Abnormal Shutdown
A shutdown is considered abnormal if the application receives an immediate shut down signal from the operating system and does not get a chance to complete shutdown processing; for example, kill -9 command in Unix.
When an abnormal shutdown occurs, the following may happen:
Some locks created to manage concurrency are not released. Such locks are created by timer tasks. These locks are automatically cleared when the server which was shutdown abnormally and which created these locks is restarted. If required, these locks can be manually cleared by deleting the following files:
- Revivifier - $MQ_HOME/Work/MqRevivify.lock
- FileWatcher - Location specified in the
FileWatcher.xml file
The workflow processes running while the process execution is aborted will be in an incomplete state. The messaging system will detect the failure of the message associated with the workflow process and tag the same message for restart. When the application comes back up online, the message will be re-delivered and the workflow process will be restarted at the point of failure. The duplicate execution of workflow activities can infrequently lead to duplicate artifact, such as duplicate work items or event log entries.