Managed client connections

When IBM MQ classes for .NET are used as a managed client, there are a number of differences from a standard IBM MQ MQI client.

The following features are not available to a managed client:

• Channel compression
• Channel exit chaining

If you try to use these features with a managed client, it will return an MQException. If the error is detected at the client end of a connection, it will use reason code MQRC_ENVIRONMENT_ERROR. If it is detected at the server end, the reason code returned by the server will be used.

Channel exits written for an unmanaged client do not work. We must write new exits specifically for the managed client. Check that there are no invalid channel exits specified in your client channel definition table (CCDT).

The name of a managed channel exit can be up to 999 characters long. However, if we use the CCDT to specify the channel exit name, it is limited to 128 characters.

Communication is supported only over TCP/IP.

When you stop a queue manager using the endmqm command, a server-connection channel to a .NET managed client can take longer to close than server-connection channels to other clients.

If we are have set NMQ_MQ_LIB to managed in order to use managed IBM MQ problem diagnostics, none of the parameters -i, -p, -s, -b, or -c of the strmqtrc command is supported.

A managed .NET application using XA transactions will not work with a z/OS queue manager. A managed .NET client attempting to connect to a z/OS queue manager fails with an error, MQRC_UOW_ENLISTMENT_ERROR (mqrc=2354), on MQOPEN call. However, a managed .NET application using XA transactions will work with distributed queue manager.

Parent topic: Connection differences