If you are actively indexing and opening new near-real-time readers, the number of segments in your index will increase, which means the number of open input files (corresponding to instances of ByteBufferIndexInput.SingleBufferImpl) will be created. So it's expected its heap usage grows, but the growth should taper as Lucene merges small segments into larger ones ...
Mike McCandless http://blog.mikemccandless.com On Fri, Dec 2, 2016 at 7:04 AM, Mukul Ranjan <mran...@egain.com> wrote: > Hi, > > We ran longevity Load testing run for 96 hour in our application using lucene > 5.5.2 for text search. We have observed that there is significant change in > heap size of > org.apache.lucene.store.ByteBufferIndexInput$SingleBufferImpl<mat://object/0x740b63a68>. > > Size of this class increased from 7 MB to 15 MB from day1 to day4 for this > class. We were adding some document in indexing in parallel. But we do not > excepted to increase in size of this class that much. > > Please suggest if this is expected behaviour or we are missing something in > our code. > > Thanks, > Mukul Ranjan > > ----- > Watch an eGain Try+Buy(tm) customer success > video<https://www.youtube.com/watch?v=ewEeyeqJRCE&feature=youtu.be/> --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
