Deploy and Configure a Multi-Node Elasticsearch Cluster

You are a system administrator who has been asked to deploy a 6-node Elasticsearch cluster with very specific configuration requirements:

• You will need to create an elastic user and then install Elasticsearch version 7.2.1 from an archive at /home/elastic/elasticsearch.
• Each Elasticsearch instance will need to listen on both the local and site-local addresses.
• Configure each node as outlined in the table below.

+----------+-----------+-------------------+--------------+----------+
| Server   | Node Name | Attributes        | Roles        | JVM Heap |
+----------+-----------+-------------------+--------------+----------+
| master-1 | master-1  | zone=1            | master       | 768m     |
+----------+-----------+-------------------+--------------+----------+
| master-2 | master-2  | zone=2            | master       | 768m     |
+----------+-----------+-------------------+--------------+----------+
| master-3 | master-3  | zone=3            | master       | 768m     |
+----------+-----------+-------------------+--------------+----------+
| data-1   | data-1    | zone=1, temp=hot  | data, ingest | 2g       |
+----------+-----------+-------------------+--------------+----------+
| data-2   | data-2    | zone=2, temp=hot  | data, ingest | 2g       |
+----------+-----------+-------------------+--------------+----------+
| data-3   | data-3    | zone=3, temp=warm | data         | 2g       |
+----------+-----------+-------------------+--------------+----------+

Don’t forget the prerequisites for Elasticsearch. Specifically, you will need to persistently increase the max open file limit (nofile) for the elastic user to 65536. Lastly, you will need to persistently increase the max map count for processes (vm.max_map_count) to 262144 via sysctl.