enable heap dump on OOME, and build the histogram by jhat. Did you try to reduce MaxRamBuffer or max buffered docs? or enable autocommit?
On Tue, Jun 24, 2014 at 7:43 PM, adfel70 <adfe...@gmail.com> wrote: > Hi, > > I am getting OOM during indexing 400 million docs (nested 7-20 children). > The memory usage gets higher while indexing until it gets to 24g. > also after OOM and stop indexing, the memory stays on 24g, *seems like a > leak.* > > > *Solr & Collection Info: * > solr 4.8 , 6 shards, 1 replicas per shard, 24g for jvm > > Thanks > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/OOM-during-indexing-nested-docs-tp4143722.html > Sent from the Solr - User mailing list archive at Nabble.com. > -- Sincerely yours Mikhail Khludnev Principal Engineer, Grid Dynamics <http://www.griddynamics.com> <mkhlud...@griddynamics.com>