-
Bug
-
Resolution: Done
-
Blocker
-
7.0.0.ER2 (Beta), 7.0.0.ER6
-
None
Scenario
There are 4 EAP servers. Cluster A with node 1 and 3 is started and queues InQueue and OutQueue are deployed to both of them. Cluster B with nodes 2 and 4 is started. Start producer which sends 1000 messages (mix of small and large messages) to InQueue. Once producer finishes, deploy MDB which reads messages from InQueue and sends to OutQueue (in XA transaction) to cluster B (node 2,4). When MDBs are processing messages, cleanly shutdown and restart nodes in following sequence 1,3,1,3,1.
Wait until all messages are processed and consume messages from OutQueue.
Test procedure
1. start cluster one containing node 1 and 3 with deployed inQueue and outQueue
2. start cluster two containing node 2 and 4
3. producer sends 1000 small and large messages to inQueue
4. wait for producer to finish
5. deploy MDBs to node-2 and node-4 which read messages from inQueue and sends them to outQueue in XA transactions
6.cleanly shutdown and start following nodes in this sequence: 1,3,1,3,1
7. wait until all messages are processed
8. start Consumer which consumes messages form outQueue
Consumer receives messages but sometimes one is missing. It does not happen always, but sometimes there are lost messages. This issue was hit only when jms server (nodes 1, 3) are cleanly shutdown, not killed.
- is blocked by
-
JBEAP-3577 Upgrade Artemis to 1.1.0.wildfly-014
- Closed
-
JBEAP-2499 Upgrade Artemis to 1.1.0.wildfly-011
- Closed
- is related to
-
JBEAP-2100 Lost large message in cluster with network failures
- Closed
-
ARTEMIS-432 Loading...
-
ARTEMIS-441 Loading...