Thanks a lot for all people which answer.
I have this "clustering on SAN" problem today and i think it could be
less harder to implement this today than it was for Oracle in 1993
(since i can find a lot of work in opensource which could be interesting
in this project : distributed lock, clustered shared memory, etc ...).
I will try to have a look about all these things btw, i haven't got
the pretention to suceed to do anything ... Just look and see
a bit more deeply how it could be hard.
Is there something like a "big picture" about postgres internal ?
What path is used from the SQL query to the hard-disk files ? About
the internal locking mechanism ? about the usage of the shared memory ?
I see there is everything described in the postgres doc but no "big
picture" which could help me to understand faster ... :(
I'm using postgres for several year and i have nerver problem with
it i'm really happy with these trigger and inheritance mechanism which
help me more than one time and i would present my respect to those guy
which write this so good piece of software.
I hope to see you later with something to show ...
Sincerly, Yannick.