On 29.11.10 12:47, "Seidel. Robert" <[email protected]> wrote:

>I think, the use-case is the same, that we have::
>
>Multi cluster environment - only one single Lucene index within the
>database for all nodes would
>    -> takes less disk space (if you have more than one node)
>    -> make backups easier (you only have to backup the database, not the
>db and some folders on three nodes)
>    -> makes it easier to add nodes (you don't have to wait, until the
>index is built up)
>    -> make the index more stable (because the database is transactional)

Ok, makes sense, but this would be an optimization purely for writes. I
would expect that the benefits are completely offset by the much worse
read performance of this setup.

Regards,
Alex

-- 
Alexander Klimetschek
Developer // Adobe (Day) // Berlin - Basel




Reply via email to