linuxea:rabbitmq3.7.9集群配置笔记 此前，简单的叙述了rabbitmq(瞎扯)的单机使用，本篇记录了rabbitmq集群的简单使用。其他参考：activemq+zookeeper集群配置笔记准备环境：CentOS Linux release 7.5.1804 (Core) erlang-21.1.4-1.el7.centos.x86_64rabbitmq-server-3.7.9-1.el7HA-Proxy version 1.6.9 2(docker)大致情况如下：先决条件编辑hosts文件[root@LinuxEA-172_25_12_16 ~]# cat /etc/hosts 172.25.12.16 rabbitmq-1 172.25.12.17 rabbitmq-2 172.25.12.18 rabbitmq-3而后重启网卡systemctl restart network如果有必要，修改主机名。否则每次登陆可能无法正常使用hostnamectl set-hostname rabbitmq-1 hostnamectl set-hostname rabbitmq-2 hostnamectl set-hostname rabbitmq-3安装在页面https://packagecloud.io/rabbitmq/erlang/packages/el/7/erlang-21.1.4-1.el7.centos.x86_64.rpm根据提示，使用如下安装：curl -s https://packagecloud.io/install/repositories/rabbitmq/erlang/script.rpm.sh | sudo bash sudo yum install erlang-21.1.4-1.el7.centos.x86_64在页面https://packagecloud.io/rabbitmq/rabbitmq-server/packages/el/7/rabbitmq-server-3.7.9-1.el7.noarch.rpm根据提示，使用如下安装curl -s https://packagecloud.io/install/repositories/rabbitmq/rabbitmq-server/script.rpm.sh | sudo bash sudo yum install rabbitmq-server-3.7.9-1.el7.noarch其他版本仓库：http://www.rabbitmq.com/releases/rabbitmq-server/配置互通配置互通主要用于传递文件，和集群没毛关系[root@LinuxEA-172_25_12_16 ~]# ssh-keygen -t rsa [root@LinuxEA-172_25_12_16 ~]# ssh-copy-id 172.25.12.17 [root@LinuxEA-172_25_12_16 ~]# ssh-copy-id 172.25.12.18防火墙-A INPUT -s 172.25.12.16/32 -p tcp -m tcp -m state --state NEW -m multiport --dports 4369,5671:5673,15672:15673,25672 -m comment --comment "RabbitMQ_cluster-1" -j ACCEPT -A INPUT -s 172.25.12.17/32 -p tcp -m tcp -m state --state NEW -m multiport --dports 4369,5671:5673,15672:15673,25672 -m comment --comment "RabbitMQ_cluster-2" -j ACCEPT -A INPUT -s 172.25.12.18/32 -p tcp -m tcp -m state --state NEW -m multiport --dports 4369,5671:5673,15672:15673,25672 -m comment --comment "RabbitMQ_cluster-3" -j ACCEPT [root@LinuxEA--172_25_12_16 ~]# iptables -L -nv|grep RabbitMQ 2158 110K ACCEPT tcp -- * * 0.0.0.0/0 0.0.0.0/0 tcp state NEW multiport dports 4369,5671:5673,15672:15673,25672 /* RabbitMQ_cluster-1 */ 0 0 ACCEPT tcp -- * * 172.25.12.16 0.0.0.0/0 tcp state NEW multiport dports 4369,5671:5673,15672:15673,25672 /* RabbitMQ_cluster-1 */ 0 0 ACCEPT tcp -- * * 172.25.12.17 0.0.0.0/0 tcp state NEW multiport dports 4369,5671:5673,15672:15673,25672 /* RabbitMQ_cluster-2 */ 0 0 ACCEPT tcp -- * * 172.25.12.18 0.0.0.0/0 tcp state NEW multiport dports 4369,5671:5673,15672:15673,25672 /* RabbitMQ_cluster-3 */我假设你关闭了selinux,如果没有，请关闭。