I have started an instance of elasticsearch in one machine (A) and entered
all my data.
Next I have started another elasticsearch on another machine (B).
Problem: As long as elasticsearch on A is on, search in B is happenning but
if I close A, search on B stops.
I thought may be this is happening to avoid split brain problem.
So I started another elasticsearch instance (C) and set the number of
master nodes to two.
But if A is closed search is not happening in either C or B.
Note: Data has been entered in instance A only when B or C instances were
not on.
As an user I want to have a single cluster ip, I need not to worry about
three instances A,B or C. Also search must go on as long as any one of A,B
or C is on.
Kindly let me know about the steps to follow.
Also any other resources, where there is a step by step guide to follow ?
I would would first revert all settings back to the default. Then bring up
all three ES nodes and check that they all can see each other and form a
cluster. If you run this on one of the nodes, you should be able to see how
the cluster looks like:
If they cannot all see each other, you have to first find out why - is it
possible that your network does not support multicast? If so, then you need
to go into the elasticsearch.yml file of all three nodes and turn off
multicast, and turn on unicast discovery and put a seed list of nodes in
there (i.e. IP addresses of 2 or 3 of your cluster nodes). Then stop them
all and start them all again. Then check again if they all form a cluster.
Once you have verified that they can all see each other, then you can start
with your experimentation of putting data in, taking nodes down, seeing the
behavior, etc.
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.