lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Oded Sofer <odedso...@yahoo.com.INVALID>
Subject Re: 1:M connectivity
Date Thu, 16 Apr 2015 20:27:05 GMT
Right, we are using that. 
The issue is the firewall setting needed for the cloud. We do not want to open all nodes to
all others nodes. However, we found that add-index to a specific node tries to access all
other nodes though we set it to index locally on that node only. 


On Apr 16, 2015 7:19 PM, Erick Erickson <erickerickson@gmail.com> wrote:
>
> You say "the SolrCloud API". Not entirely sure what that is, do you 
> mean the post.jar tool? 
>
> Because to get much more scalable throughput, you probably want to use SolrJ and 
> the CloudSolrServer class. That class takes a connection to Zookeeper and 
> "does the right thing". 
>
> Best, 
> Erick 
>
> On Thu, Apr 16, 2015 at 7:19 AM, Oded Sofer <odedsofer@yahoo.com.invalid> wrote:

> > Given that the index are always routed to specific machine, is there a way to avoid
connectivity from the node to all other node. 
> > We are using Solr 4.10; the Add/Update Index uses SolrCloud API and always added
to the node that get API request for add-index (i.e., we are sending the add index to the
appropriate node that should get it). 
> > 
> > 
Mime
View raw message