Jean-Pascal Briquet created ARTEMIS-4667:
--------------------------------------------
Summary: JMX Recovery operation triggers ClassCastException
Key: ARTEMIS-4667
URL: https://issues.apache.org/jira/browse/ARTEMIS-4667
Project: ActiveMQ Artemis
Issue Type: Bug
Components: Broker
Affects Versions: 2.30.0
Reporter: Jean-Pascal Briquet
Attachments: msg-replay-classcastexception.png
*Description:*
Impossibility to replay messages from retention folders due to a
ClassCastException.
It happens when replication ha-policy is configured and after the live role
switch back from the backup node to the primary node:
{code:java}
java.lang.ClassCastException : class
org.apache.activemq.artemis.core.replication.ReplicatedJournal cannot be cast
to class org.apache.activemq.artemis.core.journal.impl.JournalImpl
(org.apache.activemq.artemis.core.replication.ReplicatedJournal and
org.apache.activemq.artemis.core.journal.impl.JournalImpl are in unnamed module
of loader java.net.URLClassLoader @73d16e93){code}
See attached screenshot of the Artemis console for additional details.
*Configuration details:*
* Cluster of Artemis primary/backup groups and with replication ha-policy (ZK)
*Reproduction scenario:*
With a brand new primary/backup group with replica :
* Start primary and backup nodes
* The JMX replay command works well on primary
* Stop the primary node, backup becomes live
* Restart the primary, it becomes live again (failback is enabled)
* At this step, the reuse of the replay command on the primary trigger the
ClassCastException
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
