[ https://issues.apache.org/jira/browse/ARTEMIS-2511?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Justin Bertram updated ARTEMIS-2511: ------------------------------------ Component/s: (was: ActiveMQ-Artemis-Native) > Got duplicated messages when using ha-policy-autobackup > ------------------------------------------------------- > > Key: ARTEMIS-2511 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2511 > Project: ActiveMQ Artemis > Issue Type: Bug > Affects Versions: 2.10.1 > Reporter: Ning Kang > Assignee: Clebert Suconic > Priority: Major > > I tried [this > example|https://github.com/apache/activemq-artemis/tree/master/examples/features/ha/ha-policy-autobackup], > but found a bug of duplicated messages and 2 other issues. > When finish step 16 in HAPolicyAutoBackupExample, if I restart server1 and > consumer1, then all scaled down messages that already consumed in step 16 by > consumer0 will be received again by consumer1. > To reproduce this bug, just follow the setup in the {{ha-policy-autobackup}} > example, and add the follow steps after step 16 in > {{HAPolicyAutoBackupExample.java}}: > - 16a: restart server 1 > - 16b: restart consumer1 > - 16c: consumer1 will receive all scaled down messages that already consumed > by consumer0 > Is this a bug of Artemis HA, or there is some extra configurations need to be > made in {{broker.xml}} to avoid scaled down messages to be consumed again? > *Another 2 issues I found related to this example are:* > 1. After restart server1, both server0 and server1 will continuously do > RequestBackupVote actions and never stop, is this normal? See example log > {code:java} > 2019-10-03 17:26:37,916 INFO [org.apache.activemq.artemis.core.server] > AMQ221063: Sending quorum vote response: RequestBackupVote [backupsSize=1, > nodeID=1905cb70-e508-11e9-8290-00505680c5d7, backupAvailable=false] > 2019-10-03 17:26:38,926 INFO [org.apache.activemq.artemis.core.server] > AMQ221062: Received quorum vote request: RequestBackupVote [backupsSize=-1, > nodeID=null, backupAvailable=false]{code} > 2. Even if server1 is restarted, the loadbalancer is not working any more. > All messages from producer0 will only be sent to server0, instead of half of > messages sent to server0, another half sent to server1. -- This message was sent by Atlassian Jira (v8.20.1#820001)