我在 Kubernetes 中建立了一个 Artemis 集群，有 3 组主/从：
activemq-artemis-master-0 1/1 Running
activemq-artemis-master-1 1/1 Running
activemq-artemis-master-2 1/1 Running
activemq-artemis-slave-0 0/1 Running
activemq-artemis-slave-1 0/1 Running
activemq-artemis-slave-2 0/1 Running
Artemis 版本是 2.17.0。这是我在 master-0 broker.xml 中的集群配置。除了 connector-ref 更改为匹配代理外，其他代理的配置相同：
<?xml version="1.0"?>
<!--
Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements. See the NOTICE file
distributed with this work for additional information
regarding copyright ownership. The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing,
software distributed under the License is distributed on an
"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
KIND, either express or implied. See the License for the
specific language governing permissions and limitations
under the License.
-->
<configuration xmlns="urn:activemq" xmlns:xi="http://www.w3.org/2001/XInclude" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="urn:activemq /schema/artemis-configuration.xsd">
<core xmlns="urn:activemq:core" xsi:schemaLocation="urn:activemq:core ">
<name>activemq-artemis-master-0</name>
<persistence-enabled>true</persistence-enabled>
<!-- this could be ASYNCIO, MAPPED, NIO
ASYNCIO: Linux Libaio
MAPPED: mmap files
NIO: Plain Java Files
-->
<journal-type>ASYNCIO</journal-type>
<paging-directory>data/paging</paging-directory>
<bindings-directory>data/bindings</bindings-directory>
<journal-directory>data/journal</journal-directory>
<large-messages-directory>data/large-messages</large-messages-directory>
<journal-datasync>true</journal-datasync>
<journal-min-files>2</journal-min-files>
<journal-pool-files>10</journal-pool-files>
<journal-device-block-size>4096</journal-device-block-size>
<journal-file-size>10M</journal-file-size>
<!--
This value was determined through a calculation.
Your system could perform 1.1 writes per millisecond
on the current journal configuration.
That translates as a sync write every 911999 nanoseconds.
Note: If you specify 0 the system will perform writes directly to the disk.
We recommend this to be 0 if you are using journalType=MAPPED and journal-datasync=false.
-->
<journal-buffer-timeout>100000</journal-buffer-timeout>
<!--
When using ASYNCIO, this will determine the writing queue depth for libaio.
-->
<journal-max-io>4096</journal-max-io>
<!--
You can verify the network health of a particular NIC by specifying the <network-check-NIC> element.
<network-check-NIC>theNicName</network-check-NIC>
-->
<!--
Use this to use an HTTP server to validate the network
<network-check-URL-list>http://www.apache.org</network-check-URL-list> -->
<!-- <network-check-period>10000</network-check-period> -->
<!-- <network-check-timeout>1000</network-check-timeout> -->
<!-- this is a comma separated list, no spaces, just DNS or IPs
it should accept IPV6
Warning: Make sure you understand your network topology as this is meant to validate if your network is valid.
Using IPs that could eventually disappear or be partially visible may defeat the purpose.
You can use a list of multiple IPs, and if any successful ping will make the server OK to continue running -->
<!-- <network-check-list>10.0.0.1</network-check-list> -->
<!-- use this to customize the ping used for ipv4 addresses -->
<!-- <network-check-ping-command>ping -c 1 -t %d %s</network-check-ping-command> -->
<!-- use this to customize the ping used for ipv6 addresses -->
<!-- <network-check-ping6-command>ping6 -c 1 %2$s</network-check-ping6-command> -->
<!-- how often we are looking for how many bytes are being used on the disk in ms -->
<disk-scan-period>5000</disk-scan-period>
<!-- once the disk hits this limit the system will block, or close the connection in certain protocols
that won't support flow control. -->
<max-disk-usage>90</max-disk-usage>
<!-- should the broker detect dead locks and other issues -->
<critical-analyzer>true</critical-analyzer>
<critical-analyzer-timeout>120000</critical-analyzer-timeout>
<critical-analyzer-check-period>60000</critical-analyzer-check-period>
<critical-analyzer-policy>HALT</critical-analyzer-policy>
<page-sync-timeout>2244000</page-sync-timeout>
<!-- the system will enter into page mode once you hit this limit.
This is an estimate in bytes of how much the messages are using in memory
The system will use half of the available memory (-Xmx) by default for the global-max-size.
You may specify a different value here if you need to customize it to your needs.
<global-max-size>100Mb</global-max-size>
-->
<acceptors>
<!-- useEpoll means: it will use Netty epoll if you are on a system (Linux) that supports it -->
<!-- amqpCredits: The number of credits sent to AMQP producers -->
<!-- amqpLowCredits: The server will send the # credits specified at amqpCredits at this low mark -->
<!-- amqpDuplicateDetection: If you are not using duplicate detection, set this to false
as duplicate detection requires applicationProperties to be parsed on the server. -->
<!-- amqpMinLargeMessageSize: Determines how many bytes are considered large, so we start using files to hold their data.
default: 102400, -1 would mean to disable large mesasge control -->
<!-- Note: If an acceptor needs to be compatible with HornetQ and/or Artemis 1.x clients add
"anycastPrefix=jms.queue.;multicastPrefix=jms.topic." to the acceptor url.
