I'm having a search engine for source code of different repositories. In
which the cluster consist of about 500 indices and about 1.8 million
documents with its full content.
And am facing slowness issue while searching the content, its taking a
little extra time to display the results.
Those indices in the cluster were imported from the other cluster, if so
will the metadata of those indices will cause any difference in search
performance? Previously it was working fine, but after importing indices to
another cluster and searching in the new cluster makes the performance
degradation.
And also addition to this, when I import the indices to another cluster,
all the cluster wont be available in the newer cluster. What should be done
to make all the indices available over there?
On Monday, 17 November 2014 15:00:43 UTC+5:30, Anand kumar wrote:
I'm having a search engine for source code of different repositories. In
which the cluster consist of about 500 indices and about 1.8 million
documents with its full content.
And am facing slowness issue while searching the content, its taking a
little extra time to display the results.
Those indices in the cluster were imported from the other cluster, if so
will the metadata of those indices will cause any difference in search
performance? Previously it was working fine, but after importing indices to
another cluster and searching in the new cluster makes the performance
degradation.
There are so many moving factors, it is hard to help. Do both clusters have
exactly the same hardware, network infrastructure, number of nodes, ... ?
You can potentially try to capture where your nodes are spending time
during slow queries with the hot_threads API.
And also addition to this, when I import the indices to another cluster,
all the cluster wont be available in the newer cluster. What should be done
to make all the indices available over there?
On Monday, 17 November 2014 15:00:43 UTC+5:30, Anand kumar wrote:
I'm having a search engine for source code of different repositories. In
which the cluster consist of about 500 indices and about 1.8 million
documents with its full content.
And am facing slowness issue while searching the content, its taking a
little extra time to display the results.
Those indices in the cluster were imported from the other cluster, if so
will the metadata of those indices will cause any difference in search
performance? Previously it was working fine, but after importing indices to
another cluster and searching in the new cluster makes the performance
degradation.
Thank you so much for your reply. Yes both the clusters were having
the same configurations. And one more point, I missed out. I've configured
the master node and data node in separate docker container. Did the usage
of docker has anything to do with this slowness issue.
Thanks again in advance.
-Anand
On Tuesday, 18 November 2014 16:59:38 UTC+5:30, Adrien Grand wrote:
There are so many moving factors, it is hard to help. Do both clusters
have exactly the same hardware, network infrastructure, number of nodes,
... ? You can potentially try to capture where your nodes are spending time
during slow queries with the hot_threads API.
On Tue, Nov 18, 2014 at 9:58 AM, Anand kumar <anand...@gmail.com
<javascript:>> wrote:
And also addition to this, when I import the indices to another cluster,
all the cluster wont be available in the newer cluster. What should be done
to make all the indices available over there?
On Monday, 17 November 2014 15:00:43 UTC+5:30, Anand kumar wrote:
I'm having a search engine for source code of different repositories. In
which the cluster consist of about 500 indices and about 1.8 million
documents with its full content.
And am facing slowness issue while searching the content, its taking a
little extra time to display the results.
Those indices in the cluster were imported from the other cluster, if so
will the metadata of those indices will cause any difference in search
performance? Previously it was working fine, but after importing indices to
another cluster and searching in the new cluster makes the performance
degradation.
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.