Hi,

the hierarchy cache in the search index may use quite some memory depending on 
the number of nodes you have in the repository. Increasing the heap memory 
should help. You can also disable initialization of the hierarchy cache on 
startup in the configuration of your workspace. See initializeHierarchyCache 
parameter here: http://wiki.apache.org/jackrabbit/Search

Regards
 Marcel

On 04.02.2013, at 00:57, "virendra Manvar" 
<[email protected]<mailto:[email protected]>> wrote:

Hi,



We have repository with around 300GB and 16 million docs.



Jackrabbit DMS server is deployed on solaris sparc OS with 64 bits. We are
using wrapper to start DMS server.



When we start up server, it is doing Full GC after 5 minutes and then it
gets hung.



Also we are using *BundleDbPersistenceManager* for storing bundle in DB
which has more than 12 million records.



We have analyzed memory dump with Eclipse Memory Analyzer and it displays
that below class is using high memory.



*
org.apache.jackrabbit.core.query.lucene.CachingIndexReader.initializeParents
*



Please help to resolve this issue.

* *

Regards,

Viren

Reply via email to