IBM Support

New queue introduced in Refresh Pack WebSphere MQ 7.0.1 called SYSTEM.CLUSTER.HISTORY.QUEUE

Troubleshooting


Problem

You notice that there is a new queue called SYSTEM.CLUSTER.HISTORY.QUEUE. The WebSphere MQ 7.0.1 documentation only states that it is used to store clustering historical information for service purposes. It also states that it can be PUT (DISABLED) to stop history collection.

Symptom

You would like more information about this queue.

Resolving The Problem

SYSTEM.CLUSTER.HISTORY.QUEUE is used to store the history of cluster state information. This information is written to the queue when REFRESH CLUSTER is issued, just before the refresh command is processed. This collects the state of the repository cache before it is refreshed. The history may give vital information about any cluster related problems, before it is lost by issuing REFRESH CLUSTER.

Saving the contents of the cluster repository cache into the SYSTEM.CLUSTER.HISTORY.QUEUE before flushing the information using REFRESH CLUSTER is a serviceability feature introduced through WebSphere MQ 7.0.1.0. This operation is invoked only when REFRESH CLUSTER is issued, and therefore it has no performance overhead on the functioning of the queue managers.

There is no specific action required by the customer. If the queue is left untouched (Note: the queue is PUT enabled by default**), the history information will be collected transparent to the users. From serviceability perspective, we do not expect customers to disable the PUT on the SYSTEM.CLUSTER.HISTORY.QUEUE.

History messages are persistent messages. The messages expire after 90 days. However, they remain for another 60 days after that. The queue should never become full as these messages are written only when the REFRESH CLUSTER command is executed and customers should not issue REFRESH CLUSTER until consulting IBM support.

This feature was introduced to improve serviceability of the product.
The information stored in the history queue is for the consumption of IBM support. The information can not be used to rollback the cluster repository state.



**Differences for z/OS-
In z/OS for the SYSTEM.CLUSTER.HISTORY.QUEUE the default is PUT(DISABLED) because a REFRESH CLUSTER dumps the whole of the cache into the S.C.H.Q, which might fill the pageset, potentially causing all sorts of other problems. Therefore in z/OS dump the CHINIT before issuing the REFRESH command as that gives support everything they need in an easy-to-read format. instead of using the SYSTEM.CLUSTER.HISTORY.QUEUE

[{"Product":{"code":"SSFKSJ","label":"WebSphere MQ"},"Business Unit":{"code":"BU053","label":"Cloud & Data Platform"},"Component":"Documentation","Platform":[{"code":"PF002","label":"AIX"},{"code":"PF010","label":"HP-UX"},{"code":"PF016","label":"Linux"},{"code":"PF027","label":"Solaris"},{"code":"PF033","label":"Windows"},{"code":"PF035","label":"z\/OS"}],"Version":"8.0;7.5;7.1;7.0.1","Edition":"","Line of Business":{"code":"LOB45","label":"Automation"}}]

Product Synonym

WMQ MQ

Document Information

Modified date:
15 June 2018

UID

swg21429699