> Hi Remy
>
> We at Metatomix have been using Slide for sometime and would be
> interested in participating in the development in the area of database
> pooling in Slide.
> Does the slide have any database pool for handling connections at
> present.

No, but I've added it in the status page (which I just uploaded to the live
website), since it's obviously critical for scalability.

> If Not we would like to contribute and participate in developing the
> pool handling capacity of Slide. Please Let me know if U already have
> some ideas or specs regarding the mode of implementation of Database
> pools.

Thanks for volunteering. I could definitely use the help.

The plan was:
- Reuse the JDBC code from the current JDBC store (and probably refactor it
a bit, so that we can avoid code duplication).
- Use the pool from the Commons (at least default to it; we could support
additional pools in the future).
- We need to associate one connection <-> one transaction, and we need to
behave in the way the TM expects the store to bahave (which isn't too hard).
A hashtable keyed by transaction id will probably do it.
- Keep the JDBC store separate (it's supposed to be a reference store).

So I'd say it should be mostly a refactoring of the JDBC code, with the
addition of the pool handling code, and the transaction state handling code.

Remy


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to