Hello Guys,
Background:
We have developed a JMS to IDoc scenario in our PO 7.4 system. The sender system is a Weblogic BEA system.
We have set-up Quality of Service="Exactly Once In Order" with Queue ID "EOIOqueueName" in order to enable serialization which is a business requirement. In the Connector Settings we have set-up "Use Message Listener Based Connector" in order to make sure that only one node out of the 4 available in the system are looking-up/fetching incoming messages at a given time. In the CC Monitoring the rest nodes are correctly shown as inactive marked in a yellow triangle.
The issue we are facing is the following:
Whereas sometimes the Communication Channel operates smoothly, i.e. it processes messages at a high average throughput, there are several times wherein then Channel either stops fetching messages and remains stale/idle or the average throughput is very low. In the last 2 cases, there are 3K-4K messages in the sender system's queue which remain unprocessed. In order to overcome the issue a restart of the Communication Channel is needed with which the processing rate is temporarily increased and messages are flowing through again.
Do you have any idea what might be the root cause?
Please note that in earlier PI version 7.0 we weren't facing such an issue.
Kind Regards,
Evaggelos Gkatzios