I have like 5 elastic nodes in total. One node ingest and coordinating node while two are dedicated master nodes and the other two are dedicated data nodes.
I have elasticsearch install on all my five nodes with respective settings.
What i am confused about and want to know is.
Does i need to include all the elasticsearch nodes ip addresses in filebeat.yml file
which is install on seperate nodes output.elasticsearch.hosts: ['https://x.x.x.x', '' '',"" ]
or i should add only the specific type of nodes ip:port in this list. Actually i am really confuse and don't know much about this. Filebeat will be inserting data into elasticsearch so can anyone guide on this.
Having two dedicated master nodes is bad - if you have 3 or more nodes in the cluster you should ALWAYS have at least 3 master eligible nodes. Having dedicated node types is also optional and might be useful if you are deploying across different types of hosts. For most small clusters it however often make sense to not use dedicated node types and instead let all nodes have all roles, especially if they are deployed on uniform hardware.
yes, i do have three eligible master nodes one of my data node is also master eligible. Right now i am in testing stage the cluster is going to expand as there will be a lot of data. What i'm concerned right now which nodes i need to add in filebeat.yml does all of them?
Dedicated master nodes are generally smaller than other node types and shouldbe left to manage the cluster and not receive any requests. If you are using ingest pipelines it makes sense to direct indexing requests to your ingest nodes. Queries should typically go to dedicated coordinating-only nodes (if you have them) or directly to the data nodes.
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.