And Benchmark Scenarios only said 'indexes 8.6M documents (POIs from Geonames, total 2.8 GB json) ', but I can find anything named PIO at www.geonames.org.
you looked at our classic benchmarks which are not open-sourced. The new benchmarks are based on Rally (which is open source). We publish the nightly results at https://elasticsearch-benchmarks.elastic.co/geonames/ (and we run also the benchmarks for more data sets). The benchmark specifications are available in a separate Github repository (https://github.com/elastic/rally-tracks) and if you look at the track.json files there you can figure out where you can download them. Having that said, you shouldn't really download them manually and just use Rally for that. It will automatically handle the download for you and run the benchmark.
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.