Transport-memcached gives ClosedChannelException if query takes too long

Hi,

I am using Elasticsearch 0.20.5 with the transport-memcached plugin.

When I try running a long-running search request (in my estimate anything
over 5 seconds or so) using a memcached request, the request fails with the
message [ERROR][memcached.netty ] caught exception on memcached
decoder.

On further investigation, I found that the problem was being caused by a
ClosedChannelException being triggered on a channel.write() command when
trying to process the sendResponse in
org.elasticsearch.memcached.netty.MemcachedRestChannel.sendResponse. It
appears that at this point, the channel has been closed by some other
process.

The exact same request works correctly, if it completes quickly enough.
For example, I am able to get a successful call giving me the results I
want if I make the request a second time, with elasticsearch then getting
the results out of the cache, and thus making the request much quicker.

Has anyone else seen this problem, and does anyone know what I can do to
fix it?

--
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.

Is your issue perhaps this one?

--
Ivan

On Thu, Mar 28, 2013 at 7:07 AM, Serge Meunier sjmeunier@gmail.com wrote:

Hi,

I am using Elasticsearch 0.20.5 with the transport-memcached plugin.

When I try running a long-running search request (in my estimate anything
over 5 seconds or so) using a memcached request, the request fails with the
message [ERROR][memcached.netty ] caught exception on memcached
decoder.

On further investigation, I found that the problem was being caused by a
ClosedChannelException being triggered on a channel.write() command when
trying to process the sendResponse in
org.elasticsearch.memcached.netty.MemcachedRestChannel.sendResponse. It
appears that at this point, the channel has been closed by some other
process.

The exact same request works correctly, if it completes quickly enough.
For example, I am able to get a successful call giving me the results I
want if I make the request a second time, with elasticsearch then getting
the results out of the cache, and thus making the request much quicker.

Has anyone else seen this problem, and does anyone know what I can do to
fix it?

--
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.

Hi Ivan,

I had seen that issue previously, but while it seems to be a similar issue,
I am not sure it is same issue. The issue in the link shows that it has
been resolved in versions 0.20.6 and 0.90.0.RC1. I have since tried running
both of these versions of Elasticsearch, and I get the exact same result,
so I think there is something else causing this.

On Thursday, March 28, 2013 9:42:01 PM UTC+1, Ivan Brusic wrote:

Is your issue perhaps this one?
https://github.com/elasticsearch/elasticsearch/issues/2733

--
Ivan

On Thu, Mar 28, 2013 at 7:07 AM, Serge Meunier <sjme...@gmail.com<javascript:>

wrote:

Hi,

I am using Elasticsearch 0.20.5 with the transport-memcached plugin.

When I try running a long-running search request (in my estimate anything
over 5 seconds or so) using a memcached request, the request fails with the
message [ERROR][memcached.netty ] caught exception on memcached
decoder.

On further investigation, I found that the problem was being caused by a
ClosedChannelException being triggered on a channel.write() command when
trying to process the sendResponse in
org.elasticsearch.memcached.netty.MemcachedRestChannel.sendResponse. It
appears that at this point, the channel has been closed by some other
process.

The exact same request works correctly, if it completes quickly enough.
For example, I am able to get a successful call giving me the results I
want if I make the request a second time, with elasticsearch then getting
the results out of the cache, and thus making the request much quicker.

Has anyone else seen this problem, and does anyone know what I can do to
fix it?

--
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 elasticsearc...@googlegroups.com <javascript:>.
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.