Hello all,
I've just started looking at Tez. I've setup Tez locally and have run sample
MapReduce job with Tez as a replacement for Yarn MRV2. I plan to use Tez
independently (No hive or pig). I have the following questions
1. Are there performance gains to use Tez for Mapreduce jobs apart from
container reuse ? Want to make sure I'm not missing anything
2. More technically, from what should be the behavior of a processor while
generating outputs. Should it write to all KeyValueWriters returned from
getOutputs() in the processor. The WordCount example retrieves an output by
name and writes to it.
thanks,Madhu
