Is data all in the memory when running a m/r job with the es's querying result as the data source?
I have a question about the es-hadoop plugin.
If I want to run a m/r job on a hadoop cluster and the job reads data from es indexes , is the query result loaded in the memory ? And where is the job running? Since that the data source is not stored in the hdfs, which means that there is no datanode having the data source, how the job be dispatched?