统一认证配置将本地的/var/lib/rabbitmq/.erlang.cookie 复制到另外两台节点相同目录下[root@LinuxEA-172_25_12_16 ~]# scp -P22992 /var/lib/rabbitmq/.erlang.cookie root@172.25.12.17:/var/lib/rabbitmq/ .erlang.cookie 100% 20 16.3KB/s 00:00 [root@LinuxEA-172_25_12_16 ~]# scp -P22992 /var/lib/rabbitmq/.erlang.cookie root@172.25.12.18:/var/lib/rabbitmq/ .erlang.cookie 100% 20 16.5KB/s 00:00 如果有必要，可以修改下权限[root@LinuxEA-172_25_12_16 ~]# ssh -p22992 root@172.25.12.17 'chown rabbitmq:rabbitmq /var/lib/rabbitmq/.erlang.cookie' [root@LinuxEA-172_25_12_16 ~]# ssh -p22992 root@172.25.12.18 'chown rabbitmq:rabbitmq /var/lib/rabbitmq/.erlang.cookie' [root@LinuxEA-172_25_12_16 ~]# ssh -p22992 root@172.25.12.17 'chmod 400 /var/lib/rabbitmq/.erlang.cookie ' [root@LinuxEA-172_25_12_16 ~]# ssh -p22992 root@172.25.12.18 'chmod 400 /var/lib/rabbitmq/.erlang.cookie '而后关闭掉所有节点，重新启动另外两台rabbitmq节点[root@LinuxEA-172_25_12_16 ~]# rabbitmq-server -detached Warning: PID file not written; -detached was passed.[root@LinuxEA-172_25_12_17 ~]# rabbitmq-server -detached Warning: PID file not written; -detached was passed.[root@LinuxEA-172_25_12_18 ~]# rabbitmq-server -detached Warning: PID file not written; -detached was passed.加入集群将172.25.12.17加入到172.25.12.16集群中[root@LinuxEA-172_25_12_17 ~]# rabbitmqctl stop_app Stopping rabbit application on node rabbit@rabbitmq-2 ... [root@LinuxEA-172_25_12_17 ~]# rabbitmqctl join_cluster --ram rabbit@rabbitmq-1 Clustering node rabbit@rabbitmq-2 with rabbit@rabbitmq-1 [root@LinuxEA-172_25_12_17 ~]# rabbitmqctl start_app Starting node rabbit@rabbitmq-2 ... completed with 4 plugins.[root@LinuxEA-172_25_12_17 ~]# rabbitmqctl cluster_status Cluster status of node rabbit@rabbitmq-2 ... [{nodes,[{disc,['rabbit@rabbitmq-1']},{ram,['rabbit@rabbitmq-2']}]}, {running_nodes,['rabbit@rabbitmq-1','rabbit@rabbitmq-2']}, {cluster_name,<<"rabbit@DS-VM-Node172_25_12_16">>}, {partitions,[]}, {alarms,[{'rabbit@rabbitmq-1',[]},{'rabbit@rabbitmq-2',[]}]}]将172.25.12.18加入到172.25.12.16集群中[root@LinuxEA-172_25_12_18 ~]# rabbitmqctl stop_app Stopping rabbit application on node rabbit@rabbitmq-3 ... [root@LinuxEA-172_25_12_18 ~]# rabbitmqctl join_cluster --ram rabbit@rabbitmq-1 Clustering node rabbit@rabbitmq-3 with rabbit@rabbitmq-1 [root@LinuxEA-172_25_12_18 ~]# rabbitmqctl start_app Starting node rabbit@rabbitmq-3 ... completed with 4 plugins.[root@LinuxEA-172_25_12_18 ~]# rabbitmqctl cluster_status Cluster status of node rabbit@rabbitmq-3 ... [{nodes,[{disc,['rabbit@rabbitmq-1']}, {ram,['rabbit@rabbitmq-3','rabbit@rabbitmq-2']}]}, {running_nodes,['rabbit@rabbitmq-2','rabbit@rabbitmq-1','rabbit@rabbitmq-3']}, {cluster_name,<<"rabbit@DS-VM-Node172_25_12_16">>}, {partitions,[]}, {alarms,[{'rabbit@rabbitmq-2',[]}, {'rabbit@rabbitmq-1',[]}, {'rabbit@rabbitmq-3',[]}]}]集群状态：[root@LinuxEA-172_25_12_16 ~]# rabbitmqctl cluster_status Cluster status of node rabbit@rabbitmq-1 ... [{nodes,[{disc,['rabbit@rabbitmq-1']}, {ram,['rabbit@rabbitmq-3','rabbit@rabbitmq-2']}]}, {running_nodes,['rabbit@rabbitmq-3','rabbit@rabbitmq-2','rabbit@rabbitmq-1']}, {cluster_name,<<"rabbit@DS-VM-Node172_25_12_16">>}, {partitions,[]}, {alarms,[{'rabbit@rabbitmq-3',[]}, {'rabbit@rabbitmq-2',[]}, {'rabbit@rabbitmq-1',[]}]}]镜像高可用集群大概分为2种，单一的不算。