This class is a type-safe enum which contains the valid lifecycle states for a broker. It is used by ClusterConfig to set the state of the broker. State is stored as an int in JDBC and used as a typesafe enum within the code.
State | Description | Behavior |
INITIALIZING | a new broker is starting | set by ClusterConfig.addBroker() |
OPERATING | The broker is processing messages | set by Broker at startup |
QUIESCE_STARTED | The broker has started to quiesce, this means that new jms connections will not longer be accepted | called by BrokerStateHandler.quiesce() |
QUIESCE_COMPLETED | Quiesce has completed, this means that all jms connections have exited, the broker can not be safely shutdown. | called by BrokerStateHandler.quiesce() |
SHUTDOWN_STARTED | The broker is starting to shutdown. It may shutdown immediately or wait some period of time | set by BrokerStateHandler.shutdown() |
SHUTDOWN_FAILOVER | the broker has finished shutting down, all resources have been released it should be taken over (the broker may or may not shutdown, depending on whether it is running inprocess) | set by BrokerStateHandler.shutdown() |
SHUTDOWN_COMPLETE | the broker has finished shutting down, all resources have been released but is should NOT be taken over (the broker may or may not shutdown, depending on whether it is running inprocess) | set by BrokerStateHandler.shutdown() |
FAILOVER_PENDING | a takeover is pending (not stored to disk) | set by the TakeoverThread of HAMonitorService() |
FAILOVER_STARTED | another broker in the cluster has started to takeover the state of this broker (the broker handling takeover can be retrieved with HAClusteredBroker.getTakeoverBroker() | set by the TakeoverThread of HAMonitorService() |
FAILOVER_COMPLETE | takeover of a broker has completed | set by the TakeoverThread of HAMonitorService() |
FAILOVER_FAILED | a takeover failed (not stored to disk) | set by the TakeoverThread of HAMonitorService() |