[
https://issues.apache.org/jira/browse/ARTEMIS-2388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16872621#comment-16872621
]
Justin Bertram commented on ARTEMIS-2388:
-----------------------------------------
How are the cluster-connections configured? If they're using the default
reconnect-attempts (i.e. {{-1}}) then I think this is the expected behavior
since from what I recall nodes from the topology won't be removed until the
cluster-connection gives up on reconnecting. Can you test this again with
something like {{<reconnect-attempts>1</reconnect-attempts>}}?
> HA replication diagram does not remove (update) missing (killed) slaves in
> topology
> -----------------------------------------------------------------------------------
>
> Key: ARTEMIS-2388
> URL: https://issues.apache.org/jira/browse/ARTEMIS-2388
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: Broker, Web Console
> Affects Versions: 2.9.0
> Reporter: Michal Toth
> Priority: Major
>
> When we have 3 HA pairs (which are properly distinguished by master/slave
> colors ARTEMIS-1825 fixed that), there is a problem with displaying properly
> slave nodes.
> I have killed 2 out of 3 slaves and they keep being displayed as active
> slaves for given masters. This is imo a bug.
> First I thought it's a bug in diagram only, but Howard Gao told me, that
> diagram now reflects the true nature of provided topology. Is that some kind
> of feature or bug?
> I think this is a bug.
> This was discovered while checking upstream PR for 2.10.0 from Howard Gao.
> https://github.com/apache/activemq-artemis/pull/2710
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
