| Author |
Message
|
| varmamvn |
 Posted: Mon May 13, 2002 8:51 am Post subject: 2085 for an existing queue, again !!!!! |
 |
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Hi all,
I desperately need your expertise in solving my problem that I am facing in MQ-Series. I have MQ-Series 5.2 installed and configured on a Sun Solaris machine. It was giving occassional errors of 2085, saying that the Remote Queue is not defined. I did not pay much attention as it used to get up and running by itself. There were no changes in my Java program and I observed that the problem is somehow reoccuring without any specific reason. It suddenly occurs and after some time becomes alright by itself. All this happens only when I try from my MQ Java program running on WebLogic 6.1. I face no problem in locating the remote queue and putting sample messages on the queue using amqsput.
Can any one tell us what more settings have to be done for this to work. This error occurrs only while I am trying to put the message in the REMOTE queue. Please do not advice us to use the proper queue name. It’s all been checked, we need more configuration information if anything we have to do. Your earliest and valuable reply would be appreciated.
Regards,
Varma |
|
| Back to top |
|
 |
| kolban |
| Posted: Mon May 13, 2002 9:02 am Post subject: |
|
|
Grand Master
Joined: 22 May 2001
Posts: 1072
Location: Fort Worth, TX, USA
|
| I think it will be useful to understand the overall topology of the system. Where is the remote queue? How is it defined? Is clustering involved? When the error occurs, is it repeatable to capture a trace? Are any errors logged to the queue manager error logs? |
|
| Back to top |
|
|
| varmamvn |
| Posted: Mon May 13, 2002 5:57 pm Post subject: Re: 2085 problem! |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Hi Kolban,
Thanks for offering to help. This is the setup information. There is no clustering involved. The remote Q is on the IBM Mainframe. The remote Q manager serves many remote MQ servers on the Solaris and we never faced this problem earlier. Weblogic runs using the user Id uptlapp1.
# -u = defines the dead letter Queue
/opt/mqm/bin/crtmqm -d SYSTEM.DEFAULT.XMIT.Q -u PORTAL.DLQ -ld /var/mqm/log SGPORTAL_QM.UAT
# Start the Queue Manager
/opt/mqm/bin/strmqm SGPORTAL_QM.UAT
#To run the below mq-series commands given below
/opt/mqm/bin/runmqsc SGPORTAL_QM.UAT <<!
DEF QL(SGGRB_QM.UAT) PUT(ENABLED) GET(ENABLED) USAGE(XMITQ) TRIGGER INITQ(SYSTEM.CHANNEL.INITQ) TRIGDATA(SGPORTAL.SGGRB)
SCOPE(QMGR) DESCR('Transmission Queue')
DEF QL(PORTAL.SG.FROM_GRB.LISTEN) PUT(ENABLED) GET(ENABLED) USAGE(NORMAL) MAXMSGL(530000) MAXDEPTH(10001) NOTRIGGER MSGD
LVSQ(FIFO) SCOPE(QMGR) DESCR('Local Queue') DEFPSIST(YES)
DEF QR(GRB.SG.TO_GRB.REMOTE) RNAME(GRB.SG.TO_EBS_GRB.LISTEN) RQMNAME(SGGRB_QM.UAT) XMITQ(SGGRB_QM.UAT) PUT(ENABLED) SCOP
E(QMGR) DESCR('Remote Queue Definition')
DEF CHL('SGPORTAL.SGGRB') CHLTYPE(SDR) TRPTYPE(TCP) CONNAME('169.165.212.21(8110)') XMITQ('SGGRB_QM.UAT') DESCR('Sender
Channel')
DEF CHL('SGGRB.SGPORTAL') CHLTYPE(RCVR) TRPTYPE(TCP) DESCR('Receiver Channel')
START CHANNEL(SGPORTAL.SGGRB)
end
!
/opt/mqm/bin/setmqaut -m SGPORTAL_QM.UAT -n PORTAL.SG.FROM_GRB.LISTEN -t queue -p uptlapp1 +get +put +browse
/opt/mqm/bin/setmqaut -m SGPORTAL_QM.UAT -n GRB.SG.TO_GRB.REMOTE -t queue -p uptlapp1 +put +get +browse
/opt/mqm/bin/setmqaut -m SGPORTAL_QM.UAT -t qmgr -p uptlapp1 +connect +inq |
|
| Back to top |
|
|
| varmamvn |
| Posted: Mon May 13, 2002 9:50 pm Post subject: 2085 for a valid queue |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Hi all,
I just came to know that this problem has more complicated roots. Basically, I have my Queue manager Q1 along with another Queue manager Q2 on Sun Solaris. Both the remote queue definitions point to the same IBM Host MQ server (different Queue manager alaises). It seems like there is some kind of resource allocation problem. If my QM Q1 has first started sending messages from Weblogic App sever, then QM Q2's attempt to put message on the remote Queue would fail with reason 2085. And similarly if QM Q2 first starts sending messages when Weblogic server restarts, then QM Q1's attempt would get the 2085 error.
