[storm] Locking and concurrency support for PostgreSQL in Storm

[James Henstridge]

On 08/08/07, akira <nhytro-python at web.de> wrote:
> Hi guys!
>
> How is locking and concurrency support for Postgres implemented in
> Storm? Should one only use transactions?

Yep, Storm sets up the database connection in the serializable isolation level.

This means that you Storm can assume it is the only thing accessing
the database during the transaction and cache things accordingly.
This does mean that transactions may fail and need to be retried
though.

James.