AnsweredAssumed Answered

Minimum configuration for a hadoop connection to a cluster?

Question asked by MichaelSegel on Aug 12, 2016
Latest reply on Aug 12, 2016 by mufeed

What's the minimum cluster information required to uniquely identify a mapr cluster in a Configuation instance?


Normally the java code would use a line like Configuration conf = new Configuration();


Which will pick up the hadoop setting from the /opt/mapr/conf/*.xml files.


But what happens if I want a client to connect to two different clusters?

I would have to set the configuration parameters separately.

Which parameters are required and which ones are going to be mapr specific?

(Apache doesn't have a CLDB ...)


My goal is to look at both MapRFS on each cluster... and I don't want to have to rely on NFS mounting of both clusters locally.

Any sample code snippets?