| Author |
Message
|
| vijayaleads |
 Posted: Wed Apr 16, 2008 11:47 am Post subject: Viewing Clsuterqmgr and Clusterqueue. |
 |
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
Hi all,
I have encountered a strange problem on my MQ 6.0 Cluster Environment on Solaris 5.9.
We have a 2 qmgr cluster, with 1 cluster queue in each queue manager.
After some 4 days of idle operation, I logged on that box and when I did a DISPLAY QCLUSTER(*), I found 1 cluster queue instance of one qmgr and 50 same queue names of the other queue manager. I felt my qmgr is getting the wrong notifications from the other full repository. I have removed the queue manage from the cluster, suspended it, disabled the cluster receiver and then again added the qmgr to cluster with enabling the cluster attributes on the CLUSRCVR channels. Finally did a Refresh Cluster and found everything fine.
Today when I did a DISPLAY CLUSQMGR(*), I found 6 instances of the same queue manager names. Found nothing suspicious in my logs. Followed the same procedure again and then noticed everything seems later.
I did not understand the core reason for such a behavior.
Regards,
Vijaya
Last edited by vijayaleads on Thu Apr 17, 2008 6:58 am; edited 1 time in total |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Wed Apr 16, 2008 2:25 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
ask as well for the clusterqmgr's qmid
you should probably request a
dis clusqmgr(*) qmid
The cluster name should be in the output as well.
If name + QMID show duplicates in the same cluster open a PMR...
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 6:53 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
Its again a mess today morning. I felt everything was fine after I removed the qmgr from cluster and added it again.
when I do a display clusqmgr(*) qmtype qmid, I find the same queue manager name with the QMID a bit different.
The name of the qmgr is SASA02MD.
display clusqmgr(*) qmtype qmid
8 : display clusqmgr(*) qmtype qmid
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_08.40.47)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_07.22.24)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_05.11.41)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_04.49.03)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_04.34.29)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_04.21.04)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-17_03.56.40)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-16_22.10.37)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SASA02MD) CHANNEL(TO.SASA02MD)
CLUSTER(CLASASD) QMID(SASA02MD_2008-04-16_02.58.15)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(SINC001D) CHANNEL(TO.SINC001D)
CLUSTER(CLASASD) QMID(SINC001D_2008-04-15_12.38.47)
QMTYPE(REPOS)
AMQ8441: Display Cluster Queue Manager details.
CLUSQMGR(TESTQMGR) CHANNEL(TO.TESTQMGR)
CLUSTER(CLASASD) QMID(TESTQMGR_2008-04-16_21.08.29)
QMTYPE(REPOS)
There is an example in the Clusters documentation which says if we see multiple instances of the same queue manager then we need to do a Reset cluster with action force remove.
Ex:2 : reset cluster(CLASASD) qmid('SASA02MD_2008-04-16_02.58.15') action(FORCEREMOVE) queues(yes)
AMQ8559: RESET CLUSTER accepted.
Seeing the QMID's I see they are all different instances of different times of a day.
How should I choose the qmgr and its perticular qmid while resetting the cluster. I mean which latest qmgr &qmid should I leave.
As told before should I open a PMR here?..
Please  here.
Thanks,
Vijaya |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 7:06 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
| I see the count is going up with the new time extensions added to the QMID. |
|
| Back to top |
|
|
| mvic |
| Posted: Thu Apr 17, 2008 7:17 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| vijayaleads wrote: |
Please here.
Thanks,
Vijaya |
- Are there any relevant entries in /var/mqm/qmgrs/QMNAME/errors/*.LOG ?
- Are there any recent files in /var/mqm/errors ?
- What MQ code level is the queue manager running? |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 7:30 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
Ok...Here is what I see something related to my issue.
The Queue 3000RSP is hosted by the queue manager SASA02MD.
04/17/08 08:39:34 - Process(4406.1) User(mqm) Program(amqrrmfa)
AMQ9407: Cluster queue SSIF13000RSP is defined inconsistently.
EXPLANATION:
The definition of cluster queue 3000RSP on the queue manager with UUID
SASA02MD_2008-04-17_08.40.47 has different DEFPRTY, DEFPSIST and DEFBIND values
from the definition of the same cluster queue on the queue manager with UUID
SASA02MD_2008-04-16_02.58.15. Both definitions now exist in the local
repository. All definitions of the same cluster queue should be identical. In
particular, problems arise if your applications rely on a queue default value
which is defined inconsistently to determine messaging behavior. This applies,
for example, if the applications open a cluster queue with option
MQOO_BIND_AS_Q_DEF. If different instances of the queue have different DEFBIND
values the behavior of the message transfer differs depending on which instance
of the queue is selected when it is opened. In general the instance selected
varies across opens.
ACTION:
For each inconsistency decide which of the values is the correct one. Alter the
definitions of cluster queue SSIF13000RSP so that all definitions have correct
DEFPRTY, DEFPSIST and DEFBIND values.
Don't how this is related to multiple cluster qmgrs.
Are there any recent files in /var/mqm/errors ?
No..Nothing.
What MQ code level is the queue manager running?
Its 6.0.
Thanks,
Vijaya |
|
| Back to top |
|
|
| mvic |
| Posted: Thu Apr 17, 2008 7:36 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| vijayaleads wrote: |
| Its 6.0. |
Thanks, can you also post output from "dspmqver" please? |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 7:40 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
$ dspmqver
Name: WebSphere MQ
Version: 6.0.0.0
CMVC level: p000-L050519
BuildType: IKAP - (Production) |
|
| Back to top |
|
|
| mvic |
| Posted: Thu Apr 17, 2008 7:44 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| vijayaleads wrote: |
$ dspmqver
Name: WebSphere MQ
Version: 6.0.0.0
CMVC level: p000-L050519
BuildType: IKAP - (Production) |
I suggest installing the latest fix pack 6.0.2.3. There is nearly a 3-year history since that code level, in which many fixes have been produced in the area of MQ clustering, and other areas too. Start here: http://www.ibm.com/software/integration/wmq/support/ |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 8:10 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
| Its a new implemetation and don't know how soon we would install the new fix pack. |
|
| Back to top |
|
|
| mvic |
| Posted: Thu Apr 17, 2008 8:22 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| vijayaleads wrote: |
| Its a new implemetation and don't know how soon we would install the new fix pack. |
Since it's a new implementation it is good to specify and test with the latest fix pack, to ensure you get the benefit of all available fixes. There are plenty of issues fixed since 2005: some found by customers and reported in PMRs, and some found internally. Maybe the issue you are seeing is already fixed? So there are positive reasons to use up-to-date code, and finally one more - it makes it so much quicker and easier for IBM support to investigate any new issue you might find in the future.
The fix packs page is here: http://www.ibm.com/support/docview.wss?rs=171&uid=swg27006037 |
|
| Back to top |
|
|
| vijayaleads |
| Posted: Thu Apr 17, 2008 10:09 am Post subject: |
|
|
Novice
Joined: 10 Jan 2008
Posts: 22
|
Thanks mvik...I have told my manager about this issue and I was told we will be installing the new service pack very soon.
Right now, I assumed that my qmgr is receiving the wrong notifications and for this I disabled my CLUSRCVR channel and reenabled it and broght it back to the cluster.
When i do display clusqmgr(*) qmid, I see only the required qmgrs.
Thanks,
Vijaya |
|
| Back to top |
|
|
|
|