See https://issues.apache.org/jira/browse/ARTEMIS-1644 for more information. -->
<!-- Acceptor for every supported protocol -->
<acceptor name="artemis">tcp://0.0.0.0:61616?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;amqpMinLargeMessageSize=102400;protocols=CORE,AMQP,STOMP,HORNETQ,MQTT,OPENWIRE;useEpoll=true;amqpCredits=1000;amqpLowCredits=300;amqpDuplicateDetection=true</acceptor>
<!-- AMQP Acceptor. Listens on default AMQP port for AMQP traffic.-->
<acceptor name="amqp">tcp://0.0.0.0:5672?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;protocols=AMQP;useEpoll=true;amqpCredits=1000;amqpLowCredits=300;amqpMinLargeMessageSize=102400;amqpDuplicateDetection=true</acceptor>
<!-- STOMP Acceptor. -->
<acceptor name="stomp">tcp://0.0.0.0:61613?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;protocols=STOMP;useEpoll=true</acceptor>
<!-- HornetQ Compatibility Acceptor. Enables HornetQ Core and STOMP for legacy HornetQ clients. -->
<acceptor name="hornetq">tcp://0.0.0.0:5445?anycastPrefix=jms.queue.;multicastPrefix=jms.topic.;protocols=HORNETQ,STOMP;useEpoll=true</acceptor>
<!-- MQTT Acceptor -->
<acceptor name="mqtt">tcp://0.0.0.0:1883?tcpSendBufferSize=1048576;tcpReceiveBufferSize=1048576;protocols=MQTT;useEpoll=true</acceptor>
</acceptors>
<security-settings>
<security-setting match="#">
<permission type="createNonDurableQueue" roles="amq"/>
<permission type="deleteNonDurableQueue" roles="amq"/>
<permission type="createDurableQueue" roles="amq"/>
<permission type="deleteDurableQueue" roles="amq"/>
<permission type="createAddress" roles="amq"/>
<permission type="deleteAddress" roles="amq"/>
<permission type="consume" roles="amq"/>
<permission type="browse" roles="amq"/>
<permission type="send" roles="amq"/>
<!-- we need this otherwise ./artemis data imp wouldn't work -->
<permission type="manage" roles="amq"/>
</security-setting>
</security-settings>
<address-settings>
<!-- if you define auto-create on certain queues, management has to be auto-create -->
<address-setting match="activemq.management#">
<dead-letter-address>DLQ</dead-letter-address>
<expiry-address>ExpiryQueue</expiry-address>
<redelivery-delay>0</redelivery-delay>
<!-- with -1 only the global-max-size is in use for limiting -->
<max-size-bytes>-1</max-size-bytes>
<message-counter-history-day-limit>10</message-counter-history-day-limit>
<address-full-policy>PAGE</address-full-policy>
<auto-create-queues>true</auto-create-queues>
<auto-create-addresses>true</auto-create-addresses>
<auto-create-jms-queues>true</auto-create-jms-queues>
<auto-create-jms-topics>true</auto-create-jms-topics>
</address-setting>
<!--default for catch all-->
<address-setting match="#">
<dead-letter-address>DLQ</dead-letter-address>
<expiry-address>ExpiryQueue</expiry-address>
<redelivery-delay>0</redelivery-delay>
<!-- with -1 only the global-max-size is in use for limiting -->
<max-size-bytes>-1</max-size-bytes>
<message-counter-history-day-limit>10</message-counter-history-day-limit>
<address-full-policy>PAGE</address-full-policy>
<auto-create-queues>true</auto-create-queues>
<auto-create-addresses>true</auto-create-addresses>
<auto-create-jms-queues>true</auto-create-jms-queues>
<auto-create-jms-topics>true</auto-create-jms-topics>
</address-setting>
</address-settings>
<addresses>
<address name="DLQ">
<anycast>
<queue name="DLQ"/>
</anycast>
</address>
<address name="ExpiryQueue">
<anycast>
<queue name="ExpiryQueue"/>
</anycast>
</address>
</addresses>
<!-- Uncomment the following if you want to use the Standard LoggingActiveMQServerPlugin pluging to log in events
<broker-plugins>
<broker-plugin class-name="org.apache.activemq.artemis.core.server.plugin.impl.LoggingActiveMQServerPlugin">
