lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael McCandless <luc...@mikemccandless.com>
Subject Re: MergePolicy$MergeException CorruptIndexException in lucene2.4.1
Date Mon, 30 Nov 2009 13:34:09 GMT
On Mon, Nov 30, 2009 at 7:22 AM, jm <jmuguruza@gmail.com> wrote:
> No other exceptions I could spot.

OK

> OS: win2003 32bits, with NTFS. This is a vm running on vmware fusion on a mac.

That should be fine...

> jvm: I made sure, java version "1.6.0_14"

Good.

> IndexWriter settings:
>        writer.setMaxFieldLength(maxFieldLength);
>        writer.setMergeFactor(10);
>        writer.setRAMBufferSizeMB(10);
>        writer.setUseCompoundFile(true);
>        writer.setTermIndexInterval(512);

All normal.  The 512 terms index interval is interesting though --
that's not a setting most people tweak.  I can't think how it could
lead to what you're seeing though...

> In this production environment only a single thread was indexing. No
> addIndexes() is used. In some ocassions yes write.lock might be
> deleted by code.

Removing the write.lock when you shouldn't is a recipe for disaster --
is there any chance you removed it when a writer was in fact still
active?

> I have added infoStream to stdout in my code so in the future we would
> run this with debug on.

OK excellent.  Post back with the output if/when it happens... thanks!

Mike

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Mime
View raw message