I have a trigger program that runs on the AS/400 (MQS v5.x) that pumps messages to an NT server (also at MQS v5.x). The problem that we've had is that all I/O stalls for several minutes (the exact length of time varies) when a large number of messages are being "put" -- usually over 10,000 but it varies. Looking at the call stack, the program is always stalled at the MQPUT. MQSeries does not error out. It just sits there waiting.
However, there is an error message on the NT event log as follows:
Connection to host 'xxx.xxx.xxx.xxx' closed.
An error occurred receiving data from 'xxx.xxx.xxx.xxx' over TCP/IP.
The connection to the remote host has unexpectedly terminated.
Tell the systems administrator.
The IP address in the error message is the IP of the AS/400. So it looks like it's losing a connection for some reason. After a few minutes, the channel will open up again and messages will start flowing normally.
One thing I did notice and fix was that the sender/receiver channel pair from the AS/400 to the NT server was named the same as the cluster receiver channel from the NT server to the cluster server (an RS/6000).
I created a new sender/receiver pair between the AS/400 and the NT server.
I have not noticed the problem again, but I have not had that high a volume of messages again yet.
Can anyone say for sure one way or the other if the sender/receiver channel naming was the cause of the problem? Or is there some other problem that I need to correct?
I have tried increasing the number of log files and their size to give MQSeries plenty of log space as well.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum