lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Carlson <>
Subject Re: Optimization
Date Mon, 03 Jun 2002 14:39:39 GMT
If you watch what's going on in the index folder (if you are using an
FSDirectory), you will see that Lucene actually creates a brand new set of
index files with a different prefix (like current set is 2411_... And new
set is 2412_...). The current index files still exist until the new index is
fully created.

If you take a backup in the middle of optimization, I am pretty sure that
you will get the uncompleted state of the new index. This means that the new
index will not be consistent, but the current index will fine.


On 6/3/02 12:03 AM, "Nader S. Henein" <> wrote:

> Lucene and Backing up the index :
> Since we're on the concurrency issue, can I take a back up from lucene while
> it's indexing or optimizing (using the writer), will this be a consistent
> backup image
> -----Original Message-----
> From: Peter Carlson []
> Sent: Monday, June 03, 2002 10:38 AM
> To: Lucene Users List
> Subject: Re: Optimization
> There should be no problem, especially if you are working on a unix based
> machine.
> Since the reader will be using the current index and the optimizer will be
> creating a new index there should be no problem. However, you will have to
> close the reader and reopen it to have it use the new index.
> There has been some mention on the list about the optimizer deleting the old
> indexes, but if you are using a unix machine then the index files will still
> be available as long as one process is using them. I don't know how this
> would work on Windows (my guess is it wouldn't be able to delete the index
> files in use).
> At least this is my understanding.
> --Peter
> On 6/2/02 10:54 PM, "V S" <> wrote:
>> As optimization is a slightly different process,
>> basically merging segments,  I was not sure if this
>> would not block, especially if you are querying on the
>> segment that is being merged?
>> I do see indexing and searching working in parallel
> --
> To unsubscribe, e-mail:
> <>
> For additional commands, e-mail:
> <>
> --
> To unsubscribe, e-mail:   <>
> For additional commands, e-mail: <>

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

View raw message