In our current Env we are running a Single Node Elasticsearch so I set up One Shard and Cero Replica for all my indexes. But there are a suggestions to increment the number of the shards for using more the CPU usage. More exactly this question when I put only one shard and cero replica:
Won't this decrease the overall performance of the index since it will only be able to use a single core?
I read in some question that Shards and CPU core doesn't have a direct relationship and in a single node will node will not improve nothing if we increment the number of shards per Index.
It depends on the use case. If you have a search heavy use case searches will run on a number of threads and query the shard in parallel. If you have few concurrent queries you may benefit from having more shards as this allows more CPU cores to help out. If you have an index heavy use case the anser may differ further.
As a Single Node it will be more Heavy Index. So I know adding more shards Improve your Query Performance. But I want to save like 2000 Json Docs Per second. I managed to have that heavy index. So Sharding affects indexing performance but improves Searchs. Am I correct ?
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.