On Thu, Mar 17, 2011 at 5:50 PM, Geeta Subramanian <gsubraman...@commvault.com> wrote: > Here is the attached xml. > In our xml, maxBufferedDocs is commented. I hope that's not causing any issue. > The ramBufferSizeMB is 32Mb, will changing this be of any use to me?
Nope... your index settings are fine. Perhaps something in extracting request handler or tika is holding onto memory. Has anyone else experienced/reproduced this? Geeta, can you open a JIRA issue? If you're actually giving the JVM 4G of heap (is this a 64 bit JVM?), this looks like a bug somewhere. -Yonik http://lucidimagination.com