Collect IBM MQ MustGather data to solve performance problems

• IBM MQ Performance Reports

If you believe there is a defect or regression which is preventing MQ from performing in line with its reports, collect the same information you would for a hang or high CPU usage problem. If you have both a working system and a slow system, collect data from both for comparison. In addition, please ensure your system is up-to-date on operating system patches and firmware:

• MQ MustGather for Hangs and High CPU Usage on Linux, UNIX, Windows and IBM i
• MQ MustGather for Hangs and High CPU Usage on HPE NonStop

Additional Information for Performance Problems

Please provide as much detail as you can about the hardware and software on the systems:

1. A description of the system, its CPUs and available memory


2. Disk information for the MQ filesystems (e.g. SCSI, NAS, SAN, etc.)


3. Other software running on the system (e.g. databases, application servers, etc.)


4. Other traffic on your network or SAN which could affect MQ


5. Firewalls and other network devices which could affect MQ


6. The specific channels, queues, topics, and applications involved


7. Details on how you are measuring MQ performance


8. The full set of performance numbers you recorded

Performance Problems Involving Multiple Systems

If your performance problem involves queue managers, clusters, and applications across multiple systems, you must provide sufficient detail to IBM support about each system. For example, if you were seeing excessively slow round-trip messaging performance involve request and reply messages flowing through an MQ network, you would need to provide:

1. Information about all systems involved, including the CPU, memory, network, disk, operating system and MQ versions on each.


2. The names and configuration of all topics, queues, channels, queue managers and clusters used to send requests and replies through the full round-trip.


3. Details of the back-end application and its rate of processing, as well as any queue depths or channel statistics indicating the location of the bottleneck.


4. Observations on network performance, including packet traces on both ends of a connection if you suspect excessively slow transfer rates.


5. Information on any changes to MQ, the systems, the network, or the applications made before the problem started.


6. A list of any errors or unusual conditions you saw which may be relevant to the problem. 

Please refer to the IBM Software Support Handbook for more information on working with IBM support and on the types of problems which are handled by IBM support and by IBM services engagements.

Back to top