FAQ / When an application/service error is detected in and RSF-1 Cluster, is it possible to restart components locally or does failover always have to happen?

Application and service errors are detected by the RSF-1 agent framework. An agent is started as part of a service and can utilise a number of ways to test the health of the applications in that service (for instance writing and reading from a database or monitoring a port for expected responses to connection probes).

The action taken on service failure is fully configurable and can be as simple as restarting the service locally, allowing a number of restarts withing a sliding window of time, or performing a fail over of that service to another node.

Posted in: Administration