Sounds like you are begging for trouble.
Suppose that a customer calls in to the main office, and you update some
customer data.
The field salesman also updates data for this customer.
If you update the main office database with the field data, you will
lose information.
If you update the field database data with main office database data,
you will lose information.
Because we have two different changed records, it will be very difficult
to reconcile this data without human intervention.
In short, a disconnected system where data on both ends can change is
begging for trouble.
Now, if you never update the database in the main office except with
data from the field salesmen, it could be made to work. But I doubt
that this is what you are after.
If it were me, and someone proposed a model where two-way replication
was needed, I would tell them to rethink their model. It's broken.
IMO-YMMV.
-----Original Message-----
From: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org] On Behalf Of Caleb
Simonyi-Gindele
Sent: Tuesday, March 29, 2005 10:58 AM
To: pgsql-general@postgresql.org
Subject: [GENERAL] do I need replication or something else?
We have a billing system and we want the ability to send users out into
the field with an unconnected (no WAN, VPN etc) laptop containing our pg
db and software. Upon their return we need to synchronize changes to the
main db. We would like the ability to be able to have this accomplished
at the db level rather than doing this in our app.
What is the best tool to accomplish this with?
Caleb