SVN Clustering

Key Features

• Dynamically balances workload across Subversion servers at a single location over a LAN, or multiple locations thousands of miles apart with Subversion MultiSite (WAN Clustering).
• Subversion LAN Clustering has no single point of failure. There is no sharing of disk, CPU or memory between servers in a cluster. Even load balancers can be paired for redundancy using the resilient cluster option.
• Intelligent load balancing capabilities optimize performance for end users by taking each server's current load into account before routing requests. Load balancing can also be configured to take the content of the data in a request into consideration so that specific tasks can be routed to the servers best suited to handle them.
• The load balancers run as software services, not hardware appliances. This offers tremendous deployment flexibility and long-term scalability as a Subversion implementation grows.
• There are no passive standby servers. All servers are active peers simultaneously supporting user requests.
• All servers in a cluster are fully readable and writeable and always in sync, making the cluster appear to end users as a single, always available Subversion server.
• When a network or server outage takes place, failover is transparent and immediate and recovery is automatic. Users experience no downtime or data loss, and the risk of human error is eliminated.
• Automated failover and recovery capabilities can be used to take servers offline for maintenance without interrupting user access, making full 24-by-7 operation possible. When a server is brought back online after maintenance is complete, it resynchronizes automatically with the other servers in the cluster.
• Connection draining allows existing tasks to be completed, while blocking new user connections before scheduled server maintenance takes place.
• Hot deploy capability allows servers to be added to a cluster without any downtime.
• Subversion LAN Clustering can be implemented stand-alone or in combination with Subversion MultiSite (WAN Clustering).
• SSL Support Encrypts all Network Traffic to Meet the Most Stringent Security Requirements. Transactions between Subversion servers and between Subversion clients and servers can now be configured to use SSL.
• Parallel Transaction Processing. Each repository now has its own transaction queue. This allows replicated transactions to be processed in parallel across multiple repositories for improved performance and tighter control in large development environments.
• Silent Installer Non-GUI installer automates deployment for faster implementation and minimal downtime when new servers are added to a Subversion Clustering implementation.
• Complete LDAP Integration for the Admin Console Multiple administrators can log on using their individual LDAP credentials.
• New RESTful API allows third party monitoring systems and custom scripts to query Subversion Clustering's administration console and modify its configuration. Subversion Clustering operations can also be invoked externally through the API.

Can Subversion Clustering be Implemented With Just Two Nodes?

Subversion Clustering can be implemented with just two nodes. However, this will only support load balancing, not high availability. In a two node cluster, if one node fails, this leaves a single server and a single point of failure without another server to recover from. A minimum of three nodes is recommended to safely support both high availability and load balancing.

Subversion Clustering Optmizes Build Performance for Agile Development

Agile development is iterative and incremental, requiring continuous communication and continuous build integration. Subversion LAN Clustering allows build processes to be offloaded from the Subversion servers used by developers, significantly improving their productivity. At the same time, the latest changes from the development team are available on the other servers in the cluster where continuous builds are running.

A number of WANdisco's customers are implementing Subversion Clustering in combination with Subversion MultiSite (WAN Clustering) to fully enable agile development in a distributed environment. This combination eliminates WAN latency for remote developers as well as the performance bottleneck of a single Subversion server at the largest development sites.