Here is my scenario:
I am using Solr for my application where it will hit the solr servers with input query and get the results. I have 2 solr nodes, so i am using solr load balancer LBHttpSolrServer. Something like this,
LBHttpSolrServer server1 = new LBHttpSolrServer("http://ift.tt/1ETbwkE","http://ift.tt/1DZ64Pr");
When this code is executed from simple java "public static void main" method (i.e. as standalone program), the load balancing happens fine. that is every other request is alternating between endpoint A and B.
However, when the same code is wrapped inside Mapper code and deployed in AWS EMR set up, load balancing is not happening, we confirmed by observing the Server logs, all requests from mapper code is always reaching first server in the connect string (i.e. it always reaches "http://ift.tt/1ETbwkE" as per above string ).
Some More details,
Solr version: 4.10.3
Load Balancer Class used: org.apache.solr.client.solrj.impl.LBHttpSolrServer
EMR AMI version: 2.4.9
Hadoop Distribution: Amazon 1.0.3
anyone of you guys please help me out on this. Thanks much..
Aucun commentaire:
Enregistrer un commentaire