[ 
https://issues.apache.org/jira/browse/SOLR-9311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15380809#comment-15380809
 ] 

Junfeng Mu edited comment on SOLR-9311 at 7/16/16 2:50 PM:
-----------------------------------------------------------

and we find that when we shutdown the server (the caller requestor) , the 
connections on http port(8180 for example) closed  immediately. the solrj code 
of CloudSolrServer as below:

    public synchronized static CloudSolrServer getSolrCloudReadServer() {
        if (reviewSolrCloudReadServer == null) {
            ModifiableSolrParams params = new ModifiableSolrParams();
            params.set(HttpClientUtil.PROP_MAX_CONNECTIONS, 1000);
            params.set(HttpClientUtil.PROP_MAX_CONNECTIONS_PER_HOST, 100);
            HttpClient client = HttpClientUtil.createClient(params);            
            
            LBHttpSolrServer lbServer = new LBHttpSolrServer(client);
            
lbServer.setConnectionTimeout(ReviewProperties.getCloudConnectionTimeOut());
            lbServer.setSoTimeout(ReviewProperties.getCloudSoTimeOut());
            reviewSolrCloudReadServer = new 
CloudSolrServer(ReviewProperties.getZkHost(),lbServer);
            
reviewSolrCloudReadServer.setDefaultCollection(ReviewProperties.getZkReviewConnection());
            
reviewSolrCloudReadServer.setZkClientTimeout(ReviewProperties.getZkClientTimeout());
            
reviewSolrCloudReadServer.setZkConnectTimeout(ReviewProperties.getZkConnectTimeout());
        }
        return reviewSolrCloudReadServer;
    }

besides, the zookeeper version is 3.4.6.

Thanks.
Kent


was (Author: kent_mu):
and we find that when we shutdown the server (the caller requestor) , the 
connections on http port(8180 for example) closed  immediately. the solrj code 
of CloudSolrServer as below:

    public synchronized static CloudSolrServer getSolrCloudReadServer() {
        if (reviewSolrCloudReadServer == null) {
            ModifiableSolrParams params = new ModifiableSolrParams();
            params.set(HttpClientUtil.PROP_MAX_CONNECTIONS, 1000);
            params.set(HttpClientUtil.PROP_MAX_CONNECTIONS_PER_HOST, 100);
            HttpClient client = HttpClientUtil.createClient(params);            
            
            LBHttpSolrServer lbServer = new LBHttpSolrServer(client);
            
lbServer.setConnectionTimeout(ReviewProperties.getCloudConnectionTimeOut());
            lbServer.setSoTimeout(ReviewProperties.getCloudSoTimeOut());
            reviewSolrCloudReadServer = new 
CloudSolrServer(ReviewProperties.getZkHost(),lbServer);
            
reviewSolrCloudReadServer.setDefaultCollection(ReviewProperties.getZkReviewConnection());
            
reviewSolrCloudReadServer.setZkClientTimeout(ReviewProperties.getZkClientTimeout());
            
reviewSolrCloudReadServer.setZkConnectTimeout(ReviewProperties.getZkConnectTimeout());
        }
        return reviewSolrCloudReadServer;
    }

besides, the zookeeper version is 3.4.6.

> solrcloud so many connections
> -----------------------------
>
>                 Key: SOLR-9311
>                 URL: https://issues.apache.org/jira/browse/SOLR-9311
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: clients - java
>    Affects Versions: 4.9
>            Reporter: Junfeng Mu
>         Attachments: CPU.png, connections.png, shards.png
>
>
> firstly, I do not know wether it is a bug or a wrong usage. if this is a 
> wrong usage, I will opligize to disturb you, and tell me what the wrong usage 
> is.
> We are using Solrj 4.9.1 to connect to a Zookeeper. and the solr server 
> version is 4.9.0 We are currently using CloudSolrServer as a singleton to 
> operate index data, and I believe that solrj to zookeeper is a TCP 
> connection, and zookeeper to solrcloud internal is actually a httpconnection.
> we use the zabbix to monitor the solrcloud status, and we deploy solr in 
> Wildfly(JBOSS), for example the http port is 8180, we find the number that 
> connecting with solr on port 8180 is so high. for now  we find the number can 
> be around 4000, that is too large.and we find that with the increasing 
> connections, the query speed become slow. the CPU of the solr sever is 
> unstable, and the speed to commit new index data becomes slow as well.
> besides, the JDK version is 1.7.25 to running solr server.
> on the other hand, we have 3 cores with 5 shards, and each shard with one 
> leader and one replication. now the data account goes to be 100 million, and 
> it is still growing up.
> please see the screenshot of Zabbix in the attachments.
> please help me, and looking forward to your reply.
> Thanks.
> Kent



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to