Can't use ConfigMap for /opt/flink/conf because entrypoint tries to write to it

Nobody has claimed this yet.

It looks like you are relying on the ENTRYPOINT script to append the contents of OPERATOR_FLINK_CONFIG env var to `/opt/flink/conf/flink-conf.yaml`. Unfortunately, this makes it impossible to use a ConfigMap mounted as a volume to provide a default `flink-conf.yaml` and `log4j.properties`, because ConfigMap volumes are intentionally read-only. I can't see a way to make this happen.

Alternatively, you could provide a way for the Operator to control default values for checkpoint folders etc. For example to ensure that state is always stored to a specific place (eg. S3 bucket) with the Job name as a component in the path (job name as one of the folder names). Or, to provide the Zookeeper quorum info for that environment. Generally, you could allow the operator to provide (templated) default flink-conf.yaml. An appropriate `log4j.properties` can be included in a Docker base image. But it would be nice for the operator to provide job-specific config.

Start by tracing the ENTRYPOINT behavior that appends OPERATOR_FLINK_CONFIG to /opt/flink/conf/flink-conf.yaml and examine how ConfigMap volumes are mounted. Clarify whether the goal is to avoid writes to a mounted configuration directory or to add operator-controlled, templated defaults; done should support the chosen configuration path without the read-only volume failure.

Written by the indexing model from the issue text.