Sometimes happens when your zookeeper timeout is short (the 15 second timeout has been too short) and you hit a GC pause that is longer. Try lengthening your ZK timeout to 45 or 60 seconds.
It would help if you told us about your setup, memory allocated to Solr, all that stuff, see: http://wiki.apache.org/solr/UsingMailingLists Best, Erick On Mon, Mar 30, 2015 at 1:43 AM, sthita <sthit...@gmail.com> wrote: > Solr server crashes 2-3 times a day . When i looked into log i can see some > recovery failure. > > I got these errors on my solr log > > WARN - 2015-03-30 16:03:51.137; org.apache.solr.cloud.RecoveryStrategy; > Stopping recovery for zkNodeName=searcher:8980_solr_rn0core=rn0 > ERROR - 2015-03-30 16:03:56.121; org.apache.solr.update.UpdateLog; Error > inspecting tlog tlog{file=/solr/rn0/data/tlog/tlog.0000000000000000709 > refcount=2} > ERROR - 2015-03-30 16:03:56.121; org.apache.solr.update.UpdateLog; Error > inspecting tlog tlog{file=/solr/rn0/data/tlog/tlog.0000000000000000710 > refcount=2} > WARN - 2015-03-30 16:04:00.242; org.apache.solr.core.SolrCore; [rn0] > PERFORMANCE WARNING: Overlapping onDeckSearchers=2 > > I googled so many places, suggestions are related to my commit . > > My commit configuration on solrconfig.xml : > > > <updateHandler class="solr.DirectUpdateHandler2"> > > <autoCommit> > <maxTime>600000</maxTime> > <openSearcher>false</openSearcher> > </autoCommit> > > <updateLog> > <str name="dir">${solr.data.dir:}</str> > </updateLog> > </updateHandler> > > > Why is it happening ? Any idea/solution will be appreciated. > > > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Solr-went-on-recovery-multiple-time-tp4196249.html > Sent from the Solr - User mailing list archive at Nabble.com.
