Automatic Channel Creation

Automatic channel creation is one of the operational wins that sells IBM MQ clustering: instead of defining sender and receiver channels between every pair of fifty queue managers (thousands of definitions), the queue manager creates CLUSSDR and CLUSRCVR channels when routing first needs them. The first put from QM_APP to a cluster queue on QM_REPORTING may cause MQ to define a cluster sender on QM_APP targeting QM_REPORTING and ensure a matching cluster receiver exists—names often derived from queue manager names and cluster conventions. That magic does not remove security obligations—CHLAUTH, TLS, and firewall rules still govern who connects. This tutorial explains triggers for auto creation, naming patterns, relationship to repository CONNAME and CLUSQM data, CHAD-related queue manager attributes at overview level, manual versus automatic strategies, monitoring DISPLAY CHANNEL, cleaning orphaned auto channels, and troubleshooting BINDING or retry on first cluster route.

Manual Mesh Versus Auto Creation

What Happens on First Cluster Route

1. Application puts to cluster queue requiring remote instance.
2. Workload algorithm selects target queue manager.
3. Local QM checks for usable CLUSSDR to target.
4. If missing, auto process defines CLUSSDR using repository CONNAME and port data.
5. Target QM ensures CLUSRCVR can accept—inbound may auto-define symmetrically.
6. Channel starts, message flows, subsequent puts reuse channel.

Repository Data Drives Addresses

Auto channels use knowledge from cluster repository messages—IP, port, queue manager name. Wrong CLUSRCVR CONNAME on manual seed channel propagates bad data. When relocating a queue manager, update cluster definitions and refresh repository before expecting auto channels to dial correct host. NAT and firewall must allow return path to listener port on each member.

1
2
3
4
5
DISPLAY QMGR CLUSTER CLUSNL
DISPLAY CLUSQMGR('QM_REPORTING') CLUSTER('SALES')
DISPLAY CHANNEL(*) CHLTYPE(CLUSSDR) WHERE(CLUSTER EQ 'SALES')
* After first remote cluster put, new CLUSSDR may appear:
DISPLAY CHSTATUS(*) WHERE(CHLTYPE EQ CLUSSDR)

Security for Auto-Defined Channels

CHLAUTH TYPE(QMGRMAP) or ADDRESSMAP rules must allow partner queue managers. TLS with mutual authentication is common in production—auto channel picks SSL parameters from queue manager and channel template attributes. Deny-by-default CHLAUTH with explicit permits for cluster partner names beats wide open AUTO. Review MCAUSER on cluster channels—overly privileged MCAUSER on CLUSRCVR allows remote puts with that identity context per IBM rules.

Operations: Monitoring and Cleanup

• DISPLAY CHANNEL summary counts—growth when many QMs added.
• Remove obsolete auto channels after decommissioning queue managers—stale defs confuse admins.
• Alert on CLUSSDR in RETRY—firewall or CONNAME errors block cluster.
• Document expected channel naming pattern for help desk.
• Capacity plan file descriptors and channel instances on hub queue managers.

When to Pre-Define Instead

Regulated environments may require every channel in Git MQSC before production. Pre-define critical CLUSSDR with explicit HBINT, SSLCIPH, and MAXMSGL. Auto creation handles long tail of rare pairs. Hybrid: pre-define full repository interconnects; allow auto among partial application members.

Explainer: Roads Built on First Trip

Automatic channel creation is like paving a road the first time a delivery truck needs to reach a new warehouse—later trucks use the same road. Manual pre-definition is paving all roads before any truck exists.

Explain Like I'm Five: Automatic Channels

When your school needs to call another school for the first time, the phone company sets up the phone line automatically instead of you writing hundreds of line orders in advance.

Practice Exercises

Exercise 1

After first cluster put, which DISPLAY commands find new CLUSSDR?

Exercise 2

List three CHLAUTH considerations for auto channels.

Exercise 3

When would you pre-define all cluster channels?

Frequently Asked Questions

Test Your Knowledge