MQSeries.net :: View topic - MQ connection in bind state reset by a second connection

Mr Butcher

From time to time we see this connection issue, but we do not find any reason for it nor how to solve it.

MQSeries 5.3.1 on z/OS talking with MQSeries on distributed plattforms (V5, V6 on Windows, dont know exactly because these are customer systems)

In our MQSeries log on z/OS we sometimes see channels abending with TCP 461 and 467 returncodes (connection reset by peer / timeout). This is not our problem, but it sometimes causes this situation:

After the channels in both directions abended the connection from the customer to us becomes active again, the connection from us to the customer is in retry / bind state. We can not connect, no matter what we try (stop / force channel, start channel, ...), we always are stuck in "bind" - status at every manual or automatic retry interval.

Now i use a "dummy" channel on a different queuemanager, put in the IP address of the channel that is in bind state and start the dummy channel. Immediately i get a "remote channel not defined" (as expected) from the customers queuemanager.

I now return to the original channel that is in retry / bind state, stop it, start it, and it works.

We already investigated the ADOPTMCA settings at the remote end, the customer said that it is active.
What network or MQ issue can cause this bind state that is reset by just using a different channel (from a different queuemanager)?

This issue is not related to a single customer, we experience this with various remote systems.....

Any hints welcome....
_________________
Regards, Butcher