Fix PostgreSQL Support, Add AWS connectors on the Hadoop classpath and Optimize Image Size

[idirze]

This PR fixes issues with PostgreSQL, adds AWS connector on the Hadoop classpath and reduces the container size by removing unnecessary Hadoop components.

1. Fix PostgreSQL support and add ability to pass the username as a secret when needed:

• Ensures schematool correctly uses -dbType postgres
• Allows (optionally) injecting the DB username from a Kubernetes Secret for better security and flexibility

2. Add AWS connectors on the Hadoop classpath

• Use the bundles provided in ${HADOOP_HOME}/share/hadoop/tools/lib/{hadoop-aws-*.jar, aws-java-sdk-bundle-*.jar}

3. Optimize base image size by removing unnecessary Hadoop components:

• Removes unused YARN, MapReduce, client libraries, tests, examples, docs, and webapps
• Keeps only what is necessary for the Hive Metastore to operate