ActiveMQ Artemis客户端故障转移未优先连接对应备份代理的问题求助

我正在测试ActiveMQ Artemis集群的客户端故障转移功能，但实际行为和预期不符，想请大家帮忙排查下问题。

问题现象 #

我原本预期：当某个主节点故障时，连接到该节点的客户端应该自动切换到对应的备份节点。但实际观察到的情况是，客户端会直接跳转到集群里的其他主节点，只有偶尔会按预期连接到备份节点，而且看起来和消费者初始创建的位置有关。

举个具体例子：

• 消费者A初始连接到代理A，消费者B初始连接到代理B
• 关闭代理A后，消费者A并没有切换到代理A的备份节点，反而直接连到了代理B
• 重启代理A后再关闭代理B，之前跳转到代理B的消费者会切回代理A，而另一个消费者则正确连接到代理B的备份节点

环境说明 #

我用的是复制式HA架构，目前在本地部署了6个节点（3主3备，通过端口区分），后续计划部署到3台物理服务器，现在先做本地测试验证。

客户端配置 #

连接工厂代码 #

我参考了官方的客户端故障转移相关文档，在连接URL里添加了ha=true和reconnectAttempts=10参数，代码如下：

public ActiveMQConnectionFactory activeMQConnectionFactory() throws NamingException{
    ArtemisTargetConfig config = (ArtemisTargetConfig) new InitialContext().lookup("java:comp/env/bean/ArtemisConfig");
    ActiveMQConnectionFactory cf = new ActiveMQConnectionFactory("udp://" + config.getGroupAddress() + ":" + config.getGroupPort() + "?ha=true&reconnectAttempts=10");
    cf.setUser(config.getUserName());
    cf.setPassword(config.getPassword());

    return cf;
}

最终生成的连接URL是：udp://231.7.7.7:9876?ha=true&reconnectAttempts=10

JmsConfig配置 #

@EnableJms
@Configuration
public class JmsConfig {

    private static final Logger LOG = LoggerFactory.getLogger(JmsConfig.class);

    private final ActiveMQConnectionFactory activeMQConnectionFactory;

    @Autowired
    public JmsConfig(
            ActiveMQConnectionFactory activeMQConnectionFactory){
        this.activeMQConnectionFactory = activeMQConnectionFactory;
    }
    
    @Bean
    public CachingConnectionFactory cachingConnectionFactory() {
        return new CachingConnectionFactory(activeMQConnectionFactory);
    }

    @Bean
    public JmsTemplate jmsTemplate(CachingConnectionFactory cachingConnectionFactory) {
        JmsTemplate template = new JmsTemplate(cachingConnectionFactory);
        template.setPubSubDomain(false);
        return template;
    }

    @Bean
    public DefaultJmsListenerContainerFactory jmsListenerContainerFactory () {

        DefaultJmsListenerContainerFactory jmsListenerContainerFactory = new DefaultJmsListenerContainerFactory();
        jmsListenerContainerFactory.setConnectionFactory(activeMQConnectionFactory);
        jmsListenerContainerFactory.setSessionAcknowledgeMode(ActiveMQJMSConstants.INDIVIDUAL_ACKNOWLEDGE);
        jmsListenerContainerFactory.setPubSubDomain(false);
        
        MappingJackson2MessageConverter converter = new MappingJackson2MessageConverter();
        ObjectMapper objectMapper = new ObjectMapper();
        objectMapper.registerModule(new JavaTimeModule());
        converter.setObjectMapper(objectMapper);
        converter.setTargetType(MessageType.TEXT);
        converter.setTypeIdPropertyName("_type");
        jmsListenerContainerFactory.setCacheLevel(null);
        jmsListenerContainerFactory.setMessageConverter(converter);

        return jmsListenerContainerFactory;
    }
}

代理配置（关键部分） #

主节点1的broker.xml #

<name>localhost</name>

<connectors>
    <connector name="artemis">tcp://localhost:61616</connector>
</connectors>

<acceptors>
    <acceptor name="artemis">tcp://localhost:61616?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;amqpMinLargeMessageSize=102400;protocols=CORE,AMQP,STOMP,HORNETQ,MQTT,OPENWIRE;useEpoll=true;amqpCredits=1000;amqpLowCredits=300;amqpDuplicateDetection=true;supportAdvisory=false;suppressInternalManagementObjects=false 
    </acceptor>
</acceptors>

