PostgreSQL Replication
Replication is, simply put, copying data from one place to another. PostgreSQL includes several built-in replication features such as write-ahead log (WAL) shipping, warm and hot standby, and streaming replication. Some widely-used third-party replication tools include Bucardo, developed by End Point, as well as Slony, Londiste, pgbouncer, and pgpool.
Each tool has its own strengths and constraints. End Point’s consultants have experience determining the tools that work best in various situations and helping you achieve the right balance between cost, availability, and performance.
Put End Point’s PostgreSQL replication experts to work for you!
Why Replication?
Replicas for load distribution
Growing businesses often reach a point where their database systems slow unacceptably under increasingly high demand. Replication allows you to create copies of your main databases, known as slaves or replicas. These replicas can then take up some of the load from reporting jobs and other read-only activities. Bucardo and Slony are popular replication solutions that can support this arrangement. PostgreSQL 9.0 and newer provide another read-only replica option with its “hot standby” feature. End Point can help you determine which of these solutions will be best for you and your business.
Multi-master replication
Bucardo also offers two-way replication such that two databases can be complete authoritative repositories of all information, updated in near real-time. This can be achieved across multiple datacenters for risk mitigation, lower latency to local services, or to distribute load. We can help you avoid pitfalls and make the most of a multi-master replication system.
Backups
There are many ways to back up a database. The most common PostgreSQL backup utility, pg_dump, is simple and easy to use, but can add load to your production system. Some workloads benefit from replicating data to a slave database and exporting that periodically instead. We have experience with other backup solutions such as point-in-time recovery (PITR) and filesystem snapshots, and can create a solution suited to your needs.
Data warehousing
Businesses have always collected data; many businesses now want to “mine” or analyze their data for marketing, performance, and process optimization purposes. This usually involves very complex queries, which can be expensive both in time and server processing power. A common approach is to periodically transfer the data from production databases into a “data warehouse” to simplify the queries and ease load on the main database servers, in a process known as “Extraction, Transformation, and Loading”, or ETL. Bucardo allows seamless transformation of data between master and replica, and can handle ETL processes in a single package.
Disconnected servers
Sometimes businesses or applications need to distribute changes across long distances or delayed over time. In these scenarios latency and disruptions to connectivity can affect replication in unexpected ways. End Point engineers are versed in the options, and can help you build a solution that will handle your more complex needs.
