A workflow runs in 12 minutes using the Analytics user interface. Running of the same workflow from the command line (on same windows pc) takes 4 hours, resulting data is the same. Any idea what could cause this difference in the efficiency? Does the command line version use the same knime.ini -file, which has -Xmx4096m ? Thanks for any ideas…
the automatic execution of KNIME Workflow is a lot faster with KNIME Server. KNIME Batch has a big overload because KNIME needs to be started in the back first. However 4h versus 12 minutes sounds like a virus scanner blocking the start of the executor.
Best, Iris
If I remember well my workflows actually took shorter time to execute from command line (Linux) vs Analytics Platform on Windows. But those were simple and short workflows (1-2 minutes). I used argument worflowFile instead of workflowDir…
Also what I realized (again if I remember well ) is when you use preferences just once it writes them down so there is no more need to use that argument again. For example I had ojdbc driver added on my Windows Platform and after calling a workflow from command line in Linux with .epf file from Windows the following file was modified: ..\workspace\.metadata\.plugins\org.eclipse.core.runtime\.settings\org.knime.workbench.core.prefs
I am running a relative small workflow (~30s in Knime UI) but it normally takes longer in batch mode. Is it because Knime was opened in the back every time I run a cmd command even though I already open Knime UI?
) is when you use preferences just once it writes them down so there is no more need to use that argument again. For example I had ojdbc driver added on my Windows Platform and after calling a workflow from command line in Linux with .epf file from Windows the following file was modified: