[ARTEMIS-2892] Replication stopped working after many failovers

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.11.0
• Fix Version/s: 2.36.0
• Component/s: Broker
• Labels: replication
• Assignee: Clebert Suconic
• Reporter: Chris Oman
• Created: 01/Sep/20 18:06
• Updated: 24/Jul/24 16:48
• Resolved: 24/Jul/24 16:48

Description

We have 2 servers that we have been testing the failover capabilities with. We have encountered a situation where replication has completely stopped, but the master is working properly otherwise.

When the backup attempts to connect, the master has this in its logs.

2020-09-01 11:23:31,177 WARN  [org.apache.activemq.artemis.core.server] AMQ222013: Error when trying to start replication: ActiveMQInternalErrorException[errorType=INTERNAL_ERROR message=AMQ229006: journals are not JournalImpl. You can''t set a replicator!]
        at org.apache.activemq.artemis.core.persistence.impl.journal.JournalStorageManager.startReplication(JournalStorageManager.java:628) [artemis-server-2.11.0.jar:2.11.0]
        at org.apache.activemq.artemis.core.server.impl.SharedNothingLiveActivation$2.run(SharedNothingLiveActivation.java:178) [artemis-server-2.11.0.jar:2.11.0]
        at java.lang.Thread.run(Thread.java:748) [rt.jar:1.8.0_201]

How can this happen? Up until this point, replication was working properly through all of the failovers (10+).

I have attached the logs files from both systems.

Issue Links

• links to: GitHub Pull Request #4978