Multicast application monitoring
Use this information to learn about administering and monitoring IBM MQ Multicast.
The status of the current publishers and subscribers for multicast traffic (for example, the number of messages sent and received, or the number of messages lost) is periodically transmitted to the server from the client. When status is received, the COMMEV attribute of the COMMINFO object specifies whether or not the queue manager puts an event message on the SYSTEM.ADMIN.PUBSUB.EVENT. The event message contains the status information received. This information is an invaluable diagnostic aid in finding the source of a problem.
Use the MQSC command DISPLAY CONN to display connection information about the applications connected to the queue manager. For more information on the DISPLAY CONN command, see DISPLAY CONN.
Use the MQSC command DISPLAY TPSTATUS to display the status of your publishers and subscribers. For more information on the DISPLAY TPSTATUS command, see DISPLAY TPSTATUS.
COMMEV and the multicast message reliability indicator
The reliability indicator, used in conjunction with the COMMEV attribute of the COMMINFO object, is a key element in the monitoring of IBM MQ Multicast publishers and subscribers. The reliability indicator (the MSGREL field that is returned on the Publish or Subscribe status commands) is an IBM MQ indicator that illustrates the percentage of transmissions that have no errors Sometimes messages have to be retransmitted due to a transmission error, which is reflected in the value of MSGREL. Potential causes of transmission errors include slow subscribers, busy networks, and network outages. COMMEV controls whether event messages are generated for multicast handles that are created using the COMMINFO object and is set to one of three possible values:
- DISABLED
- Event messages are not written.
- ENABLED
- Event messages are always written, with a frequency defined in the COMMINFO MONINT parameter.
- EXCEPTION
- Event messages are written if the message reliability is under the reliability threshold. A
message reliability level of 90% or less indicates that there might be a problem with the network
configuration, or that one or more of the Publish/Subscribe applications is running too slowly:
- A value of MSGREL(100,100) indicates that there have been no issues in either the short term, or the long-term time frame.
- A value of MSGREL(80,60) indicates that 20% of the messages are currently having issues, but that it is also an improvement on the long-term value of 60.
Clients might continue transmitting and receiving multicast traffic even when the unicast connection to the queue manager is broken, therefore the data might be out of date.