Allow storing the metastore configuration on Hadoop (HDFS, S3, etc)

[yruslan]

Background

Currently, metastore config is a part of the configuration that is passed to Pramen on 'spark-submit'.

Sometimes it is very helpful to store parts of the configuration in a central location, like Hadoop

Feature

Add an ability to store metastore or some other parts of configuration in Hadoop.

When Pramen starts it can start the Spark session based on the local configuration, and then load configuration from Hadoop (could be multiple files), and then load it alltogether.