To my surprise the results are pretty much the same, it looks as rally has run the same workload 3 times instead of 1 workload shared amongst the 3 nodes.
I think I was a bit confused about the results. Since there are some that improve like index append median that goes from 5500 docs/s to 12000 docs/s, wile others stay pretty much the same like index-stats that is around 100 ops/s in both
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.