I did some deeper annalise and I guess/suppose that this cluster is optimized for number of shares instead of disk usage. Is there any simple way to change it?
That's correct, and this is deliberate. Balancing by size is expensive and unnecessary. All your nodes have ample free space, so why does anything need to be done differently?
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.