What are my options for clustering (load balancing, failover, etc.)?
The OpenSRF framework is inherently cluster-friendly, allowing the ad-hoc addition of inexpensive commodity hardware when system load dictates. The PINES implementation currently uses approximately 25 servers in production, and machines are added and removed on a regular basis as needed. Evergreen uses PostgreSQL, the most advanced and robust Open Source RDBMS available, as its back-end data store. In order to provide load balancing for the Evergreen database, PINES has implemented Slony-I with PGPool as a clustering solution. Other clustering solutions exist for PostgreSQL, and may make sense in other environments. In order to provide load balancing and high-availability at the OPAC and Staff Client level, PINES has implemented a Linux Virtual Server environment with five independent mini-clusters. This allows live updates of the entire system with no perceived downtime or interruption in service.
Related Questions
- How does performance typically compare between Windows Network Load Balancing (Active-Active) and Windows Failover Clustering (Active-Passive)?
- Does the RoboSuite runtime server support load balancing and failover scenarios? Can it spread client requests across multiple servers?
- What are my options for clustering (load balancing, failover, etc.)?