As I understand from the docs, data from a single shard goes to same path.data directory.
So, let's suppose, my path.data looks like ["/disk1", "/disk2"]. and I want maximum throughput for indexing. To achieve that, what I would ideally do is, assign 2 shards per node and expect ES to create these two shards on different path directories. However, I'm observing that both shards are created on "/disk1".
Why is that? Should this change? Is this an optimization that we can do in future versions?
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.