saga-actuator only provide a central recovery mechanism. recovery thread will be started independent on server side. but this could not work for sharding-sphere, physical datasource only kept on the application side. so we need make every application instance recovery independently. if one instance have crashed, it can be failover to another available instance.
saga-actuator only provide a central recovery mechanism. recovery thread will be started independent on server side. but this could not work for sharding-sphere, physical datasource only kept on the application side. so we need make every application instance recovery independently. if one instance have crashed, it can be failover to another available instance.