On 12/16/2013 2:34 AM, Torben Greulich wrote: > we get a OutOfMemoryError in RamUsageEstimator and are a little bit > confused about the error. > We are using solr 4.6 and are confused about the Lucene42DocValuesProducer. > We checked current solr code and found that Lucene42NormsFormat will be > returned as NormFormat in Lucene46Codec and so the Lucene42DocValuesProducer > will be returned. Is there any special reason not to have a higher version > of NormsFormat?
The format for that part of the index apparently hasn't changed since Lucene 4.2. Something else must have changed about the default index format, or I expect that there would not be a 4.6 specific Codec object. From what I've seen, the verion number encoded in the class name doesn't change until there's a new class needed for the default codec. A java OutOfMemoryError means that you need to increase the max heap when you start the program. It looks like you're trying to get a core status, so this is probably happening when using the admin UI. To put it quite simply, your Java max heap is not big enough to handle what Solr has been asked to do. http://wiki.apache.org/solr/SolrPerformanceProblems#Java_Heap Thanks, Shawn