就剩下默认的集群模式和镜像模式。镜像模式可以完善普通集群的弊端，所有消息实体会主动在镜像节点同步，而不是单单的临时拉取。但这样的方式对性能有一定的损耗，但是胜在可靠性高。[root@LinuxEA-172_25_12_16 ~]# rabbitmqctl set_policy ha-all "^" '{"ha-mode":"all"}' Setting policy "ha-all" for pattern "^" to "{"ha-mode":"all"}" with priority "0" for vhost "/" ...配置用户添加用户[root@LinuxEA-172_25_12_16 ~]# rabbitmqctl add_user admin admin set管理员状态权限[root@LinuxEA-172_25_12_16 ~]# rabbitmqctl set_user_tags admin administrator授权读写所有权限[root@LinuxEA-172_25_12_16 ~]# rabbitmqctl set_permissions -p "/" admin ".*" ".*" ".*" Setting permissions for user "admin" in vhost "/" ...WEB管理界面插件安装[root@LinuxEA-172_25_12_16 ~]# rabbitmq-plugins enable rabbitmq_management创建一个host[root@LinuxEA-172_25_12_17 ~]# rabbitmqctl add_vhost test Adding vhost "test" ...将test所有队列标记为镜像队列，即队列会被复制到各个节点，各个节点状态保持一致[root@LinuxEA-172_25_12_17 ~]# rabbitmqctl set_policy ha-all "^" '{"ha-mode":"all"}' -p test Setting policy "ha-all" for pattern "^" to "{"ha-mode":"all"}" with priority "0" for vhost "test" ...haproxyglobal log 127.0.0.1 local3 log 127.0.0.1 local1 notice maxconn 5000 pidfile /run/haproxy.pid defaults log global mode tcp option tcplog option dontlognull retries 3 option redispatch #在使用基于cookie定向时，一旦后端某一server宕机时，会将会话重新定向至某一上游服务器，必须使用 的选项 maxconn 4000 #每个server最大的连接数 timeout http-request 10s #在客户端建立连接但不请求数据时，关闭客户端连接 timeout queue 1m #等待最大时长 timeout connect 10s #定义haproxy将客户端请求转发至后端服务器所等待的超时时长 timeout client 1m #客户端非活动状态的超时时长 timeout server 1m #客户端与服务器端建立连接后，等待服务器端的超时时长 timeout http-keep-alive 10s #定义保持连接的超时时长 timeout check 10s #健康状态监测时的超时时间，过短会误判，过长资源消耗 listen stats bind :9000 mode http stats enable stats uri /haproxy_stats stats realm HAProxy\ Statistics #统计页面密码框上提示文本，默认为Haproxy\ Statistics stats auth admin:ZTVmNTlhYzczOGM0 #统计页面用户名和密码设置 stats admin if TRUE #如果认证通过就做管理功能，可以管理后端的服务器 stats hide-version #隐藏统计页面上HAProxy的版本信息 stats refresh 30s #统计页面自动刷新时间 frontend rabbitmq bind *:5672 mode tcp log global default_backend rabbitmq backend rabbitmq balance roundrobin server rabbitmq1-slave 172.25.12.18:5672 check port 5672 rise 1 fall 2 maxconn 4000 weight 1 # mem data server rabbitmq2-slave 172.25.12.17:5672 check port 5672 rise 1 fall 2 maxconn 4000 weight 1 # mem data #server rabbitmq3-master 172.25.12.16:5672 check port 5672 rise 1 fall 2 maxconn 4000 weight 1 #disk data172.25.12.16在rabbitmqctl cluster_status查看时候是disc节点，就不做为读写了，之作为存储盘，而其他两个内存节点就作为读写[root@DT_Node-172_25_12_16 ~]# rabbitmqctl cluster_status Cluster status of node rabbit@rabbitmq-1 ... [{nodes,[{disc,['rabbit@rabbitmq-1']}, {ram,['rabbit@rabbitmq-3','rabbit@rabbitmq-2']}]}, {running_nodes,['rabbit@rabbitmq-3','rabbit@rabbitmq-2','rabbit@rabbitmq-1']}, {cluster_name,<<"rabbit@DS-VM-Node172_25_12_16">>}, {partitions,[]}, {alarms,[{'rabbit@rabbitmq-3',[]}, {'rabbit@rabbitmq-2',[]}, {'rabbit@rabbitmq-1',[]}]}]

