Hello, Andi. AV> There is a thread currently on the Lucene user list titled 'Opening up one AV> large index takes 940M or memory' that may bring some answers to your problem.
they tune memory usage by increasing TermInfosWriter.indexInterval. Can we do this in PyLucene? AV> There may also be ways to increase the amount of memory gcj gets to use, I AV> would look on the [EMAIL PROTECTED] mailing list archives for this. I suspect that gcj has some kind of memory limitation... AV> Andi.. AV> On Tue, 15 Feb 2005, Yura Smolsky wrote: >> Hello, Andi. >> >> I have index of 51Gb. When I try to create IndexSearcher with >> FSDirectory of this index, I receive exception: >> >> GC Warning: Out of Memory! Returning NIL! >> Traceback (most recent call last): >> File "./RemoteSearcherServer.py", line 275, in ? >> server.processCommandLine(sys.argv) >> File "/home/search/lib/myCORBA.py", line 81, in processCommandLine >> self.start() >> File "/home/search/lib/myCORBA.py", line 51, in start >> self.setUp() >> File "./RemoteSearcherServer.py", line 240, in setUp >> self.objServant = RemoteSearcher_i(num, path, luceneSearcher) >> File "./RemoteSearcherServer.py", line 132, in __init__ >> luceneSearcher._reloadIndex() >> File "./RemoteSearcherServer.py", line 18, in _reloadIndex >> self.searcher = IndexSearcher(self.directory) >> File "/usr/lib/python2.4/site-packages/PyLucene.py", line 2266, in __init__ >> newobj = _PyLucene.new_IndexSearcher(*args) >> ValueError: java.lang.OutOfMemoryError >> <<No stacktrace available>> >> >> Why? >> >> Yura Smolsky. >> >> >> Yura Smolsky. _______________________________________________ pylucene-dev mailing list [email protected] http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
