Hi Michael, I was indexing documents about 10,000 or more, but the error occurred shortly at early stage.
The physical memory is 2Gb on the machine, and at the time of the OOM occurred is about 768MB, I'm pretty sure that memory was enough for system operated. When indexing was running, there were no other search operations going. I googled for a while, and found some information( http://issues.apache.org/jira/browse/LUCENE-1191 see the first comment), things about corrupted index segment file. I quickly rebuilt index files from database records, and tried to do index again. The OOM situation disappear and never occurr again. I guess there must be something wrong when indexing. Perhaps the error message could be adjusted to tell the root problem is corrupted index files not OOM insteaed. FYI, Thanks Floyd 2010/3/10 Michael Garski <[email protected]> > Floyd, > > Can you provide more details? Things such as how many documents you > were indexing, how much memory was being consumed at the time of the > OOM, whether you were indexing and searching concurrently, and whether > you are using sorts and filters are a good place to start. > > Michael > > -----Original Message----- > From: Floyd Wu [mailto:[email protected]] > Sent: Tuesday, March 09, 2010 12:46 AM > To: [email protected] > Subject: The Writer hit an OutOfMemory Error > > Hi all, > > I just received a message "The writer hit an OutOfMemory Error ..." this > afternooon while indexing some documets. > > I searchd the Internet and seem this *bug* had been solved in Java > Lucene > 2.3. > Currently I use Lucne.Net-2.9.1, this issue probably has been solved in > this > release.(am I right?) > If not yet been solved, how to prevent this problem and what's the > reason > for? > If has been solved, why? > > Thanks in advanced > > Floyd > >
