Author |
Message
|
learnmqnow |
Posted: Tue Oct 04, 2005 10:59 pm Post subject: Solved!!! MQ clustering across a different network |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
Hi.
I have a cluster setup which includes 3 different machines 2 Windows and one Linux. Linux machine and one windows machine is in the same network located at Bangalore. The other Windows machine is located at Mumbai.
Linux machine has 5 queue managers of which two are repositories. I have created cluster queues in these 5 queue managers and the cluster is working fine. From the windows machine in the same network i joined the linux cluster and this also succesfully joined the cluster and i can see all the cluster queues.
Then i tried to join the Mumbai windows queue manager to the Linux queue manager. But it's not clustered properly. When i issue
dis clusqmgr(*) from linux machine, it's showing the queue manager name at mumbai. But when i issue dis qcluster(*) its not showing the cluster queue in mumabi queue manager.
When issuing dis clusqmgr(*) from the Mumbai windows mc, it's showing some system.temp...queue manager.Whereas from the explorer window, all the queue managers in the cluster can be seen.But again, the cluster queues in the linux queue managers are not visible there.
Is this a network issue..Access is enabled only for the listening port(1414) for the queue manager. ie, 'telnet MumbaiIP 1414' is working fine.
Pls reply ASAP.
Thanks in advance.
Last edited by learnmqnow on Thu Nov 03, 2005 2:04 am; edited 1 time in total |
|
Back to top |
|
 |
ChrisW |
Posted: Wed Oct 05, 2005 2:06 am Post subject: |
|
|
Voyager
Joined: 20 May 2001 Posts: 78 Location: UK
|
Have you remembered that you cannot see cluster queues from a partial repository Queue Manager until you try to use them? |
|
Back to top |
|
 |
learnmqnow |
Posted: Wed Oct 05, 2005 3:08 am Post subject: |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
Yes. I know that cluster queues will not be visible unless you put some message.
I tried to put the message, but it's giving 8085 error.(Unknown object name).
Anything else i need to check
Thanks. |
|
Back to top |
|
 |
happyj |
Posted: Wed Oct 05, 2005 3:54 am Post subject: |
|
|
Voyager
Joined: 07 Feb 2005 Posts: 87
|
system.temp.qmgr... means that there is a problem
in your CLUSRCVR or CLUSSDR definitions.
Check the local logs and the logs on the full repository qm.
this is in the cluster manual - troubleshooting section |
|
Back to top |
|
 |
learnmqnow |
Posted: Wed Oct 05, 2005 4:08 am Post subject: |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
Which is the cluster manual? I have a queue manager cluster book in the MQ documentation. But i didnt find any troubleshooting section in that. And the program (formatLog) for converting binary log files to text is not working for me. Can u pls help me out in this. where can i find the logs for MQ in linux and how to view those?
Thanks. |
|
Back to top |
|
 |
Nigelg |
Posted: Wed Oct 05, 2005 4:34 am Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004 Posts: 1046
|
system.temp...qmgr may mean that there is a problem in the cluster defs, or that the cluster qmgr cannot be contacted becuase the network is down.
Either way, the cluster channels have not started such that the cluster qmgrs can communicate. Check the qmgr CLUSRCVR and CLUSSDR defs to see whether they are correct. particularly CONNAME for the correct host and port, and check the qmgr error logs for channel errors. _________________ MQSeries.net helps those who help themselves.. |
|
Back to top |
|
 |
happyj |
Posted: Wed Oct 05, 2005 6:29 am Post subject: |
|
|
Voyager
Joined: 07 Feb 2005 Posts: 87
|
the troubleshooting section of the cluster manual is in chapter 9
it is easy to get this error with a typo in the channel definitions.
you need to check file AMQERR01.LOG for both the partial and full
repository QMGRS - the directory structure for unix and windows is
explained in the system adminstatation manual.
post any errors if you need help with them. |
|
Back to top |
|
 |
learnmqnow |
Posted: Wed Oct 05, 2005 10:12 pm Post subject: |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
The cluster channel definitions are correct. we are using mirrored ips for connection.
