Index maintenance - best practices

[Abhay Saswade <ay...@i.frys.com>]

Hi Folks,

My search app is dealing with multiple indices each contain 1 to 20
fields, 100k to 1 million documents and size ranging from 50 megs to 10
gb.

Wanted to know best practices to deal with maintenance of
indices/collections like this with no down time?

Background: I wrote a very simple search application some time back and it
is running fine. Presently it’s running on Lucene 1.2. A separate index
creation application creates whole index @ different location and copy it
to actual search application location. Copy process points search app to
new index location after copying is done. All this takes place with no
downtime.

Now I have upgraded search app to Lucene 1.4 and thought it’s a good time
to ask pros how they deal with this especially with 1.4, where you can add
documents to/from FSDirectory and RAMDirectory.

Some issues are
1. Effectiveness of RAMDirectory under diff O.S. especially on Linux.
2. Strategy for whole and incremental indexing
3. Backup of RAMDirectory or FSDirectory
4. Anything else pertaining to this topic you can think of

Thanks in advance,
Abhay



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