<property key="LOG_ALL_EVENTS" value="true"/>
<property key="LOG_CONNECTION_EVENTS" value="true"/>
<property key="LOG_SESSION_EVENTS" value="true"/>
<property key="LOG_CONSUMER_EVENTS" value="true"/>
<property key="LOG_DELIVERING_EVENTS" value="true"/>
<property key="LOG_SENDING_EVENTS" value="true"/>
<property key="LOG_INTERNAL_EVENTS" value="true"/>
</broker-plugin>
</broker-plugins>
-->
<cluster-user>clusterUser</cluster-user>
<cluster-password>aShortclusterPassword</cluster-password>
<connectors>
<connector name="activemq-artemis-master-0">tcp://activemq-artemis-master-0.activemq-artemis-master.svc.cluster.local:61616</connector>
<connector name="activemq-artemis-slave-0">tcp://activemq-artemis-slave-0.activemq-artemis-slave.svc.cluster.local:61616</connector>
<connector name="activemq-artemis-master-1">tcp://activemq-artemis-master-1.activemq-artemis-master.svc.cluster.local:61616</connector>
<connector name="activemq-artemis-slave-1">tcp://activemq-artemis-slave-1.activemq-artemis-slave.svc.cluster.local:61616</connector>
<connector name="activemq-artemis-master-2">tcp://activemq-artemis-master-2.activemq-artemis-master.svc.cluster.local:61616</connector>
<connector name="activemq-artemis-slave-2">tcp://activemq-artemis-slave-2.activemq-artemis-slave.svc.cluster.local:61616</connector>
</connectors>
<cluster-connections>
<cluster-connection name="activemq-artemis">
<connector-ref>activemq-artemis-master-0</connector-ref>
<retry-interval>500</retry-interval>
<retry-interval-multiplier>1.1</retry-interval-multiplier>
<max-retry-interval>5000</max-retry-interval>
<initial-connect-attempts>-1</initial-connect-attempts>
<reconnect-attempts>-1</reconnect-attempts>
<message-load-balancing>ON_DEMAND</message-load-balancing>
<max-hops>1</max-hops>
<!-- scale-down>true</scale-down -->
<static-connectors>
<connector-ref>activemq-artemis-master-0</connector-ref>
<connector-ref>activemq-artemis-slave-0</connector-ref>
<connector-ref>activemq-artemis-master-1</connector-ref>
<connector-ref>activemq-artemis-slave-1</connector-ref>
<connector-ref>activemq-artemis-master-2</connector-ref>
<connector-ref>activemq-artemis-slave-2</connector-ref>
</static-connectors>
</cluster-connection>
</cluster-connections>
<ha-policy>
<replication>
<master>
<group-name>activemq-artemis-0</group-name>
<quorum-vote-wait>12</quorum-vote-wait>
<vote-on-replication-failure>true</vote-on-replication-failure>
<!--we need this for auto failback-->
<check-for-live-server>true</check-for-live-server>
</master>
</replication>
</ha-policy>
</core>
<core xmlns="urn:activemq:core">
<jmx-management-enabled>true</jmx-management-enabled>
</core>
</configuration>
我的使用者在 Spring Boot 应用中被定义为 JmsListener。在队列中的消息消费过程中，Spring Boot 应用程序崩溃，导致 kubernetes 删除 pod 并重新创建一个新的 pod。但是，我注意到新的 pod 没有连接到同一个 Artemis 节点，因此之前连接的剩余消息永远不会被消耗。
我认为使用集群的全部意义在于让所有 Artemis 节点作为一个单元向消费者传递消息，而不管它连接到哪个节点。我错了吗？如果集群无法将消费者连接重新路由到正确的节点（该节点保存来自前一个消费者的剩余消息），那么推荐的处理这种情况的方法是什么？
答案 0 :(得分：2)
首先，需要注意的是，没有任何功能可以让客户端重新连接到客户端崩溃/重启后断开的代理。一般来说，客户端不应该真正关心它连接到哪个代理；这是水平可扩展性的主要目标之一。
还值得注意的是，如果代理上的消息数量和连接的客户端数量足够少，以至于这种情况经常出现，那几乎肯定意味着您的集群中有太多的代理。
也就是说，我相信您的客户端没有收到它期望的消息的原因是因为您使用了默认的 redistribution-delay（即 -1），这意味着消息不会 重新分配到集群中的其他节点。如果您想启用重新分发（这似乎是您所做的），那么您应该将其设置为 >= 0，例如：
<address-setting match="#">
...
<redistribution-delay>0</redistribution-delay>
...
</address-setting>
您可以在 the documentation 中阅读有关再分配的更多信息。
除此之外，您可能需要重新考虑您的拓扑。通常，如果您处于类似云的环境中（例如使用 Kubernetes 的环境），其中基础设施本身将重启失败的 pod，那么您不会使用主/从配置。您只需将日志挂载在单独的 pod 上（例如使用 NFSv4），这样当节点出现故障时，它会重新启动，然后重新连接回其持久存储。这有效地提供了代理高可用性（这是为云环境之外设计的主/从）。
此外，单个 ActiveMQ Artemis 实例每秒可以处理数百万条消息，具体取决于用例，因此您实际上可能不需要 3 个活动节点来满足预期负载。
请注意，这些是关于您的整体架构的一般建议，与您的问题没有直接关系。