I have three node cluster and my current shard setting is 2 primary and 1 replica so in total, 4 shards for three node cluster.
In this case, a single node is having one extra shard everytime?
Can i have 1 primary and 2 replica shards? So that each node will have only one shard to have less storage footprint.
Also, it is mentioned in documentation that we can reduce no. of shards for rollover index, is it possible with three node cluster and above current shard setting (2 primary,1 replica) to have less storage footprint.
I am asking these questions to save storage space.
Can someone confirm that there will be no disadvantage of having 1 primary and 2 repliacs on three node cluster?
Am i missing something.
Here the aim is not really performance (paraller read etc.) but to have shards equally spread on each node ( no extra shard allocated but in above case currently its happening).
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.