keepalived搭建zabbix server双机高可用

头像
Buttonwood
    阅读 8 分钟
    1

    零. 准备工作

    最近公司zabbix服务器升级，鉴于以前单机负载压力较大，宕机之后系统瘫痪，趁机改成双机高可用架构。
    以下是用keepalived做HA的搭建过程，以备参考！

    两台主机，三个ip，两台主机配置zabbix和keepalived用于高可用架设。

    IP地址 主机名 备注
    192.168.3.141 zabbix_master zabbix/keepalived
    192.168.3.141 zabbix_slaver zabbix/keepalived
    192.168.3.144 VIP ---

    一. keepalived安装

    1. Installation

    centOS 7下安装很简单，直接yum即可。

    $ su root 
$ yum install keepalived
$ /usr/sbin/keepalived -D
$ pstree |grep keepalived
        |-keepalived---2*[keepalived]

    可以看到实际启动了三个进程。

    2. Base Configuration

    制作开机启动

    pgrep keepalived |xargs kill -15
 #重新加载
systemctl daemon-reload  
 #设置开机自动启动
systemctl enable keepalived.service
 #取消开机自动启动  
systemctl disable keepalived.service 
 #启动
systemctl start keepalived.service
 #停止
systemctl stop keepalived.service
 #状态检查
systemctl status keepalived.service
    3. Advance Configuration
    su root
chown zabbix:zabbix /usr/sbin/keepalived
chown zabbix:zabbix /lib/systemd/system/keepalived.service
chown zabbix:zabbix /etc/keepalived/keepalived.conf
su zabbix 
vim /lib/systemd/system/keepalived.service
vim /etc/keepalived/keepalived.conf
chmod +x /home/zabbix/keepalived/src/zabbix_status_check2.sh

    主要是keepalived.conf和zabbix_status_check2.sh两个脚本，网上资料很多，这里不细讲。

    二. 双机启动

    1. 主机启动keepalived
    Mar 13 17:27:57 vm1184 Keepalived_vrrp[5026]: VRRP_Script(zabbix_status_check) succeeded
Mar 13 17:28:00 vm1184 Keepalived_vrrp[5026]: VRRP_Instance(VI_1) Transition to MASTER STATE
Mar 13 17:28:01 vm1184 Keepalived_vrrp[5026]: VRRP_Instance(VI_1) Entering MASTER STATE
Mar 13 17:28:01 vm1184 Keepalived_vrrp[5026]: VRRP_Instance(VI_1) setting protocol VIPs.
Mar 13 17:28:01 vm1184 Keepalived_vrrp[5026]: VRRP_Instance(VI_1) Sending gratuitous ARPs on ens160 for 192.168.3.144
Mar 13 17:28:01 vm1184 Keepalived_healthcheckers[5025]: Netlink reflector reports IP 192.168.3.144 added
Mar 13 17:28:01 vm1184 avahi-daemon[637]: Registering new address record for 192.168.3.144 on ens160.IPv4.
Mar 13 17:28:06 vm1184 Keepalived_vrrp[5026]: VRRP_Instance(VI_1) Sending gratuitous ARPs on ens160 for 192.168.3.144
    2. 主机地址,VIP已经绑定到网卡
    $  ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:94:eb:ad brd ff:ff:ff:ff:ff:ff
    inet 192.168.3.141/32 brd 192.168.3.141 scope global ens160
       valid_lft forever preferred_lft forever
    inet 192.168.3.144/24 scope global ens160
       valid_lft forever preferred_lft forever
    inet6 fe80::ff15:1cc9:5bd0:b06e/64 scope link
       valid_lft forever preferred_lft forever
    3. 备机日志，keepalived已启动。
    $  tail -f  /var/log/messages
Mar 14 09:24:02 vm1185 Keepalived_healthcheckers[7588]: Registering Kernel netlink command channel
Mar 14 09:24:02 vm1185 Keepalived_healthcheckers[7588]: Opening file '/etc/keepalived/keepalived.conf'.
Mar 14 09:24:02 vm1185 Keepalived_healthcheckers[7588]: Configuration is using : 8087 Bytes
Mar 14 09:24:02 vm1185 Keepalived_healthcheckers[7588]: Using LinkWatch kernel netlink reflector...
    4. 备机ip地址，VIP实际并没有绑到网卡上。
    $  ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:94:88:c9 brd ff:ff:ff:ff:ff:ff
    inet 192.168.3.142/32 brd 192.168.3.142 scope global ens160
       valid_lft forever preferred_lft forever
    inet6 fe80::e00c:f722:cd30:ad8b/64 scope link
       valid_lft forever preferred_lft forever

