Regarding elasticsearch load balancer

Hi All,

i know a concept of load balancer in elastic search which is HTTP 

enabled and never be a master and doesn't hold any data. my doubt is if we
introduce client node called as loadbalancer in cluster is there any need
to setup haproxy for the cluster to forward request.if there is no need of
haproxy please let me know how to achieve fail over using load balancer.

suppose in my cluster i don't have load balancer nodes how can we manage
request forward settings . let say node 1 is down but we configured node1
in client application.users will hit node1 from client application but
node1 is down in this case how can we forward request to client with out
haproxy?

Thanks
phani.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/d0043350-001e-46a2-9800-ee9b0328aae5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

It depends on the language/platform you use.

If Java, all is very easy - NodeClient connects to more than one node, and
also TransportClient "sniff" mode. So Java clients are using fault tolerant
connection mode.

Same do the official clients, but only if configured properly. Please study
how to connect official ES clients to more than one node of a cluster.

"Load balancing" can be achieved by adding more connections to the client.
Since each node knows the cluster state and can find out how to route
requests to the involved nodes, this is not always required. This is an
internal ES mechanism.

No haproxy needed. Even if haproxy is used, haproxied requests are
balanced/routed once more in the ES internals. So you will end up in
"double load balancing".

Jörg

On Tue, Jan 27, 2015 at 1:03 PM, phani.nadiminti@goktree.com wrote:

Hi All,

i know a concept of load balancer in elastic search which is HTTP

enabled and never be a master and doesn't hold any data. my doubt is if we
introduce client node called as loadbalancer in cluster is there any need
to setup haproxy for the cluster to forward request.if there is no need of
haproxy please let me know how to achieve fail over using load balancer.

suppose in my cluster i don't have load balancer nodes how can we
manage request forward settings . let say node 1 is down but we configured
node1 in client application.users will hit node1 from client application
but node1 is down in this case how can we forward request to client with
out haproxy?

Thanks
phani.

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/d0043350-001e-46a2-9800-ee9b0328aae5%40googlegroups.com
https://groups.google.com/d/msgid/elasticsearch/d0043350-001e-46a2-9800-ee9b0328aae5%40googlegroups.com?utm_medium=email&utm_source=footer
.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAKdsXoHLeOhDfE1p2585Eiib6W3doTP12aWKwBHioAhTMZOPAA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Hi Jorg

Thank you for the quick reply. let say i am establishing small cluster i
don't have client node on it in this case can i use haproxy to forward
requests between servers.

Thanks
phani

On Tuesday, January 27, 2015 at 5:33:34 PM UTC+5:30, phani.n...@goktree.com
wrote:

Hi All,

i know a concept of load balancer in elastic search which is HTTP 

enabled and never be a master and doesn't hold any data. my doubt is if we
introduce client node called as loadbalancer in cluster is there any need
to setup haproxy for the cluster to forward request.if there is no need of
haproxy please let me know how to achieve fail over using load balancer.

suppose in my cluster i don't have load balancer nodes how can we
manage request forward settings . let say node 1 is down but we configured
node1 in client application.users will hit node1 from client application
but node1 is down in this case how can we forward request to client with
out haproxy?

Thanks
phani.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/004fc17e-e151-4cea-adad-1018072f0f9f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.