I am considering pgsql as the RDBMS in a project with the following constraints: - there's a master and reserve instance of the RDBMS on every remote location - there's a master and reserve instance of the RDBMS on a central location - the connections are DSL connections and therefore unreliable - all changes have to be propagated to all servers (multimaster replication) - if the connection between a remote location and the central location fails, the local server continues working and resynchronizes with the central server when the connection is restored - if any master fails, the reserve instance takes over and the rest of the system acts as though nothing happened The master/reserve instance is, from what I read, standard functionality, but I'm not so sure about the resynchronization part of a failed link...I imagine something like WAL shipping might be of use here, but it's just an uneducated guess. Does code exist to support this on pgsql or is it considered application specific functionality? Do other RDBMSs support similar functionality? TIA, t.n.a. ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
