While optimizing it's pretty easy to determine the number of segments by
executing Optimize command with max_num_segments parameter.
My question is about indexing prior to optimizing.
I checked the http://www.elasticsearch.org/guide/en/elasticsearch/reference/0.90/index-modules-merge.html#tiered section
and I still couldn't find - how do I control the number of segments during
indexing? Is there a simple way to tell to ES that I am willing to pay with
indexing time and frequent merges but keep number of segments below 10, for
example?
It's a bit more complex. You can use the "segments_per_tier" to limit the
number of segments to be created at a tier. Together with
"max_merge_at_once" you can control the amount of running merges.
Mike McCandless is the authority who can explain how the tiered merge
policy works:
While optimizing it's pretty easy to determine the number of segments by
executing Optimize command with max_num_segments parameter.
My question is about indexing prior to optimizing.
I checked the Elasticsearch Platform — Find real-time answers at scale | Elastic section
and I still couldn't find - how do I control the number of segments during
indexing? Is there a simple way to tell to ES that I am willing to pay with
indexing time and frequent merges but keep number of segments below 10, for
example?
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.