<cluster-user>cluster-admin</cluster-user>
<cluster-password>apassword</cluster-password>
<broadcast-groups>
    <broadcast-group name="bg-group1">
        <group-address>231.7.7.7</group-address>
        <group-port>9876</group-port>
        <broadcast-period>5000</broadcast-period>
        <connector-ref>artemis</connector-ref>
    </broadcast-group>
</broadcast-groups>

<discovery-groups>
    <discovery-group name="dg-group1">
        <group-address>231.7.7.7</group-address>
        <group-port>9876</group-port>
        <refresh-timeout>10000</refresh-timeout>
    </discovery-group>
</discovery-groups>

<cluster-connections>
    <cluster-connection name="my-cluster">
        <connector-ref>artemis</connector-ref>
        <message-load-balancing>ON_DEMAND</message-load-balancing>
        <max-hops>1</max-hops>
        <discovery-group-ref discovery-group-name="dg-group1"/>
    </cluster-connection>
</cluster-connections>

<ha-policy>
    <replication>
        <primary>
            <vote-on-replication-failure>true</vote-on-replication-failure>
            <check-for-active-server>true</check-for-active-server>
        </primary>
    </replication>
</ha-policy>

备份节点1的broker.xml #

<name>localhost</name>

<connectors>
    <connector name="artemis">tcp://localhost:61617</connector>
</connectors>

<acceptors>
    <!-- Acceptor for every supported protocol -->
    <acceptor name="artemis">tcp://localhost:61617?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;amqpMinLargeMessageSize=102400;protocols=CORE,AMQP,STOMP,HORNETQ,MQTT,OPENWIRE;useEpoll=true;amqpCredits=1000;amqpLowCredits=300;amqpDuplicateDetection=true;supportAdvisory=false;suppressInternalManagementObjects=false 
    </acceptor>
</acceptors>

<cluster-user>cluster-admin</cluster-user>
<cluster-password>aPassword</cluster-password>

<broadcast-groups>
    <broadcast-group name="bg-group1">
        <group-address>231.7.7.7</group-address>
        <group-port>9876</group-port>
        <broadcast-period>5000</broadcast-period>
        <connector-ref>artemis</connector-ref>
    </broadcast-group>
</broadcast-groups>

<discovery-groups>
    <discovery-group name="dg-group1">
        <group-address>231.7.7.7</group-address>
        <group-port>9876</group-port>
        <refresh-timeout>10000</refresh-timeout>
    </discovery-group>
</discovery-groups>

<cluster-connections>
    <cluster-connection name="my-cluster">
        <connector-ref>artemis</connector-ref>
        <message-load-balancing>ON_DEMAND</message-load-balancing>
        <max-hops>1</max-hops>
        <discovery-group-ref discovery-group-name="dg-group1"/>
    </cluster-connection>
</cluster-connections>

<ha-policy>
    <replication>
        <backup>
          <allow-failback>true</allow-failback>
        </backup>
    </replication>
</ha-policy>
</core>

集群拓扑信息 #

通过代理的listNetworkTopology操作得到的结果：

[{"nodeID":"19f26f12-ee1f-11ef-b81c-00155df5022d","live":"localhost:61616","primary":"localhost:61616","backup":"localhost:61617"},
    {"nodeID":"b071e7f0-ee3f-11ef-b10d-00155df5022d","live":"localhost:61620","primary":"localhost:61620","backup":"localhost:61621"},
    {"nodeID":"d46e2693-ee1f-11ef-9cd3-00155df5022d","live":"localhost:61618","primary":"localhost:61618","backup":"localhost:61619"}]

已尝试的排查操作 #

• 确认复制功能正常：主节点日志能看到AMQ221025: Replication: sending NIOSequentialFile，备份节点日志有AMQ221031: backup announced的输出
• 完全移除了CachedConnectionFactory，问题依旧
• 删除并重建过所有代理，问题短暂解决但之后又重现

有没有朋友遇到过类似的情况？或者能帮我指出配置里可能存在的问题？

备注：内容来源于stack exchange，提问作者Sergio Scaramuzzi