On 3/4/2014 2:23 AM, Angel Tchorbadjiiski wrote:
in the last couple of weeks one of my machines is experiencing
OutOfMemoryError: Java heap space errors. In a couple of hours after
starting the SOLR instance queries with execution times of unter 100ms
need more than 10s to execute and many Java heap space erros appear in
the logs. I would be grateful for any hints, how to localize/fix the
problem.
The system is a single shard SOLR 4.6.1 instance with two cores in the
default jetty container. The CORE1 has ~45M documents (disk size of
32GB) with 40 fields (all stored and indexed). CORE2 has ~20M
documents (disk size of 18GB) with 60 fields (both stored and indexed
also). All the fields are relatively short with a maximal length of
100 characters. In both cores 20 of the fields are used for faceting,
have a cache populating query on "newSearcher" event and the following
cache configurations:
It may be your facets that are killing you here. As Toke mentioned, you
have not indicated what your max heap is.20 separate facet fields with
millions of documents will use a lot of fieldcache memory if you use the
standard facet.method, fc.
Try adding facet.method=enum to all your facet queries, or you can put
it in the defaults section of each request handler definition.
Alternatively, you can add docValues to all your facet fields, but that
will require a full reindex.
http://wiki.apache.org/solr/SolrPerformanceProblems#Java_Heap
Thanks,
Shawn
