MQSeries.net :: View topic - Sender channel failure between NT and AS/400 -Tricky problem

mountazar

Hi everybody,
we encountered an MQSeries channel communication serious problem with V5.2 on Windows NT 4.0 that is connected to an AS/400 machine that have also WMQ V5.2

The sender channel from NT -> AS/400 failed suddenly and couldn't be restarted even after doing "Reset Channel" and "Resolve Channel".

The communciation between both machines on TCP/IP level is working well (we can ping each machine from the other side). Also the listener on the AS/400 is up and running on the default port 1414 (we can see it clearly in netstat; and we can ensure that by doing a telnet from NT -> AS/400 on port 1414).

Following is the result of the investigation done:

1. Having the sender in the stopped state, if we try to do a 'Ping Channel' (from within MQ Explorer), we get the following error:

"A configuration error was detected in the channel definition or communication subsystem; allocation of a conversation was not possible. (AMQ4057)".

2. In AMQERR01.log for the concerned queue manager, the following error is being logged each time we try to start the sender channel:

08/13/03 09:55:21
AMQ9002: Channel program started.

EXPLANATION:
Channel program 'LNT2LAS' started.
ACTION:
None.
-------------------------------------------------------------------------------
08/13/03 09:55:22
AMQ9203: A configuration error for TCP/IP occurred.

EXPLANATION:
Error in configuration for communications to host 10.0.1.1. Allocation of a
TCP/IP conversation to host 10.0.1.1 was not possible.
ACTION:
The configuration error may be one of the following:

1.If the communications protocol is LU 6.2, it may be that one of the
transmission parameters (Mode, or TP Name) is incorrect. Correct the error
and try again. The mode name should be the same as the mode defined on host
10.0.1.1. The TP name on 10.0.1.1 should be defined.

2.If the communications protocol is LU 6.2, it may be that an LU 6.2 session
has not been established. Contact your systems administrator.

3.If the communications protocol is TCP/IP, it may be that the host name
specified is incorrect. Correct the error and try again.

4.If the communications protocol is TCP/IP, it may be that the host name
specified cannot be resolved to a network address. The host name may not be
in the nameserver.

The return code from the TCP/IP (connect) call was 10049 (X'2741').

Record the error values and tell the system administrator.
-------------------------------------------------------------------------------
08/13/03 09:55:22
AMQ9999: Channel program ended abnormally.

EXPLANATION:
Channel program 'LNT2LAS' ended abnormally.
ACTION:
Look at previous error messages for channel program 'LNT2LAS' in the error
files to determine the cause of the failure.
-------------------------------------------------------------------------------

5. I tried to delete the channel and recreate a new one with the same name and properties, but I got the same errors

6. I tried to create another receiver with a new name on the AS/400 and its corresponding sender on the NT, but still got the same errors.

7. I stopped the queue manager on the NT machine and created another one with a sender channel that points to the same receiver residing on the AS/400 side, and it worked fine. So I recreated all the queues and other channels on the new one and reput the application in live mode.
However I kept the old queue manager with its bad sender channel to work on this issue to know how to resolve it in the future if ever we re-encounter it. Since it's not practical to recreate everything from scratch once such problem occurs.

Please advise,