[Bucardo-general] Replication scenario and Bucardo application

[Robert Scott Davis]

I have a scenario where there is one central server located on the  
Internet.  There are multiple potentially on-line and off-line  
laptop/desktop systems in the field that have at one point the db  
contents of the central server.  There is no guaranteed Internet  
connection for the fielded systems to replicate with the central  
server.  When they do come on-line they will be sync'ed with the  
central server via a replication program.  In this scenario I believe  
each of the fielded systems would be called a master.  There are more  
than two fielded systems which I believe may be a issue with bucardo  
since I believe you can only have two masters currently.  I think  
bucardo would work for my scenario but I am concerned about the only  
two masters are supported.  In my scenario is there a work around  
and/or am I using the term master incorrectly and potentially they are  
slaves.

In summary the fielded systems can be connected to internet and in  
"real-time" be replicating with the central server and in some cases  
there can be some fielded systems that are not connected to the  
internet and operating with other client systems in a LAN that at some  
point would be connected to the internet and would want to replicate  
with the central server.

Is the use of bucardo viable for this situation.

I am using Linux 2.6.16 (open SuSE), PostgreSql 8.1, and latest Perl RPMs.
Scott