jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexander Klimetschek <aklim...@day.com>
Subject Re: Jackrabbit and web application
Date Sun, 28 Mar 2010 17:34:31 GMT
On Sun, Mar 28, 2010 at 16:44, Ilya Skorik <ilya@skorik.me> wrote:
> I plan to use jackrabbit in a Web application. Prompt, what strategy is
> necessary for selecting for the connection organisation to base?
>
> Considering specificity of Web applications when the server handles set of
> queries from different users.
>
> Whether there is something like connection pooling for jackrabbit?

In JCR you have sessions that are per-user. Unlike JDBC, you typically
actually login with the front-end users to the repository as JCR has a
good ACL model (so no "db-admin" user for all connections). Also,
acquiring sessions is very fast and light-weight, hence a new session
per request is best practice.

Experiments with session pooling (eg. in Apache Sling) didn't benefit,
they even tended to create more problems due to unwanted side-effects
of re-using old sessions.

With regard to backend connections inside Jackrabbit to eg. a
database, those connections are already pooled and safely handled by
Jackrabbit 2.0 (using the standard database persistence managers). See
http://wiki.apache.org/jackrabbit/ConnectionPooling

Regards,
Alex

-- 
Alexander Klimetschek
alexander.klimetschek@day.com

Mime
View raw message