You are viewing a plain text version of this content. The canonical link for it is here.

Posted to solr-user@lucene.apache.org by "Joshi, Shital" <Sh...@gs.com> on 2014/02/05 17:59:51 UTC

Solr4 performance

Hi,

We have SolrCloud cluster (5 shards and 2 replicas) on 10 dynamic compute boxes (cloud). We're using local disk (/local/data) to store solr index files. All hosts have 60GB ram and Solr4 JVM are running with max 30GB heap size. So far we have 470 million documents. We are using custom sharding and all shards have ~9-10 million documents. We have a GUI sending queries to this cloud and GUI has 30 seconds of timeout.

Lately we're getting many timeouts on GUI and upon checking we found that all timeouts are happening on 2 hosts. The admin GUI for one of the hosts show 96% of physical memory but the other host looks perfectly good. Both hosts are for different shards. Would increasing ram of these two hosts make these timeouts go away? What else we can check?

Many Thanks!