Apache Branch Progress

• 03/08/2019

I can run NNA inside its own Jetty HTTP server within a Standby or Observer NameNode. I can also fully run the legacy NNA API / interface as shown in NamenodeAnalyticsMethods.java for Hadoop 3.x builds. All that is missing is to prune the builds down to only 3.x builds and provide the ServicePlugin implementation that will launch NNA for you. From there I would like to refactor and re-write NNA, changing its API to... well... actually be an API... I am thinking of adopting the WebHDFS standard as that will be more common with other Hadoop folks. At that time I would like to refactor NNA and possibly re-write as there is heavy code duplication currently within NNA. I can see NNA keeping the same features but with different URL mappings since API will change.

• 01/05/2020

HDFS-15763 created. I have several patches with an implemented NameNode plugin. Just need to massage them to work for 'trunk' in Apache.