Resynchronizing the IMS bridge

The IMS bridge is automatically restarted whenever the queue manager, IMS, or OTMA are restarted.

The first task undertaken by the IMS bridge is to resynchronize with IMS. This involves WebSphere MQ and IMS checking sequence numbers on every synchronized Tpipe. A synchronized Tpipe is used when persistent messages are sent to IMS from a WebSphere MQ-IMS bridge queue using commit mode zero (commit-then-send).

If the bridge cannot resynchronize with IMS, the IMS sense code is returned in message CSQ2023E and the connection to OTMA is stopped. If the bridge cannot resynchronize with an individual IMS Tpipe, the IMS sense code is returned in message CSQ2025E and the Tpipe is stopped. If a Tpipe has been cold started, the recoverable sequence numbers are automatically reset to 1.

If the bridge discovers mismatched sequence numbers when resynchronizing with a Tpipe, message CSQ2020E is issued. Use the WebSphere MQ command RESET TPIPE to initiate resynchronization with the IMS Tpipe. You need to provide the XCF group and member name, and the name of the Tpipe; this information is provided by the message.

We can also specify:

• A new recoverable sequence number to be set in the Tpipe for messages sent by WebSphere MQ, and to be set as the partner's receive sequence number. If you do not specify this, the partner's receive sequence number is set to the current WebSphere MQ send sequence number.

• A new recoverable sequence number to be set in the Tpipe for messages received by WebSphere MQ, and to be set as the partner's send sequence number. If you do not specify this, the partner's send sequence number is set to the current WebSphere MQ receive sequence number.

If there is an unresolved unit of recovery associated with the Tpipe, this is also notified in the message. Use the WebSphere MQ command RESET TPIPE to specify whether to commit the unit of recovery, or back it out. If you commit the unit of recovery, the batch of messages has already been sent to IMS, and is deleted from the bridge queue. If you back the unit of recovery out, the messages are returned to the bridge queue, to be subsequently sent to IMS.

Commit mode 1 (send-then-commit) Tpipes are not synchronized.