You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "Ryan Yeats (Jira)" <ji...@apache.org> on 2022/04/27 16:27:00 UTC
[jira] [Commented] (ARTEMIS-3064) Disable useTopologyForLoadBalancing for non HA federation connections between clusters
[ https://issues.apache.org/jira/browse/ARTEMIS-3064?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17528878#comment-17528878 ]
Ryan Yeats commented on ARTEMIS-3064:
-------------------------------------
[~jbertram] I am wondering if it would have been better to bring the circuit breaking configuration (useTopologyForLoadBalancing and
circuit-breaker-timeout) into the connector URI instead. So that it is common to Artemis' existing configuration for connections and not tied to federation.
I am thinking this because as I tried to use federation and i did not find it intuitive to realize topology is turned off by default since ha is off by default in federation.
It is also unintuitive that ha=true in a connector UI would be ignored and i would have to use an undocumented <ha>true</ha> in my federation configuration.
> Disable useTopologyForLoadBalancing for non HA federation connections between clusters
> --------------------------------------------------------------------------------------
>
> Key: ARTEMIS-3064
> URL: https://issues.apache.org/jira/browse/ARTEMIS-3064
> Project: ActiveMQ Artemis
> Issue Type: Improvement
> Components: Federation
> Affects Versions: 2.16.0
> Reporter: Gary Tully
> Assignee: Gary Tully
> Priority: Major
> Fix For: 2.17.0
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> The core client respects the topology for loadbalancing by default. A federation connection configures and reuses a server locator via it's circuit breaker.
> With ha = false, it looks like we should disable useTopologyForLoadBalancing such that the circuit reconnect gets back to the original broker even when clustered.
> With ha=true, the reconnect can be to any broker in the cluster in the normal way.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)