Hi ,

I am using JBoss EAP 5.1 server in clustered in environment for my application. Let me give a brief idea about the structure:

Node A
Master (Performs write)

Node B
Slave - Performs write via Queuing message to Master and performs search on Local copy

Node C
Slave - Performs write via Queuing message to Master

Node D
Slave - Performs write via Queuing message to Master

The hibernatesearch queue is set up as a clustered queue using services 1) jboss.messaging:service=ServerPeer and 2) jboss.messaging:service=PostOffice
The queue is also set up as persistant and is backed in oracle DB.

The problem that is happening is few Index update messages are remaining in DB which are never processed. Even a re-start is not pushing them through. The new incoming message are going through while the old ones are still there in DB.
This is causing a considerable pile up over a period of time thus loosing valuable index updates.

This problem is not happening for other Queues which is forcing me to think that something might be wrong with Hibernate Search.
Please help.

Regards
Arnab

Hi,
Can you tell which type of messages (insert,update,delete) are stuck on the queue? Is there anything unusual in the logs? How exactly does your JMS configuration look like? Or even better can you create a test case?

--Hardy