SQL-Ledger offers a nice inventory control module (a.k.a. Goods & Services) that provides inventory count, and making its data available to our online store allows real-time inventories and more. For our servers that are at different locations, this could be done by querying the SL server in real time, but a lost of communications could disable our production system in this synchronous setup. So we use Slony-I to replicate data.
Slony-I is a Postgresql database replication tool and it utilizes table triggers and watchdog daemons to maintain the replication cluster in sync. In addition, it also comes with a simplified set of tools to help setup and maintain the cluster. Slony-I replicates its cluster asynchronously, and its daemon is capable of picking up where it left off just prior to communications failure–a main reason that Slony was implemented in our software architecture, and you can read up on Slony-I here.
At the end, using Slony-I, I was able to make SL’s AR and inventories available to the production server, and replicate sales and customer data to SL via a two-way (clarification, in a sense via different table sets not like a true two-way Oracle replication) replication.