We built a new Execution Context (EC) Docker image and deployed it successfully in Kubernetes, but the executor cannot start and keeps restarting.
Observed Behavior
-
Executor pods continuously cycle (
Creating → Running → Terminating). -
The EC status remains Starting up and never becomes Ready.
-
The same deployment procedure works with another EC image.
Key Errors in Logs
The failing EC consistently shows these errors during KNIME startup:
java.io.FileNotFoundException:
/home/knime/knime/configuration/.../artifacts.xml (Permission denied)
followed by:
java.lang.IllegalStateException:
The platform metadata area could not be written:
/home/knime/knime-workspace/.metadata
What This Means
The KNIME runtime cannot write to its configuration and workspace directories:
/home/knime/knime/configuration
/home/knime/knime-workspace/.metadata
Because of this filesystem permission failure, the KNIME platform initialization aborts and the executor process exits.
Therefore the issue is not the path itself, but the write permissions or ownership of these directories in the failing Docker image/runtime.
I am trying to set up the HUB to test with potential customer (who has downloaded the latest AP, which is 5.10), but stuck here….. Please help, thank in advance.