[Artemis 2.x upgrade] Undelivered/Lost messages in remote JCA scenario with killing remote servers -------------------------------------------------------------------------------------------------- Key: WFWIP-120 URL: https://issues.jboss.org/browse/WFWIP-120 Project: WildFly WIP Issue Type: Bug Components: Artemis Reporter: Miroslav Novak Assignee: Martyn Taylor Priority: Blocker There undelivered/lost messages in sceario where MDB is connected to remote Artemis cluster in which nodes are killed and restarted. Issue is intermittent. Test Scenario: * start cluster containing node 1 and 3 with deployed inQueue and outQueue ** JDBC store with Oracle 12c is used * start node 2 and 4 * producer sends 5000 small and large messages to inQueue to node-1 * wait for producer to finish * deploy MDBs to node-2 and node-4 which read messages from inQueue and sends them to outQueue in XA transactions * kill and start following nodes in this sequence: 1,3,1,3,1 * wait until all messages are processed * start Consumer which consumes messages form outQueue * Check that number of messages sent to InQueue and received from OutQueue is the same Result: 7 messages were missing in OutQueue.