Cluster management

Elastic Stack Elasticsearch
1

Hi, new to ElasticSearch and this may be a noob question. When a node in a
cluster goes down, how does the client realize that it's not supposed to
hit that IP or more importantly that port in case of more than one nodes on
a server? Seems to me like the clients need to maintain metadata about the
cluster and loop through that if the request fails.

Thanks
Chethan

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

2

Hello Chethan,

Yes, I think you'll need to loop from your application, or use a load
balancer. Although, if the load balancer goes down, you're back to square 1.

Some clients (pyes, for example) accept an array of addresses, and handle
that for you.

Best regards,
Radu

http://sematext.com/ -- Elasticsearch -- Solr -- Lucene

On Tue, May 14, 2013 at 2:35 PM, Chethan Prasad chethan.p.04@gmail.comwrote:

Hi, new to Elasticsearch and this may be a noob question. When a node in a
cluster goes down, how does the client realize that it's not supposed to
hit that IP or more importantly that port in case of more than one nodes on
a server? Seems to me like the clients need to maintain metadata about the
cluster and loop through that if the request fails.

Thanks
Chethan

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

3

Hello Radu,

Thanks for the response. Although I understand adding something like a name
node is not all that wise because then it brings in a single point of
failure. But I think techniques used by Cassandra can be adopted here. It
may be a good idea for a fork.

Thanks
Chethan

On Tue, May 14, 2013 at 5:38 PM, Radu Gheorghe
radu.gheorghe@sematext.comwrote:

Hello Chethan,

Yes, I think you'll need to loop from your application, or use a load
balancer. Although, if the load balancer goes down, you're back to square 1.

Some clients (pyes, for example) accept an array of addresses, and handle
that for you.

Best regards,
Radu

http://sematext.com/ -- Elasticsearch -- Solr -- Lucene

On Tue, May 14, 2013 at 2:35 PM, Chethan Prasad chethan.p.04@gmail.comwrote:

Hi, new to Elasticsearch and this may be a noob question. When a node in
a cluster goes down, how does the client realize that it's not supposed to
hit that IP or more importantly that port in case of more than one nodes on
a server? Seems to me like the clients need to maintain metadata about the
cluster and loop through that if the request fails.

Thanks
Chethan

--
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.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the
Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit
https://groups.google.com/d/topic/elasticsearch/ZKg3fbm_Z2Q/unsubscribe?hl=en-US
.
To unsubscribe from this group and all its topics, send an email to
elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

4

Hi Chethan

What client are you talking about? The elasticsearch Java node client? The
transport client? One of the REST clients?

The Java clients can sniff a list of live nodes and so will remove dead
nodes. (client.transport.sniff: true)

Some of the REST clients do this as well and there are plans afoot to
support this in all REST clients

clint

On Tue, May 14, 2013 at 2:24 PM, chethan chethan.p.04@gmail.com wrote:

Hello Radu,

Thanks for the response. Although I understand adding something like a
name node is not all that wise because then it brings in a single point of
failure. But I think techniques used by Cassandra can be adopted here. It
may be a good idea for a fork.

Thanks
Chethan

On Tue, May 14, 2013 at 5:38 PM, Radu Gheorghe <radu.gheorghe@sematext.com

wrote:

Hello Chethan,

Yes, I think you'll need to loop from your application, or use a load
balancer. Although, if the load balancer goes down, you're back to square 1.

Some clients (pyes, for example) accept an array of addresses, and handle
that for you.

Best regards,
Radu

http://sematext.com/ -- Elasticsearch -- Solr -- Lucene

On Tue, May 14, 2013 at 2:35 PM, Chethan Prasad chethan.p.04@gmail.comwrote:

Hi, new to Elasticsearch and this may be a noob question. When a node in
a cluster goes down, how does the client realize that it's not supposed to
hit that IP or more importantly that port in case of more than one nodes on
a server? Seems to me like the clients need to maintain metadata about the
cluster and loop through that if the request fails.

Thanks
Chethan

--
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.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the
Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit
https://groups.google.com/d/topic/elasticsearch/ZKg3fbm_Z2Q/unsubscribe?hl=en-US
.
To unsubscribe from this group and all its topics, send an email to
elasticsearch+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

5

Hi,

I was talking about just using REST.