    三. 主备切换

    1. 杀掉主机上的keepalived之后，备机现在自动切换成主机了。
    Mar 14 09:29:04 vm1185 Keepalived_vrrp[7589]: VRRP_Instance(VI_1) Transition to MASTER STATE
Mar 14 09:29:05 vm1185 Keepalived_vrrp[7589]: VRRP_Instance(VI_1) Entering MASTER STATE
Mar 14 09:29:05 vm1185 Keepalived_vrrp[7589]: VRRP_Instance(VI_1) setting protocol VIPs.
Mar 14 09:29:05 vm1185 Keepalived_vrrp[7589]: VRRP_Instance(VI_1) Sending gratuitous ARPs on ens160 for 192.168.3.144
Mar 14 09:29:05 vm1185 Keepalived_healthcheckers[7588]: Netlink reflector reports IP 192.168.3.144 added
Mar 14 09:29:05 vm1185 avahi-daemon[674]: Registering new address record for 192.168.3.144 on ens160.IPv4.
Mar 14 09:29:10 vm1185 Keepalived_vrrp[7589]: VRRP_Instance(VI_1) Sending gratuitous ARPs on ens160 for 192.168.3.144
    2. 现主机（原备机）现在的网卡情况，VIP已经绑定上去了。
    $  ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:94:88:c9 brd ff:ff:ff:ff:ff:ff
    inet 192.168.3.142/32 brd 192.168.3.142 scope global ens160
       valid_lft forever preferred_lft forever
    inet 192.168.3.144/24 scope global ens160
       valid_lft forever preferred_lft forever
    inet6 fe80::e00c:f722:cd30:ad8b/64 scope link
       valid_lft forever preferred_lft forever
    3. 再启动原主机上的keepalived，原主机变备机，VIP绑定被清除了。
    $  tail -f /var/log/messages
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) Entering MASTER STATE
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) setting protocol VIPs.
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) Sending gratuitous ARPs on ens160 for 10.18      .8.144
Mar 14 09:30:12 vm1184 avahi-daemon[637]: Registering new address record for 192.168.3.144 on ens160.IPv4.
Mar 14 09:30:12 vm1184 Keepalived_healthcheckers[11856]: Netlink reflector reports IP 192.168.3.144 added
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) Received higher prio advert
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) Entering BACKUP STATE
Mar 14 09:30:12 vm1184 Keepalived_vrrp[11857]: VRRP_Instance(VI_1) removing protocol VIPs.
Mar 14 09:30:12 vm1184 Keepalived_healthcheckers[11856]: Netlink reflector reports IP 192.168.3.144 removed
Mar 14 09:30:12 vm1184 avahi-daemon[637]: Withdrawing address record for 192.168.3.144 on ens160.
    4. 原主机变备机，VIP绑定确实被清除了。
     $  ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:50:56:94:eb:ad brd ff:ff:ff:ff:ff:ff
    inet 192.168.3.141/32 brd 192.168.3.141 scope global ens160
       valid_lft forever preferred_lft forever
    inet6 fe80::ff15:1cc9:5bd0:b06e/64 scope link
       valid_lft forever preferred_lft forever

    四. zabbix_server 宕机实验

      # 原server启动方式
$  ps -ef |grep zabbix_server

zabbix   31668 31643  0 Mar13 ?        00:00:11 ./sbin/zabbix_server: escalator #1 [processed 0 escalations in 0.001148 sec, idle 3 sec]
zabbix   31669 31643  0 Mar13 ?        00:00:03 ./sbin/zabbix_server: proxy poller #1 [exchanged data with 0 proxies in 0.000066 sec, idle 5 sec]
zabbix   31670 31643  0 Mar13 ?        00:00:09 ./sbin/zabbix_server: self-monitoring [processed data in 0.000046 sec, idle 1 sec]
zabbix   31671 31643  0 Mar13 ?        00:00:03 ./sbin/zabbix_server: task manager [processed 0 task(s) in 0.000289 sec, idle 5 sec]

  # 杀掉进程，操作宕机
$  pgrep zabbix_server |xargs kill -15

  # 检查
$  ps -ef |grep zabbix_server

zabbix   14053 14028  0 09:59 ?        00:00:00 /home/zabbix/zabbix/sbin/zabbix_server: history syncer #4 [synced 0 items in 0.000043 sec, idle 1 sec]
zabbix   14054 14028  0 09:59 ?        00:00:00 /home/zabbix/zabbix/sbin/zabbix_server: escalator #1 [processed 0 escalations in 0.001202 sec, idle 3 sec]
zabbix   14055 14028  0 09:59 ?        00:00:00 /home/zabbix/zabbix/sbin/zabbix_server: proxy poller #1 [exchanged data with 0 proxies in 0.000057 sec, idle 5 sec]
zabbix   14056 14028  0 09:59 ?        00:00:00 /home/zabbix/zabbix/sbin/zabbix_server: self-monitoring [processed data in 0.000040 sec, idle 1 sec]
zabbix   14057 14028  0 09:59 ?        00:00:00 /home/zabbix/zabbix/sbin/zabbix_server: task manager [processed 0 task(s) in 0.000433 sec, idle 5 sec]

    说明keepalived的健康检查脚本确实执行了，检查不到zabbix_server进程的情况下会自动重启。

    zabbixkeepalivedcentos7
    Buttonwood
    17 声望1 粉丝

    Float like a butterfly! Stand like a buttonwood!

    « 上一篇
    自动检测硬盘，格式化及挂载
    下一篇 »
    zabbix3.2和oracle数据库安装记录

    引用和评论

    推荐阅读
    头像
    分区partition知识点

    Buttonwood阅读 2.4k

    头像
    【Docker】基本概念及语法与环境搭建

    neojayway阅读 4.9k

    头像
    Keepalived 实现 Ambari-Server 高可用

    LnEoi阅读 261

    头像
    Zabbix 数据对接观测云最佳实践

    观测云阅读 203

    头像
    Zabbix 和 Prometheus 选型对比

    SRETALK阅读 182

    头像
    WGCLOUD使用笔记功能介绍 - agent可以采集主机的哪些指标信息

    幽影绝尘阅读 161

    头像
    Zabbix 和 Prometheus 选型对比

    千杯不醉的瀑布阅读 145

    0 条评论
    得票最新