Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
MIME-Version: 1.0
In-Reply-To: <9ac0c6aa0904241411h168d8edaid96d5bc07075bf69@mail.gmail.com>
References: <49E4573E.7070208@aduna-software.com>
	 <9ac0c6aa0904140306t4235c120n73629596899256bc@mail.gmail.com>
	 <49E6FD9F.5080104@aduna-software.com>
	 <9ac0c6aa0904160510l6cf559c6yfc6b0c1deadbf1de@mail.gmail.com>
	 <49EE41F9.3030007@aduna-software.com>
	 <49EE4B71.8030105@aduna-software.com>
	 <9ac0c6aa0904230302p568892f2p1cf365f4a1d61207@mail.gmail.com>
	 <49F228C8.3020002@aduna-software.com>
	 <9ac0c6aa0904241411h168d8edaid96d5bc07075bf69@mail.gmail.com>
Date: Fri, 24 Apr 2009 18:02:08 -0400
Message-ID: <9ac0c6aa0904241502u1e8e781an889114eb978b4dcf@mail.gmail.com>
Subject: Re: semi-infinite loop during merging
From: Michael McCandless <lucene@mikemccandless.com>
To: java-user@lucene.apache.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

OK I opened https://issues.apache.org/jira/browse/LUCENE-1611.

Christiaan, could you try out that patch to see if it fixes the
semi-infinite merging?  Thanks.  (You'll need to back-port to 2.4.1,
but it's a very small patch so hopefully not a problem).

Mike

On Fri, Apr 24, 2009 at 5:11 PM, Michael McCandless
<lucene@mikemccandless.com> wrote:
> On Fri, Apr 24, 2009 at 5:02 PM, Christiaan Fluit
> <christiaan.fluit@aduna-software.com> wrote:
>
>> Rollback does not work for me, as my IW is in auto-commit mode. It gives=
 an
>> IllegalStateException when I invoke it.
>>
>> A workaround that does work for me is to close and reopen the IndexWrite=
r
>> immediately after an OOME occurs.
>
> Ahh, right: that's (close the writer) the correct workaround (when
> autoCommit=3Dtrue). =A0Though... it'd be best to not hit OOME so often.
>
>> I added message(String) invocations in all places where the IW.hitOOM fl=
ag
>> is set, to see which method turns it on. It turned out to be addDocument
>> (twice). These OOME's only happen with the JET build, which explains why=
 the
>> Java build does not show the exploding index behavior: the hitOOM flag i=
s
>> simply never set and the merge is allowed to proceed normally.
>
> Why does JET hit OOME but straight JRE doesn't?
>
>> The flag is definitely not set while the IW is merging, nor do any OOME'=
s
>> appear in my log files during merging. Therefore, there must be a proble=
m in
>> how the merge operation responds to the flag being set.
>
> OK; likely a merge eventually tried to start after the OOME.
>
> Mike
>

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