Thanks
Chethan

On Wed, May 15, 2013 at 2:53 PM, Clinton Gormley clint@traveljury.comwrote:

Hi Chethan

What client are you talking about? The elasticsearch Java node client? The
transport client? One of the REST clients?

The Java clients can sniff a list of live nodes and so will remove dead
nodes. (client.transport.sniff: true)

Some of the REST clients do this as well and there are plans afoot to
support this in all REST clients

clint

On Tue, May 14, 2013 at 2:24 PM, chethan chethan.p.04@gmail.com wrote:

Hello Radu,

Thanks for the response. Although I understand adding something like a
name node is not all that wise because then it brings in a single point of
failure. But I think techniques used by Cassandra can be adopted here. It
may be a good idea for a fork.

Thanks
Chethan

On Tue, May 14, 2013 at 5:38 PM, Radu Gheorghe <
radu.gheorghe@sematext.com> wrote:

Hello Chethan,

Yes, I think you'll need to loop from your application, or use a load
balancer. Although, if the load balancer goes down, you're back to square 1.

Some clients (pyes, for example) accept an array of addresses, and
handle that for you.

Best regards,
Radu

http://sematext.com/ -- Elasticsearch -- Solr -- Lucene

On Tue, May 14, 2013 at 2:35 PM, Chethan Prasad chethan.p.04@gmail.comwrote:

Hi, new to Elasticsearch and this may be a noob question. When a node
in a cluster goes down, how does the client realize that it's not supposed
to hit that IP or more importantly that port in case of more than one nodes
on a server? Seems to me like the clients need to maintain metadata about
the cluster and loop through that if the request fails.

Thanks
Chethan

--
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.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the
Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit
https://groups.google.com/d/topic/elasticsearch/ZKg3fbm_Z2Q/unsubscribe?hl=en-US
.
To unsubscribe from this group and all its topics, send an email to
elasticsearch+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the
Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit
https://groups.google.com/d/topic/elasticsearch/ZKg3fbm_Z2Q/unsubscribe?hl=en-US
.
To unsubscribe from this group and all its topics, send an email to
elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

6

I add to Clint's answer the following link that should help you as well: Elasticsearch Platform — Find real-time answers at scale | Elastic

--
David Pilato | Technical Advocate | Elasticsearch.com
@dadoonet | @elasticsearchfr | @scrutmydocs

Le 15 mai 2013 à 11:23, Clinton Gormley clint@traveljury.com a écrit :

Hi Chethan

What client are you talking about? The elasticsearch Java node client? The transport client? One of the REST clients?

The Java clients can sniff a list of live nodes and so will remove dead nodes. (client.transport.sniff: true)

Some of the REST clients do this as well and there are plans afoot to support this in all REST clients

clint

On Tue, May 14, 2013 at 2:24 PM, chethan chethan.p.04@gmail.com wrote:
Hello Radu,

Thanks for the response. Although I understand adding something like a name node is not all that wise because then it brings in a single point of failure. But I think techniques used by Cassandra can be adopted here. It may be a good idea for a fork.

Thanks
Chethan

On Tue, May 14, 2013 at 5:38 PM, Radu Gheorghe radu.gheorghe@sematext.com wrote:
Hello Chethan,

Yes, I think you'll need to loop from your application, or use a load balancer. Although, if the load balancer goes down, you're back to square 1.

Some clients (pyes, for example) accept an array of addresses, and handle that for you.

Best regards,
Radu

http://sematext.com/ -- Elasticsearch -- Solr -- Lucene

On Tue, May 14, 2013 at 2:35 PM, Chethan Prasad chethan.p.04@gmail.com wrote:
Hi, new to Elasticsearch and this may be a noob question. When a node in a cluster goes down, how does the client realize that it's not supposed to hit that IP or more importantly that port in case of more than one nodes on a server? Seems to me like the clients need to maintain metadata about the cluster and loop through that if the request fails.

Thanks
Chethan

--
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.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/ZKg3fbm_Z2Q/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to elasticsearch+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.