| Author |
Message
|
| Empeterson |
 Posted: Tue Dec 02, 2003 7:10 am Post subject: Moving queue's in and out of a cluster... |
 |
|
Centurion
Joined: 14 Apr 2003
Posts: 125
Location: Foxboro, MA
|
Is it ok to move a queue in and out of a cluster on the fly, or are there potential pitfalls with doing that? Will all repositories immediately get updated and notified when a queue is removed from the cluster?
_________________
IBM Certified Specialist: MQSeries
IBM Certified Specalist: Websphere MQ Integrator |
|
| Back to top |
|
 |
| Empeterson |
| Posted: Tue Dec 02, 2003 8:02 am Post subject: |
|
|
Centurion
Joined: 14 Apr 2003
Posts: 125
Location: Foxboro, MA
|
Or maybe I should just put-inhibit the queue?
_________________
IBM Certified Specialist: MQSeries
IBM Certified Specalist: Websphere MQ Integrator |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Tue Dec 02, 2003 4:22 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
What are you trying to accomplish?
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| Empeterson |
| Posted: Thu Dec 04, 2003 12:52 pm Post subject: |
|
|
Centurion
Joined: 14 Apr 2003
Posts: 125
Location: Foxboro, MA
|
Basically, we just want to be able to inhibit messages from going to specific queue managers in the cluster. If we put-inhibit one of the queue's, will workload balancing still occur across the remaining queue's in the cluster?
_________________
IBM Certified Specialist: MQSeries
IBM Certified Specalist: Websphere MQ Integrator |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu Dec 04, 2003 12:59 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
Yes, that's the correct and easy way to do it. PUT INHIBIT it. No messages will be sent to it (it gets removed from the round robin), but all other queues function normally.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| mqdev |
| Posted: Tue Dec 09, 2003 1:40 pm Post subject: Cluster changes crash the Full Repos boxes |
|
|
Centurion
Joined: 21 Jan 2003
Posts: 136
|
We have a huge MQ cluster (1000+ QMgrs) implemented in production. We had bad experiences with adding cluster queues (what we were attempting was basically we had 2 instances of a group of cluster queues - we were adding a 3rd instance of the same queues to balance load and other accomplish other tasks...). The Full Repo boxes just ran out of paging space (our env is: AIX 5.2, 4.3 with MQ v5.3 throughout the cluster).
Moral of the story: When attempting cluster updates (esp in large clusters) which are likely to be propagated to the whole cluster, ensure that the Full Repo boxes have enough resources (memory, cpu etc) to handle the updates. |
|
| Back to top |
|
|
|
|
