I am implementing Elastic cluster which will have 3 data node, 1 primary master node & 1 node for both client & secondary master.
while pushing the data to cluster whether i have to specify data nodes or client node?
from my point am using nginx for proxy,security & load balance purpose so in nginx i ll specify 3 data nodes so that nginx will manage the HA of data nodes.
also i want to disable direct http access of data nodes, so user can access client node and admin will manage through master. is it sense better? pls suggest.
Second one,
I want to create 1 primary index & 1 secondary index. the primary index will keep last 30 days data. Documents which are present more than 30 days in primary index should be moved to secondary index based on time stamp, then the secondary index will keep last 6 month documents
How to configure this in Elasticsearch end? because my client servers are pushing the data directly to elastic using Curl cmd through nginx proxy, my custom time stamp format is 2016-08-24:17:21:54.
The 3d master will be any one of data node which will take master role also.
Could you please explain, why should I avoid primary and secondary index with time based concept..
so I can understand from your perspective. bez its isnecessary for my requirement.
Also can I configure elasticsearch to create an index every three month or one month? i can do this using logstash but here am not using logstash.
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.