构建高可用LVS+keepalived+httpd和双主模型的keepalived方案

构建高可用LVS+keepalived+httpd和双主模型的keepalived方案

keepalived简介

keepalived是一个类似于layer3, 4 & 5交换机制的软件，也就是我们平时说的第3层、第4层和第5层交换。Keepalived的作用是检测web服务器的状态，如果有一台web服务器死机，或工作出现故障，Keepalived将检测到，并将有故障的web服务器从系统中剔除，当web服务器工作正常后Keepalived自动将web服务器加入到服务器群中，通过ping来监测对方的状态。 拓扑图：

----------------------------------------------------------------------------------------- 实验环境： centos-6.4 所需的安装包：

主的和备的Director上都安装上keepalived

后端的两个realserver上都安装上htppd服务 ----------------------------------------------------------------------------------------- 1.先同步各点上的时间 [root@yulong]# ntpdate 172.16.0.1 2.在主的和从的上面安装keepalived [root@yulong ~]# yum install keepalived -y 3.编辑keepalived配置文件 [root@yulongkeepalived]# vim keepalived.conf 修改一下参数为自己所用的 global_defs { notification_email { root@localhost收件人 } notification_email_fromkeep@localhost发件人 smtp_server 127.0.0.1 邮件服务器 smtp_connect_timeout 10 链接超时时间 router_id LVS_DEVEL 添加如下字段设置可以手动切换故障转移: vrrp_scriptchk_maintainace { script \ interval 1 weight -2 } 这段脚本的意思就是如果在这个目录下存在down的这个一个文件就把vip给转移到BAKPUP节点.这个文件可以手动创建来实现调试. 以下用文字注释的都是需要修改的地方 vrrp_instance VI_1 { state MASTER 设置主节点 interface eth0 网卡类型 virtual_router_id 51 虚拟路由ID priority 100 设置权重,权重必须大于BACKUP的权重 advert_int 1 authentication { auth_type PASS 加密方法 auth_pass 1111 密码 } virtual_ipaddress { 172.16.8.100 设置访问的虚拟IP } track_script { chk_maintainace这个是定义刚才添加的那段代码生效 } } 4.配置完毕以上内容后复制到从节点上一份 [root@yulongkeepalived]# scpkeepalived.confyu:/etc/keepalived/ 5.然后在从节点上需要修改下如下两项 [root@www ~]# vim /etc/keepalived/keepalived.conf state BACKUP 设置为备用节点 priority 99 权重设置为比主节点的小 6.然后主备都启动keepalived服务 [root@yulongkeepalived]# servicekeepalived restart Stopping keepalived: [ OK ] Starting keepalived: [ OK ] 7.然后查看master上的状态 [root@yulongkeepalived]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:0e:d7:4d brdff:ff:ff:ff:ff:ff inet 172.16.8.17/16 brd 172.16.255.255 scope global eth0 inet 172.16.8.100/32 scope global eth0 刚才配置的虚拟ip已经生效 inet6 fe80::20c:29ff:fe0e:d74d/64 scope link valid_lft forever preferred_lft forever 8.然后我们把master上的节点给停掉看看能不能转移 root@yulongkeepalived]# touch down [root@yulongkeepalived]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:0e:d7:4d brdff:ff:ff:ff:ff:ff inet 172.16.8.17/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe0e:d74d/64 scope link valid_lft forever preferred_lft forever 在keepalived目录下Touch down 文件是我们写的那条脚本,很强大哟 在次ipaddr show查看vip已经没有了 9.在备节点上查看vip有没有转移过来 [root@www ~]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:6a:51:21 brdff:ff:ff:ff:ff:ff inet 172.16.8.18/16 brd 172.16.255.255 scope global eth0 inet 172.16.8.100/32 scope global eth0 OK,已经转移过来了,说明是正常的. inet6 fe80::20c:29ff:fe6a:5121/64 scope link valid_lft forever preferred_lft forever 10.下面在主节点上配置keepalived的realserver能调度上游的服务器 [root@yulongkeepalived]# vim keepalived.conf virtual_server 172.16.8.100 80 { 虚拟IP,就是一会要访问的IP delay_loop 6 lb_algorr调度算法 lb_kind DR lvs的模式 nat_mask 255.255.0.0 persistence_timeout 50

