance by varying max sessions/max partition sessions and min sessions.
I would like to know if some thing can be tuned further with 'requested sessions' and 'sessions per player'. But I am not sure how it works with Teradata connector in TPT and what are the environment variables for these? Any insight on this would be greatly appreciated. Thanks!