lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Carlton <carl...@sumologic.com>
Subject Re: any tips for upgrading Lucene 3.0.3 -> 3.5.0?
Date Thu, 19 Jan 2012 23:25:42 GMT
Awesome, thanks - that puts me a lot more at ease.

On Thu, Jan 19, 2012 at 2:59 PM, Uwe Schindler <uwe@thetaphi.de> wrote:

> Lucene 3.5 can read any index going back to 2.0. The IndexUpgrader is only
> needed to "forcefully" upgrade indexes for maximum performance and safe
> migration to Lucene 4.0 (that can only read indexs >= 3.0).
>
> -----
> Uwe Schindler
> H.-H.-Meier-Allee 63, D-28213 Bremen
> http://www.thetaphi.de
> eMail: uwe@thetaphi.de
>
>
> > -----Original Message-----
> > From: David Carlton [mailto:carlton@sumologic.com]
> > Sent: Thursday, January 19, 2012 8:02 PM
> > To: java-user@lucene.apache.org
> > Subject: any tips for upgrading Lucene 3.0.3 -> 3.5.0?
> >
> > I'm hoping to upgrade Lucene on a local code base from 3.0.3 to 3.5.0; is
> there
> > a good guide out there for particular pitfalls that I should worry about?
> I've
> > skimmed the ChangeLogs; the mention of an index upgrade tool made me
> > wonder: has the index format changed between those versions? If so,
> what's
> > the best way to handle that without reindexing?  (I was assuming that I'd
> be
> > able to be careful about where I pass in VERSION_30 and where I pass in
> > VERSION_35, but looking through the current code, there actually are
> basically
> > no non-test situations where we pass in a version number at all - e.g. we
> only
> > use QueryParser for test code - so maybe I don't have to worry about
> > incompatibilities?) Any other gotchas that won't be obvious from
> compilation
> > failures and straightforward sanity check test failures?
> > (We do have tests that run basic queries on sample input.) Are there
> > intermediate versions that we should try, e.g. 3.0.3 -> 3.4.0 -> 3.5.0
> instead of
> > going straight from 3.0.3 to 3.5.0?
> >
> > If it matters: the one unusual Lucene usage that we have is that we've
> written
> > our own Tokenizer.
> >
> > Thanks for any pointers that people have.
> >
> > --
> > David Carlton
> > carlton@sumologic.com
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>


-- 
David Carlton
carlton@sumologic.com

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message