PLEASE HELP - Inconsistent AQ Behavior!

Question

HELP!!!! HELP!!!! HELP!!!!
We are using Java 1.3 and Oracle 8.1.7's aqapi.jar. A frontend application enqueues messages OK. The other "side" of the queue dequeues these messages MOST of the time, but not all the time. If we change it from blocking forever to polling (timing out every 10 seconds and reconnecting), this still happens.
We are seeing this with RAW payloads (as well as Object payloads).
Storage space should not be an issue nor should network traffic with regards to the staging environment box we are dealing with.
Here are the enqueue and dequeue options we are using:
// Use most default Enqueue options
AQEnqueueOption enqueueOption = new AQEnqueueOption();
enqueueOption.setVisibility(AQEnqueueOption.VISIBILITY_IMMEDIATE)
AQDequeueOption dequeueOption = new AQDequeueOption();
dequeueOption.setDequeueMode(AQDequeueOption.DEQUEUE_LOCKED);
dequeueOption.setNavigationMode(AQDequeueOption.NAVIGATION_FIRST_MESSAGE);
dequeueOption.setWaitTime(AQDequeueOption.WAIT.FOREVER); //(or set with a time)
dequeueOption.setVisibility(AQDequeueOption.VISIBILITY_ONCOMMIT);
We are using CorrelationIDs to make sure we are getting the proper response to our request.
I know JDBC drivers are not certified for JDK 1.3, but...why would it work most of the time? An example would be enqueuing 71 similar messages using the same calls and only 65 being dequeued on the other side. This example happened all within the span of 5 minutes. None of the Java calls we are using are synchronized and the enqueuing message expirations are set high enough(60+ seconds).
We ARE NOT using JMS. We are NOT using OCI. It queue is setup as a single consumer queue. Pure Java to AQ over JDBC. There is a separate in and out queue table each with one queue (so we have separated the connections when one is committed, the other queue is not affected).
If you need more information to help with this, please let me know. We've been trying at this for MONTHS!! HELLLPP!!!!
null

Answer 1

We are seeing similar behavior and would greatly appreciate any help.
Our application enqueues messages without any problems, and the registered message listener receives the sent messages just fine, at first... After a while (can be minutes, hours, days), the listener just stops receiving messages. We can still enqueue messages, but the are not dequeued.
If we stop our application and start it again, we DO receive all of the messages that were previously stuck in the queue. However, unfortunately we are not able to receive any new messages.
If we shutdown Oracle and restart it, everything works fine again... at least until AQ wedges again.
Oddly enough, the messages are in the 'READY' state, so you would think that the listener would receive them, but it does not.
All indicators are that this is an Oracle AQ problem, as the only way to make things work correctly again is to stop and restart Oracle.
We checked for trace (.trc) files when this problem occurred, but no new ones where generated.
Any thoughts, suggestions, workarounds, tracing options that we can set, queries we can run to get more debug info, etc would be greatly appreciated.
Please feel free to email me directly: [email protected] We would prefer to use Oracle AQ, but are on the verge of trying our luck with WebLogic JMS...
Thank you in advance!