linuxea:activemq+zookeeper集群配置笔记 activemq官网提供三种集群方式，Shared File System Master Slave，JDBC Master Slave，Replicated LevelDB Store，本章采用最后一种http://activemq.apache.org/masterslave.html ,官网是不建议使用LevelDB，推荐使用kahaDB集群方式:依靠Apache ZooKeeper协调节点是否为主，当主节点启动后接受链接时，其他节点接入等候状态并同步持久状态，并不接受客户端链接，所有持久操作都是复制到等候状态的节点，如果主宕机，最近状态的节点将被提升为主。节点重新恢复正常，自动进入从模式。当然，我们在使用中也可以在上游加上代理层，如下：下载安装包：wget http://mirror.rise.ph/apache/activemq/5.14.5/apache-activemq-5.14.5-bin.tar.gz wget http://mirror.rise.ph/apache/zookeeper/zookeeper-3.4.10/zookeeper-3.4.10.tar.gz [root@DS-VM-Node49-linuxea /usr/local]# tar xf apache-activemq-5.14.5-bin.tar.gz [root@DS-VM-Node49-linuxea /usr/local]# tar xf zookeeper-3.4.10.tar.gz [root@DS-VM-Node49-linuxea /usr/local]# mkdir /data/{activemq,zookeeper,logs} -p [root@DS-VM-Node49-linuxea /usr/local]# ln -s zookeeper-3.4.10 zookeeper [root@DS-VM-Node49-linuxea /usr/local]# ln -s apache-activemq-5.14.5 activemqdownload jdk:http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html axel -n 30 https://mirrors.dtops.cc/java/8/8u111-b14/jdk-8u111-linux-x64.rpm yum install jdk-8u111-linux-x64.rpm -y 防火墙规则：iptables -I INPUT -p tcp -m tcp -m state --state NEW -m multiport --dports 2181,2888,3888,1883,61613:61617,62621,5672,8161 -m comment --comment "mq_zook" -j ACCEPTzookeeper——node1:[root@DS-VM-Node49-linuxea ~]# vim /usr/local/zookeeper/conf/zoo.cfg tickTime=2000 initLimit=10 syncLimit=5 dataDir=/data/zookeeper dataLogDir=/data/logs clientPort=2181 maxClientCnxns=0 autopurge.snapRetainCount=3 autopurge.purgeInterval=5 server.1=10.0.1.49:2888:3888 server.2=10.0.1.61:2889:3889 server.3=10.10.240.113:2890:3890 [root@DS-VM-Node49-linuxea /data/zookeeper]# echo 1 > myid [root@DS-VM-Node49-linuxea /data/zookeeper]# ls myid [root@DS-VM-Node49-linuxea /data/zookeeper]# cat myid 1 [root@DS-VM-Node49-linuxea /data/zookeeper]# zookeeper——node2:[root@DS-VM-Node113-linuxea /usr/local]# cat /usr/local/zookeeper/conf/zoo.cfg tickTime=2000 initLimit=10 syncLimit=5 dataDir=/data/zookeeper dataLogDir=/data/logs clientPort=2181 maxClientCnxns=0 autopurge.snapRetainCount=3 autopurge.purgeInterval=5 server.1=10.0.1.49:2888:3888 server.2=10.0.1.61:2889:3889 server.3=10.10.240.113:2890:3890 [root@DS-VM-Node113-linuxea /usr/local]# echo 2 > /data/zookeeper/myid [root@DS-VM-Node113-linuxea /usr/local]# cat /data/zookeeper/myid 3 [root@DS-VM-Node113-linuxea /usr/local]# zookeeper——node3:[root@DS-VM-Node61-linuxea /usr/local/zookeeper]# cat /usr/local/zookeeper/conf/zoo.cfg tickTime=2000 initLimit=10 syncLimit=5 dataDir=/data/zookeeper dataLogDir=/data/logs clientPort=2181 maxClientCnxns=0 autopurge.snapRetainCount=3 autopurge.purgeInterval=5 server.1=10.0.1.49:2888:3888 server.2=10.0.1.61:2889:3889 server.3=10.10.240.113:2890:3890 [root@DS-VM-Node61-linuxea /usr/local/zookeeper]# echo 3 > /data/zookeeper/myid [root@DS-VM-Node61-linuxea /usr/local/zookeeper]# cat /data/zookeeper/myid 2 [root@DS-VM-Node61-linuxea /usr/local/zookeeper]# 全部启动后zkServer status，查看状态，分别是follow和leader摘在网络：Zookeeper的通信架构在Zookeeper整个系统中，有3中角色的服务，client、Follower、leader。