Please kindly help as this is driving us crazy. There is nothing in common between the QM's Q1 and Q2. They have their listeners running on different ports.
Thanks in advance
Varma |
|
| Back to top |
|
|
| mrlinux |
| Posted: Tue May 14, 2002 3:20 am Post subject: |
|
|
Grand Master
Joined: 14 Feb 2002
Posts: 1261
Location: Detroit,MI USA
|
How is the application getting the queue name to open, it is possible under
certain circumstances that the queue name it gets is not the correct one ???
_________________
Jeff
IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries |
|
| Back to top |
|
|
| varmamvn |
| Posted: Tue May 14, 2002 5:04 pm Post subject: |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
hi jeff,
no it is not possible for the application to get the wrong queue name, as I can see in my logs the Q manager and the Q name that the application is trying to access. They show the right names.
Thanks for your attention
Varma |
|
| Back to top |
|
|
| amurugan |
| Posted: Tue May 14, 2002 9:51 pm Post subject: |
|
|
Voyager
Joined: 10 Sep 2001
Posts: 75
Location: india
|
hi varma ,
Could it be that the output process on the remote queue manager is limited to 1 . Just a thought . I havent come across this problem so far . Or if you have number of output process then the it might exceeded the max number of channels opened . Is it just two applications or several instances of the 2 applications that are sending the messages . let me know . try to simulate the error by sending 10 messages and 20 messages simulteanulsy and scale to the numbner equiavalent to the max number of channel connections . let me know what happened .
regards,
Murugan |
|
| Back to top |
|
|
| amurugan |
| Posted: Tue May 14, 2002 9:52 pm Post subject: |
|
|
Voyager
Joined: 10 Sep 2001
Posts: 75
Location: india
|
hi varma ,
Could it be that the output process on the remote queue manager is limited to 1 . Just a thought . I havent come across this problem so far . Or if you have number of output process then the it might exceeded the max number of channels opened . Is it just two applications or several instances of the 2 applications that are sending the messages . let me know . try to simulate the error by sending 10 messages and 20 messages simulteanulsy and scale to the numbner equiavalent to the max number of channel connections . let me know what happened .
regards,
Murugan |
|
| Back to top |
|
|
| varmamvn |
| Posted: Wed May 15, 2002 6:07 pm Post subject: 2085 |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Murugan,
Thanks for your interest. I do not think that it is to do with the number of processes. We are usually just testing with one process for QM1 and one for QM2.
I took the trace on QM1 by strmqtrc. It seems that somehow it is attaching itself to QM2 process id. When the 2085 error occurs, the process id for the QM is shown as that of QM2, and not that of QM1 as it should be. And indeed QM2 does not have that remote Queue defined. I think this is an MQ Series or Weblogic bug. But I dont know how to solve this problem. Any help will be greatly appreciated.
Thanks
varma |
|
| Back to top |
|
|
| amurugan |
| Posted: Wed May 15, 2002 6:57 pm Post subject: |
|
|
Voyager
Joined: 10 Sep 2001
Posts: 75
Location: india
|
hi varma ,
IS your messaging applications residing in weblogic . Are they justa stand alone java classes or any EJBS or MDBS . let me know . Since your are telling that both the applications are running under the same process i think this could be race condition in mq channels sending messages . I doubt this . let me know in detail
Regards,
Murugan |
|
| Back to top |
|
|
| amurugan |
| Posted: Wed May 15, 2002 6:57 pm Post subject: |
|
|
Voyager
Joined: 10 Sep 2001
Posts: 75
Location: india
|
hi varma ,
IS your messaging applications residing in weblogic . Are they justa stand alone java classes or any EJBS or MDBS . let me know . Since your are telling that both the applications are running under the same process i think this could be race condition in mq channels sending messages . I doubt this . let me know in detail
Regards,
Murugan |
|
| Back to top |
|
|
| varmamvn |
| Posted: Wed May 15, 2002 10:47 pm Post subject: |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Hi Murugan,
My Java programs are stand alone classes which are running inside the Weblogic app server. The Queue managers are different and their listener ports are different but both their sender channels point to the same IP and port. We have different QM aliases listening on the same port at the remote side.
Regards
Varma |
|
| Back to top |
|
|
| varmamvn |
| Posted: Sat May 18, 2002 4:32 am Post subject: |
|
|
Novice
Joined: 16 Mar 2002
Posts: 21
|
Thanks everyone for trying to solve this problem,
I gave up. I have merged both the 2 QMs into one. I think this is a bug with Weblogic and MQSeries. If you have two different QM's put under one application server JVM, then only one of them would work. Beware!
_________________
Regards
Varma
varmamvn@mailcity.com |
|
| Back to top |
|
|
|
|