protocol TCP 使用的协议 real_server 172.16.8.10 80 { realserver的ip地址这个别写错了 weight 1 权重 TCP_CHECK { realserver的状态监测 connect_timeout 3 响应超时时间,默认3秒 nb_get_retry 3 重试次数 delay_before_retry 3 重试间隔时间 } } real_server 172.16.8.11 80 { 第二个realserver的定义和上面的一样 TCP_CHECK { connect_timeout 3 nb_get_retry 3 delay_before_retry 3 } } } 11.配置完毕以后复制到备服务器上一份 [root@yulongkeepalived]# scpkeepalived.confyu:/etc/keepalived/ 然后在备的上面编辑keepalived.conf修改如下两项 state BACKUP 设置为备用节点 priority 99 权重设置为比主节点的小 12.上面的配置好以后我们来配置realserver服务器,两个realserver上都配置如下命令 [root@node1 ~]# ifconfig lo:0 172.16.8.100 broadcast 172.16.8.100 netmask 255.255.255.255 up 由于我们使用的DR模型配置DR模型的vip,在lo上配置就行 [root@node1 ~]# echo \[root@node1 ~]# echo \[root@node1 ~]# echo \[root@node1 ~]# echo \关闭arp广播 [root@node1 ~]# echo \设置访问的网页主页内容 13.配置完成后两个realserver服务器上都启动web服务器进行测试 [root@node1 ~]# service httpd restart Stopping httpd: [ OK ] Starting httpd: [ OK ] 多点击几下看看是不是均衡

访问成功,看来是均衡的 14.下面我们来通过ipvsadm查看下访问的状态 [root@yulongkeepalived]# yum install ipvsadm [root@yulongkeepalived]# ipvsadm -L -n --stats IP Virtual Server version 1.2.1 (size=4096) ProtLocalAddress:Port Conns InPktsOutPktsInBytesOutBytes ->RemoteAddress:Port TCP 172.16.8.100:80 3 3 0 156 0 -> 172.16.8.10:80 2 2 0 104 0 -> 172.16.8.11:80 1 1 0 52 0 Ok 看来是通过keepalived调用的 下面来配置keepalived双主模型 1.在主的上面添加如下内容 [root@yulongkeepalived]# vimkeepalived.conf vrrp_instance VI_2 { state BACKUP interface eth0 virtual_router_id 51 priority 99 advert_int 1 authentication { auth_type PASS auth_pass 2111 } virtual_ipaddress { 172.16.8.200 } track_script { chk_maintainace } } 2.在备的上面添加如下内容 vrrp_instance VI_2 { state MASTER interface eth0 virtual_router_id 51 priority 100 advert_int 1 authentication { auth_type PASS auth_pass 2111 } virtual_ipaddress { 172.16.8.200 } track_script { chk_maintainace } } 3.重启主备的keepalived的服务 [root@yulongkeepalived]# servicekeepalived restart 4.查看备的上面虚拟主ip有木有生效 [root@www ~]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:6a:51:21 brdff:ff:ff:ff:ff:ff inet 172.16.8.18/16 brd 172.16.255.255 scope global eth0 inet 172.16.8.200/32 scope global eth0 已经生效 inet6 fe80::20c:29ff:fe6a:5121/64 scope link valid_lft forever preferred_lft forever 5.下面我们把备的master给down掉看看是否会转移 root@www ~]# cd /etc/keepalived/ [root@wwwkeepalived]# touch down [root@wwwkeepalived]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:6a:51:21 brdff:ff:ff:ff:ff:ff inet 172.16.8.18/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe6a:5121/64 scope link valid_lft forever preferred_lft forever 172.16.8.200 的虚拟ip已经看不到了 6.下面在主的上面看看是否转移过来 [root@yulongkeepalived]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:0e:d7:4d brdff:ff:ff:ff:ff:ff inet 172.16.8.17/16 brd 172.16.255.255 scope global eth0 inet 172.16.8.100/32 scope global eth0 inet 172.16.8.200/32 scope global eth0 inet6 fe80::20c:29ff:fe0e:d74d/64 scope link valid_lft forever preferred_lft forever 两个虚拟ip都在,证明是正常的,此次双主模型搞定！

inet6 fe80::20c:29ff:fe6a:5121/64 scope link valid_lft forever preferred_lft forever 172.16.8.200 的虚拟ip已经看不到了 6.下面在主的上面看看是否转移过来 [root@yulongkeepalived]# ipaddr show 2: eth0: mtu 1500 qdiscpfifo_fast state UP qlen 1000 link/ether 00:0c:29:0e:d7:4d brdff:ff:ff:ff:ff:ff inet 172.16.8.17/16 brd 172.16.255.255 scope global eth0 inet 172.16.8.100/32 scope global eth0 inet 172.16.8.200/32 scope global eth0 inet6 fe80::20c:29ff:fe0e:d74d/64 scope link valid_lft forever preferred_lft forever 两个虚拟ip都在,证明是正常的,此次双主模型搞定！

本文来源：https://www.bwwdw.com/article/fg26.html