Yes if you have high ingest volumes / rate and multiple data nodes setting more than 1 primary shard will parallelize this input which can allow higher ingestion throughput.
This makes most sense when you have multiple data nodes with the primary shards distributed across the nodes.
So if you set to 2 primary shards... the your index will rollover at a total of 60GB = 2 x 30GB primary shards.
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.