Hi,
Why would the Application messages be stuck in a Processing status (IN queue) ? There is no error message either. When the service is re-started these get processed. After a little while the same problem after processing a few new messages.
I did not resolve it, but the TSM has managed to fix it by changing the batch queues that the methods processing this message. The Batch queues looked to have been having duplicate messages or too many in it.
+20Do you have any events on FNDCN_APPLICATION_MESSAGE_TAB? I have seen this behaviour in the past that errors in the event action on this table stop the integration cluster nodes from updating the status of the message.
Also, the integration cluster nodes use a connection pool. If you deploy things to the database without restarting the integration cluster these connections in the pool can get corrupted and get stuck in an indefinite cycle of “existing state of package has been discarded” without refreshing themselves which can also lead to messages getting stuck. The solutions here are:
- Do not deploy anything without a delivery
- If you deploy something without a delivery restart the integration cluster afterwards
There are also two bugfixes that will avoid processing to get stuck:
- Apply bugfix 145775 if you are on UPD3 or earlier.
- Apply the latest MWS Binary Patch if you are on MWS BP version 10.16.18.0 (UPD15) or earlier
1 person likes this
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.