lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael McCandless <>
Subject Re: Minimum HD usage during an optimize() call
Date Mon, 30 Mar 2009 19:01:19 GMT
In fact the documentation is wrong.

With autoCommit=false (I assume you used that?) the usage will peak at
3X, because the writer keeps the initial index "alive" exactly as an
open reader would do.

I'll fix the docs.


On Mon, Mar 30, 2009 at 1:09 PM, Lebiram <> wrote:
> Hi all,
> I was trying to determine if the documentation for optimize() is true:
> Testing was done using Lucene 2.4
> I basically have 2 lucene index,
> Index A) one with no Searcher open during optimize() and
> Index B) another with a searcher open during optimize().
> Also, I have daemon threads setup to monitor the size of the entire index. At the end
of the execution, the daemon thread will display the max HD usage and the current HD usage
(after optimize() )
> Now, I found out that upon completion,
> Index A has the optimized size,
> Index B has x2 of the optimized size.
> which is as what is in the documentation.
> However,
> Index A max HD usage during optimize() was the same as max HD usage of Index B.
> That is, triple the original size of the index (at one point during close() after optimize()
was called) .
> According to the documentation, Index A should just use twice HD max.
> This is important for me because I want to be able to anticipate the HD requirement for
a really big Lucene Index.
> Is it always the case that the HD requirement for Lucene Index across all versions is
x3 of the expected max Index size?

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message