其中client负责发起应用的请求，Follower接受client发起的请求，参与事务的确认过程，在leader crash后的leader选择。而leader主要承担事务的协调，当然leader也可以承担接收客户请求的功能，为了方便描述，后面的描述都是client与Follower之间的通信，如果Zookeeper的配置支持leader接收client的请求，client与leader的通信跟client与Follower的通信模式完全一样。Follower与leader之间的角色可能在某一时刻进行转换。一个Follower在leader crash掉以后可能被集群（Quorum）的Follower选举为leader。而一个leader在crash后，再次加入集群（Quorum）将作为Follower角色存在。在一个集群（Quorum）中,除了在选举leader的过程中没有Follower和leader的区分外，其他任何时刻都只有1个leader和多个Follower。更多的可以参考：http://zoutm.iteye.com/blog/708447[root@DS-VM-Node61-linuxea /usr/local/zookeeper]# ./bin/zkServer.sh status ZooKeeper JMX enabled by default Using config: /usr/local/zookeeper/bin/../conf/zoo.cfg Mode: leader [root@DS-VM-Node61-linuxea /usr/local/zookeeper]# MQ:备份下配置文件，因为要修改一些[root@DS-VM-Node61-linuxea /usr/local/activemq/conf]# cp activemq.xml activemq.xml.bak修改第一个地方：brokerName，如下： <broker xmlns="http://activemq.apache.org/schema/core" brokerName="linuxea" dataDirectory="${activemq.data}">修改第二个地方，这里虽然被启用了leveldb，但是还是可以用的： <persistenceAdapter> <replicatedLevelDB directory="${activemq.data}/leveldb" replicas="3" bind="tcp://0.0.0.0:51621" zkAddress="10.0.1.49:2181,10.0.1.61:2181,10.10.240.113:2181" hostname="10.0.1.49" zkPath="/activemq/leveldb-stores" /> </persistenceAdapter>node1-10.0.1.49: <broker xmlns="http://activemq.apache.org/schema/core" brokerName="linuxea" dataDirectory="${activemq.data}"> <persistenceAdapter> <replicatedLevelDB directory="${activemq.data}/leveldb" replicas="3" bind="tcp://0.0.0.0:51621" zkAddress="10.0.1.49:2181,10.0.1.61:2181,10.10.240.113:2181" hostname="10.0.1.49" zkPath="/activemq/leveldb-stores" /> </persistenceAdapter>node2-10.0.1.61: <broker xmlns="http://activemq.apache.org/schema/core" brokerName="linuxea" dataDirectory="${activemq.data}"> <persistenceAdapter> <replicatedLevelDB directory="${activemq.data}/leveldb" replicas="3" bind="tcp://0.0.0.0:51621" zkAddress="10.0.1.49:2181,10.0.1.61:2181,10.10.240.113:2181" hostname="10.0.1.61" zkPath="/activemq/leveldb-stores" /> </persistenceAdapter>node3-10.10.240.113: <broker xmlns="http://activemq.apache.org/schema/core" brokerName="linuxea" dataDirectory="${activemq.data}"> <persistenceAdapter> <replicatedLevelDB directory="${activemq.data}/leveldb" replicas="3" bind="tcp://0.0.0.0:51621" zkAddress="10.0.1.49:2181,10.0.1.61:2181,10.10.240.113:2181" hostname="10.10.240.113" zkPath="/activemq/leveldb-stores" /> </persistenceAdapter>配置完成基本上启动即可