Possibly a pointer in a wrong direction, but: what's your commit strategy? Is it possible that Solr is doing hard commits too often and that is holding up the threads. You could switch to soft-commits with time-based hard commits and see if that helps.
Regards, Alex. Personal blog: http://blog.outerthoughts.com/ LinkedIn: http://www.linkedin.com/in/alexandrerafalovitch - Time is the quality of nature that keeps events from happening all at once. Lately, it doesn't seem to be working. (Anonymous - via GTD book) On Fri, May 31, 2013 at 3:00 PM, ltenny <lte...@gmail.com> wrote: > I have a 10 node, 4.3 SolrCloud with 5 shards. It's a heavy write (100x more > writes than reads) environment. After less than 10k docs in 20 minutes or > so I get tons of EofExceptions from what appears to be the synchronization > traffic between the nodes. When I do queries during this exception storm I > get inconsistent number of rows for a simple *:* query. Seems to suggest > things are not synchronizing. I've changed a handful of RHEL sysctl > parameters like net.core.somaxconn=4096, net.ipv4.tcp_max_syn_backlog=8192 > with no luck. This sure seems like something basic I'm missing. > > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Getting-tons-of-EofException-with-jetty-SolrCloud-tp4067427.html > Sent from the Solr - User mailing list archive at Nabble.com.