Symphony fault tolerance

Symphony has no single point of failure. Every component in the system has a recovery operation—every component is monitored by another component, and can automatically recover from a failure.

Fault tolerance for resources

All services managed by EGO (master lim, session director, repository service, PERF loader controller, and others) can be configured to fail over to another management host.

Fault tolerance for workload

Symphony provides the following fault tolerance for workload:
  • The service instance is monitored and recovered by its service instance manager

  • The service instance manager is monitored and recovered by its session manager together with EGO

  • The session manager is monitored and recovered by session director together with EGO

  • Workload (sessions and tasks) can be specified as recoverable or not

  • Task delivery is guaranteed

Fault tolerance among applications

Symphony provides the following fault tolerance for applications:
  • All service-oriented application properties are defined in application profiles throughout the cluster

  • Administrators are in full control of Symphony applications for their assigned profiles

  • If EGO encounters an issue, it does not affect running Symphony applications

  • If one service instance fails, it does not affect any others

  • Persistence, history logging, and data cache are managed per application

Related concepts
Symphony fault tolerance