High Availability

Tungsten provides simple, automated failover procedures to allow applications to switch to replicas quickly, seamlessly, and without risk of data loss. This helps address not just failures but also maintenance, which is the main source of downtime in modern IT environments.

Continuous Operation during Emergency Failure

Automated failure detection and recovery. Tungsten failover is a one-step procedure that automatically switches masters and reconfigures the rest of the cluster, including directing other slaves to the new master and updating load-balancing.

Master recovery and failback. Tungsten can promote any slave to be master without data loss or requiring other slaves to be stopped or reprovisioned. Masters can be recovered with a single command.

Failback to the recovered master is likewise a single command. Fail-safe data application on slave. Slaves remain consistent even following crashes thanks to transactional data replication, which ensures that SQL updates are neither dropped nor repeated.

Automatic Routing of SQL. Tungsten provides a variety of options to route SQL from application to an appropriate database server, including Java routing libraries, language-independent proxies, and floating IP addresses. Tungsten automatically shifts new connections as masters move and as administrators add and remove databases in the cluster.

Continuous Operation During Maintenance

Seamless, transparent switchover for maintenance. Tungsten allows users to switch master databases without data loss or halting applications, thereby facilitating daylight maintenance operations.

Zero-downtime upgrades. Run upgrade and maintenance processes on slaves, then switch masters after the upgrade has finished successfully. This eliminates the risk of data loss due to upgrade mistakes.