I use elasticsearch for some time, but I still have ideas that are not too
Clere for me, and I would like to have a Response:
if we have nodes running on different machines, how one will know the
other,how they will communicate and the notion of scalability how it will
be respected
There is something called "multi casting". ES uses that to see if there are
other es instances in the network with the same cluster name.
If there is , it will join that es instance and form a cluster.
Now about scalability - Say if there are 2 instances of ES in network and
they have joined to form a cluster , when you give 1000 documents to ES to
index , it will store 500 documents in one instance and 500 documents in
another instance. Here when a query is made , the query is executed on both
the instances and result is aggregated to form the final result. Hence just
by increasing the number of instances in network , ES can work as a
distributed application and divide the load between all the nodes.
Going ahead , look into what is shards and replicas. That should give you a
complete idea.
I use elasticsearch for some time, but I still have ideas that are not too
Clere for me, and I would like to have a Response:
if we have nodes running on different machines, how one will know the
other,how they will communicate and the notion of scalability how it will
be respected
Le mercredi 6 mars 2013 18:06:31 UTC+1, Ammar Yahia a écrit :
Hi all,
I use elasticsearch for some time, but I still have ideas that are not too
Clere for me, and I would like to have a Response:
if we have nodes running on different machines, how one will know the
other,how they will communicate and the notion of scalability how it will
be respected
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.