Random disconnecting issues


(Curtis Ruck) #1

I've got a pure java logstash-like client using the
elasticsearch-0.90.x.jar. I'm building my client like "new
TransportClient().addTransportAddress(new
InetSocketTransportAddress("localhost", 9300));". Randomly at some point I
end up with either a NodeDisconnectedException or
a ReceiveTimeoutTransportException. Is there another more reliable way to
bulk upload data for long lived processes? I'm dealing with logs, and 5-10
million events a day. I was thinking that a River plugin might solve this,
but i've been trying to keep complexity down and don't want to add yet
another component for temporary queuing into ES (as a river plugin would
require another service to do transient data storage). Additionally, my
custom logstash-like client already has threading and persistent queuing
built-in so adding something external would increase latency and complexity.

--
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.


(Alexander Reelsen) #2

Hey,

are there any logfiles, which could help to debug that issue? Or
stacktraces? I have never really had any problems with a TransportClient
(it should even reconnect, so something special is happening here).

--Alex

On Tue, Nov 5, 2013 at 2:55 PM, Curtis Ruck curtis.ruck@gmail.com wrote:

I've got a pure java logstash-like client using the
elasticsearch-0.90.x.jar. I'm building my client like "new
TransportClient().addTransportAddress(new
InetSocketTransportAddress("localhost", 9300));". Randomly at some point I
end up with either a NodeDisconnectedException or
a ReceiveTimeoutTransportException. Is there another more reliable way to
bulk upload data for long lived processes? I'm dealing with logs, and 5-10
million events a day. I was thinking that a River plugin might solve this,
but i've been trying to keep complexity down and don't want to add yet
another component for temporary queuing into ES (as a river plugin would
require another service to do transient data storage). Additionally, my
custom logstash-like client already has threading and persistent queuing
built-in so adding something external would increase latency and complexity.

--
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.


(system) #3