Nope.

Just don't expect beowulf to buy you anything with database intensive
apps. I presume you have read the beowulf FAQ's that used to be at
www.beowulf.org.

By its nature, using firebird as backing store for beowulf architecture
programs should not be an issue - each system in the cluster is just
another client connection to the firebird server.

The only gotchas from the Firebird side are the usual performance
gotchas. Since the first three priorities of beowulf computing are (in
order) speed, speed, and speed, pay close attention to anything Helen
has up her sleeve on Firebird performance.

It is not usual to use any RDBMS as backing store for beowulf systems
because they are considered too slow. With that said, parcelling out
data files manually really sucks - an RDBMS makes sense provided it is
used as input and data sink for the processes, and not in any place
where it impacts computational performance. As long as your computation
is CPU intensive enough that the RDBMS data interface is not a handicap
you should be okay.

Note that Firebird will function over TCP/IP, not any of the
replacements used to boost communications performance in beowulf
architectures. Mind you, a PVM communications plugin for Vulcan might
be interesting ... never mind the manic gleam in my eye ... get that
whit coat away from me ... aaaaah!

On Mon, 2005-08-22 at 21:00 -0700, Rocky Castaneda wrote:
> Hi,
>
> Anybody tried installing firebird on a clustered system using beowulf?
>
> Any tips and any effects on firebird? Any warning I should know?
>
> All the best,
>
> rocky
>