on Postgres high availability -- it is also possible to have a
database on one machine act as master, and one or more databases on
other machines act as slaves, updating data from the master. How you
implement this depends on the nature and needs of your application.
You can also use the slave machines to serve clients that have read-only
needs. If you design your application well, you can switch one of the
slave machines to master if the master goes down.
>>>>> "NH" == Nicolas Huillard <[EMAIL PROTECTED]> writes:
NH> "That way" ?
NH> Is there another way you think about ?
NH> I think the solution proposed by Jorge Garcia could work : CODA as a distributed
file system, in order to be able to retreive the database installation in case of
crash of one machine, and an idle postgres that can quickly serve this installation
(with a few minutes gap). Is there something else to do
NH> Nicolas Huillard
