Trying to deploy a workflow on a server with a Conda Environment Propagation node.
Using the Analystics Platform on the server (executor) the node works well.
As soon as I Execute as a job, I have the following error:
Node: Conda Environment Propagation: Failed to reach out to the Conda installation located at ’ '.
The preferences are exported and inserted in the executor.epf. I checked if it’s read by the server using the link defined in the service and the preferences are well defined.
The server is running on Windows Server. The preference is set as follow:
It seems that the line is not properly read by the server.
Any advice on that issue ?
Thanks a lot
I think what has to be put into the preferences has changed. Maybe you got this from an old documentation (but do not quote me on that)? I thought our Python settings start with /instance/org.knime.python.python2 or something like that. What I would suggest is that you export your preferences from the local Analytics Platform and only keep all the lines related to Python, then change the values according to what you want to have on the server, then deploy that epf to the server. This makes sure you have proper preferences.
Thanks for your answer.
I have already exported the .epf from the Knime Analytics Platform on the server, with the correct version (same as the executor). And extracted the python part of the file. Which is different from the previous version I agree.
If this is still an issue, can you let us know what are the contents of your combined-preferences file found in your-executor-workspace/.metadata/.plugins/org.knime.product/combined-preferences.epf?
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.