Setting preferences for distributed executor decision on cores

Hi.
I am currently setting up a distributed environment for the KNIME server, having one server and one executor.
In the past I set the preferences for the executor in the workflow repository. So that the python environment was installed on the same server as the executor.

  1. When executing a node using python where is the node executed, on the server or on the machine hosting the executor.
  2. Where and how do I set the preferences for the distributed executor?
  3. In case I set it on the KNIME server, which path do I need to enter, the one of the executor machine?
  4. How do I specify the executor machine in the config file.

I tried to make sense out of the documentation, but for somebody who is new to KNIME it is not self explanatory.

Thanks a lot for your help.

Preferences for the executors is now configured via the Client Profiles provided by the server. The profiles also can distributed configurations to the KNIME AP instances as well.

https://docs.knime.com/2020-12/server_admin_guide/index.html#management-client-preferences

Thank you very much for the link. That answers my question where I need to configure the executor epf file. However is does not answer on which machine I need to install anaconda, on the server or on the executor?
Normally the EPF file would look like this:
/instance/org.knime.workbench.core/database_timeout=120
/instance/org.knime.ext.r.bin/knime.r.home=C:\Program Files\R\R-4.0.5
/instance/org.knime.python2/condaDirectoryPath=C:\ProgramData\Anaconda3
/instance/org.knime.python2/python2CondaEnvironmentDirectoryPath=C:\ProgramData\Anaconda3
/instance/org.knime.python2/python3CondaEnvironmentDirectoryPath=C:\ProgramData\Anaconda3\envs\pullplanning
/instance/org.knime.python2/pythonEnvironmentType=conda
/instance/org.knime.python2/serializerId=org.knime.serialization.flatbuffers.column

So how does a distributed executor know where to access Python? How do I reference in the EPF file to the distributed executor? Is the file path entered the path on the executor? This all is not clear in the documentation.

The documentation is mentioning Linux, what about Windows Server? How should the Client Profile look there?

On windows I would suggest using the KNIME AP (client) to setup the preferences within the GUI, then you can export those settings to an epf file. Then you can simply copy that file as a client profile, or if you want to break the file up into smaller profiles you can.

If you want to manually create the files, you will need to escape the windows paths separator.

Python needs to be installed with the executor, which is where the actual processing takes place.

The server is where the executor gets it’s profiles. This link talks through how to setup the client profiles. This guide is for 4.12.x, 4.13 will work with this setup, but now has other options.

Sorry I was on vacation. Thank you very much for your support.