Handling protected conversation messages across IMS restart in shared queues environment

摘要

During phase one of each protected conversation transaction, one or more messages are inserted. These messages are logged for recoverability. IMS system could crash before phase two of commit. When IMS restarted and eventually reached phase 2, it could vote to abort or commit the conversation. During IMS restart, the retry logic is used to put all protected conversation messages out to central facility with a specification that the first message segment be put on the staging queue using the recovery token as the queue name. If it is an abort request, all messages under the same commit scope identified by the recovery token, will be deleted. If it is a commit request, the first segment of all messages under the same commit scope will be moved to its appropriate ready queue with its actual queue name.