Question

Workspace runner dillema - different results

  • 24 January 2020
  • 3 replies
  • 4 views

Badge +10

Whenever I run my child workspace by itself I get correct results on my translation written to a SQL Server table.

If I run this workspace from its parent workspace I lose attributes coming from a specific output stream going into a featuremerger with multiple suppliers..

Top picture is result from child workspace; bottom is if child is ran from a runner. At the bottom I lost values to HTTP response success, FeatureLayerCount and Selected Features.

These 3 attributes come from junction 6d: Workflow stats.

I added a Feature holder before featuremerger hoping that can solve the issue but it did not.

Seeking ideas or anyone who have similar experiences with workspacerunner.

Thanks!

 


3 replies

Badge +2

Do you have the option "Wait for job to complete" checked in your WorkspaceRunner transformer?

Badge +10

Do you have the option "Wait for job to complete" checked in your WorkspaceRunner transformer?

It is set yo Yes. The thing about using workspacerunner is you don't have a visual of what's happening with the child workspace until the translation is finished and you open the log file. perhaps I should compare the two translation logs and see where it's deviating.

Badge +10

Looking at the logs of two successful runs of the same child workspace and just searching for Junction_15 where the dropped attributes are coming from, this is how they compare.

Examining the second log file at 56.5 seconds I found out the root of the problem.

Is there a way to optimize memory before I run the parent? I understand two FME.exe are simultaneously running so that's an instant memory overhead. An option is to stay in one workbench file and run my second workspace from the shutdown script with python.

 

Reply