I checked the error log in the full repository. I am pasting the relevant portion of error log here.
-------------------------------------------------------------------------------
10/05/2005 03:48:47 AM
AMQ9202: Remote host '192.168.40.31 (1414)' not available, retry later.
EXPLANATION:
The attempt to allocate a conversation using TCP/IP to host '192.168.40.31
(1414)' was not successful. However the error may be a transitory one and it
may be possible to successfully allocate a TCP/IP conversation later.
ACTION:
Try the connection again later. If the failure persists, record the error
values and contact your systems administrator. The return code from TCP/IP is
110 (X'6E'). The reason for the failure may be that this host cannot reach the
destination host. It may also be possible that the listening program at host
'192.168.40.31 (1414)' was not running. If this is the case, perform the
relevant operations to start the TCP/IP listening program, and try again.
-------------------------------------------------------------------------------
But when i telnet to this IP using pert 1414, it's connecting. connectivity is enabled only to the listening port of the queue manager, all others are blocked.
requesting ur help.
Thanks. |
|
Back to top |
|
 |
Nigelg |
Posted: Wed Oct 05, 2005 11:56 pm Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004 Posts: 1046
|
The error msg is quite clear. TCP is returning an error 110, so you have to find out why that is and fix it. This is not a WMQ problem, it is in the TCP or network stack. _________________ MQSeries.net helps those who help themselves.. |
|
Back to top |
|
 |
learnmqnow |
Posted: Thu Oct 06, 2005 1:04 am Post subject: |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
Hi
can u pls tell me how to find details about these tcp/ip error codes?
Since telnet is working , my network admin is assuming that connectivity is fine.
Thanks. |
|
Back to top |
|
 |
learnmqnow |
Posted: Thu Oct 27, 2005 1:03 am Post subject: |
|
|
Novice
Joined: 04 Oct 2005 Posts: 13 Location: Bangalore, India
|
Hi
The problem is solved. Thought i'll post the details as it will b useful for someone. Problem was with IP itself. As i mentioned before i am using a mirrored IP to connect to a different network.I'll explain in detail.
Machines A and B are in two different networks.
Machine A has say, the IP 192.168.17.12
Machine B has the IP 172.10.10.36
A and B should be part of a single cluster. Access was enabled between these networks only for port 1414 were the queue manager was listening.
But for connecting from A to B and viveversa, we were given mirrored IPs which are as below.
From Machine A, Machine B 's IP is 192.168.17.36 ie. it appears to be in the same network.
Similarly from Machine B, Machine A's IP is 172.10.14.12.
What i did initially was, i created a receiver channel in A 'TO.A' with ip 192.168.17.12 which was conceptually correct. But while joining cluster, machine B will automatically create sender channels with this IP which it cannot recognise. So i changed the receiver channel IP to 172.10.14.12 which is the mirrored IP. Initially when we create the channel, it wont be able to recognise the IP and channel will be in retrying state, but once a message comes from machine B to A, it will start running.
Hope i've made it clear.
ie, in the network 192.x.x.x i created a receiver channel with 172.x.x.x for enabling machines from second network to join the cluster. Similarly in network 172.x.x.x i created a receiver channel with 192.x.x.x.
Thanks all for ur help. |
|
Back to top |
|
 |
hguapluas |
Posted: Thu Oct 27, 2005 9:08 am Post subject: |
|
|
Centurion
Joined: 05 Aug 2004 Posts: 105 Location: San Diego
|
Depending on your network/router/firewall configurations, it might have been easier (or more complicated) to have just configured NATs on both sides. Then you wouldn't have had to make those changes in the channel defs.
Just a thought.
Cheers, |
|
Back to top |
|
 |
fjb_saper |
Posted: Thu Oct 27, 2005 12:26 pm Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
This is why I usually like all that stuff to be resolved through the DNS.
So you would have used conname('mymachinename(1414') and not have to worry about all these contact admin with the IP.
Enjoy  |
|
Back to top |
|
 |
|