Channel definition commands

Cluster attributes that can be specified on channel definition commands.

The DEFINE CHANNEL, ALTER CHANNEL, and DISPLAY CHANNEL commands have two specific CHLTYPE parameters for clusters: CLUSRCVR and CLUSSDR. To define a cluster-receiver channel we use the DEFINE CHANNEL command, specifying CHLTYPE(CLUSRCVR). Many attributes on a cluster-receiver channel definition are the same as the attributes on a receiver or sender-channel definition. To define a cluster-sender channel we use the DEFINE CHANNEL command, specifying CHLTYPE(CLUSSDR), and many of the same attributes as we use to define a sender-channel.

It is no longer necessary to specify the name of the full repository queue manager when you define a cluster-sender channel. If you know the naming convention used for channels in your cluster, we can make a CLUSSDR definition using the +QMNAME+ construction. The +QMNAME+ construction is not supported on z/OS®. After connection, IBM MQ changes the name of the channel and substitutes the correct full repository queue manager name in place of +QMNAME+. The resulting channel name is truncated to 20 characters.

For more information on naming conventions, see Cluster naming conventions.

The technique works only if your convention for naming channels includes the name of the queue manager. For example, you define a full repository queue manager called QM1 in a cluster called CLUSTER1 with a cluster-receiver channel called CLUSTER1.QM1.ALPHA. Every other queue manager can define a cluster-sender channel to this queue manager using the channel name, CLUSTER1.+QMNAME+.ALPHA.

If we use the same naming convention for all your channels, be aware that only one +QMNAME+ definition can exist at one time.

The following attributes on the DEFINE CHANNEL and ALTER CHANNEL commands are specific to cluster channels:

CLUSTER

The CLUSTER attribute specifies the name of the cluster with which this channel is associated. Alternatively use the CLUSNL attribute.

CLUSNL

The CLUSNL attribute specifies a namelist of cluster names.

NETPRTY

Cluster-receivers only.

The NETPRTY attribute specifies a network priority for the channel. NETPRTY helps the workload management routines. If there is more than one possible route to a destination, the workload management routine selects the one with the highest priority.

CLWLPRTY

The CLWLPRTY parameter applies a priority factor to channels to the same destination for workload management purposes. This parameter specifies the priority of the channel for the purposes of cluster workload distribution. The value must be in the range zero through 9, where zero is the lowest priority and 9 is the highest.

CLWLRANK

The CLWLRANK parameter applies a ranking factor to a channel for workload management purposes. This parameter specifies the rank of a channel for the purposes of cluster workload distribution. The value must be in the range zero through 9, where zero is the lowest rank and 9 is the highest.

CLWLWGHT

The CLWLWGHT parameter applies a weighting factor to a channel for workload management purposes. CLWLWGHT weights the channel so that the proportion of messages sent down that channel can be controlled. The cluster workload algorithm uses CLWLWGHT to bias the destination choice so that more messages can be sent over a particular channel. By default all channel weight attributes are the same default value. The weight attribute allows you to allocate a channel on a powerful UNIX machine a larger weight than another channel on small desktop PC. The greater weight means that the cluster workload algorithm selects the UNIX machine more frequently than the PC as the destination for messages.

CONNAME

The CONNAME specified on a cluster-receiver channel definition is used throughout the cluster to identify the network address of the queue manager. Take care to select a value for the CONNAME parameter that resolves throughout your IBM MQ cluster. Do not use a generic name. Remember that the value specified on the cluster-receiver channel takes precedence over any value specified in a corresponding cluster-sender channel.

These attributes on the DEFINE CHANNEL command and ALTER CHANNEL command also apply to the DISPLAY CHANNEL command. Note: Auto-defined cluster-sender channels take their attributes from the corresponding cluster-receiver channel definition on the receiving queue manager. Even if there is a manually defined cluster-sender channel, its attributes are automatically modified to ensure that they match the attributes on the corresponding cluster-receiver definition. Beware that we can, for example, define a CLUSRCVR without specifying a port number in the CONNAME parameter, while manually defining a CLUSSDR that does specify a port number. When the auto-defined CLUSSDR replaces the manually defined one, the port number (taken from the CLUSRCVR ) becomes blank. The default port number would be used and the channel would fail. Note: The DISPLAY CHANNEL command does not display auto-defined channels. However, we can use the DISPLAY CLUSQMGR command to examine the attributes of auto-defined cluster-sender channels.

Use the DISPLAY CHSTATUS command to display the status of a cluster-sender or cluster-receiver channel. This command gives the status of both manually defined channels and auto-defined channels.

The equivalent PCFs are MQCMD_CHANGE_CHANNEL, MQCMD_COPY_CHANNEL, MQCMD_CREATE_CHANNEL, and MQCMD_INQUIRE_CHANNEL.

Omitting the CONNAME value on a CLUSRCVR definition

In some circumstances we can omit the CONNAME value on a CLUSRCVR definition. You must not omit the CONNAME value on z/OS.

On Multiplatforms, the TCP/IP connection name parameter of a cluster-receiver channel is optional. If you leave the connection name blank, IBM MQ generates a connection name for you, assuming the default port and using the current IP address of the system. We can override the default port number, but still use the current IP address of the system. For each connection name leave the IP name blank, and provide the port number in parentheses; for example:

(1415)

The generated CONNAME is always in the dotted decimal (IPv4) or hexadecimal (IPv6) form, rather than in the form of an alphanumeric DNS host name.

This facility is useful when we have machines using Dynamic Host Configuration Protocol (DHCP). If we do not supply a value for the CONNAME on a CLUSRCVR channel, we do not need to change the CLUSRCVR definition. DHCP allocates you a new IP address.

If you specify a blank for the CONNAME on the CLUSRCVR definition, IBM MQ generates a CONNAME from the IP address of the system. Only the generated CONNAME is stored in the repositories. Other queue managers in the cluster do not know that the CONNAME was originally blank.

If you issue the DISPLAY CLUSQMGR command you see the generated CONNAME. However, if you issue the DISPLAY CHANNEL command from the local queue manager, you see that the CONNAME is blank.

If the queue manager is stopped and restarted with a different IP address, because of DHCP, IBM MQ regenerates the CONNAME and updates the repositories accordingly.