Scenario:

• There are two Wildfly servers in cluster. Both use JDBC persistence store.
• Producer sends 6000 messages to server 1.
• After all messages are sent, server 2 is killed and restarted.
• Receiver receives messages from server 2.

Expectation: All messages sent by producer are received by receiver from server 2.
Reality: Only 5055 messages from 6000 are received.

Blocker priority was set because it is regression against previous versions of Wildfly.

The issue was hit with Artemis 2.5.0 with https://github.com/jmesnil/wildfly/tree/WFLY-9407_upgrade_artemis_2.4.0_with_prefix (commit 51dd8102f103ccb0470a3cfc8713d3f9bdb1b65d)