Skip to main content

Running some truncate and loads from DB2 to ORACLE.

Since the query is basically 'select *', is there any way at all to increase performance of FME selects from DB2?

One table of 72k rows takes about 30 seconds. That isn't monstrous but as we get to larger tables the time just increases. The same truncate and load in Informatica takes 4 seconds.

Looking for any 'out of the box' ideas here. i would like to use FME because dev time is much faster but i can't justify the 'job run time and resource' usage with multiple tables.

I have read most of the suggestions that i have found with general searching. User where statements, order by, select less columns..etc.

Also tried odbc connection and straight jdbc string connection but no difference.

 

Thanks.

@robotix The article Performance Tuning FME will give you some tips on where to look to see where the time is going, especially the section on the log files - Post-Mortem Performance Tuning. The other questions to ask are;

  • where is FME installed relative to the databases (and Informatica), separate machine entirely, on DB2 machine on Oracle machine?
  • In the log - is there a difference between session duration and CPU time?
  • Do you have Feature Caching On|Off?
  • Any other transformers or processing in the FME Workbench?

If you attach the workspace and log we could take a closer look.


@robotix The article Performance Tuning FME will give you some tips on where to look to see where the time is going, especially the section on the log files - Post-Mortem Performance Tuning. The other questions to ask are;

  • where is FME installed relative to the databases (and Informatica), separate machine entirely, on DB2 machine on Oracle machine?
  • In the log - is there a difference between session duration and CPU time?
  • Do you have Feature Caching On|Off?
  • Any other transformers or processing in the FME Workbench?

If you attach the workspace and log we could take a closer look.

They are all VMs. I will take a look at the article and examine the log.

Thanks!


They are all VMs. I will take a look at the article and examine the log.

Thanks!

All VMs on separate hosts in the same location.


Reply