码迷，mamicode.com
首页 > 系统相关 > 详细

Linux 高可用keepalive

时间：2018-06-05 17:03:33      阅读：175      评论：0      收藏：0      [点我收藏+]

标签：linux keepalive

技术分享图片
vrrp:虚拟路由冗余协议(Virtual Router Redundancy Protocol),冗余路由单点故障，实现把一个地址绑定到n台服务器上（只生效一台），通过组播传递信号，当一个点故障时候，选举另外点来当路由。
keepalive工作流程：监视进程watchDog监测 vrrp跟checkers，监测过程中触发系统调用跟邮件服务，来调整浮动ip，内置ipvs规则装饰跟启动ipvs脚本。

安装

网卡组播必须开启

ens34: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500

ip link set multicast on  dev ens33
ip link set multicast off dev ens33

安装

[root@node1 ~]# yum -y install keepalived

vrrp使用

主节点配置

[root@node1 keepalived]# vim /etc/keepalived/keepalived.conf
global_defs {
   notification_email {
        root@localhost
   }
   notification_email_from Alexandre.Cassen@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node1                   #单个id
   vrrp_mcast_group4 224.1.101.33    #组播地址
}

vrrp_instance VI_1 {
    state MASTER                  #当前节点在此虚拟路由器上的初始状态;只能有一个是MASTER，余下的都应该为BACKUP
    interface ens34               #通告进行所需 接口
    virtual_router_id 51          #虚拟路由器标识:VRID(0-255)，唯一标识虚拟路由器
    priority 100                  #优先级
    advert_int 1                  #vrrp通告的时间间隔，默认1s
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.1.199 
    }
    #状态改变时候出发钩子函数
    notify_master "/etc/keepalived/notify.sh master"  #当前节点成为主节点时触发的脚本
    notify_backup "/etc/keepalived/notify.sh backup"  #当前节点转为备节点时触发的脚本
    notify_fault "/etc/keepalived/notify.sh fault"   #当前节点转为“失败”状态时触发的脚本
}

从节点（区别主节点有3个地方）

global_defs {
   notification_email {
        root@localhost
   }
   notification_email_from Alexandre.Cassen@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node2                         #1.标识
   vrrp_mcast_group4 224.1.101.33
}

vrrp_instance VI_1 {
    state BACKUP                           #2.备用
    interface ens34
    virtual_router_id 51
    priority 96                            #3.优先级
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.1.199 
    }
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

邮件通知脚本(如果是双主，脚本后面加参数区分)

[root@node1 keepalived]# vim notify.sh
#!/root/.virtualenvs/shellenv/bin/python

from raven import Client
import subprocess

import sys

if len(sys.argv) < 2:
    exit()

notify = sys.argv[1]

interface = ‘ens34‘
res = subprocess.getstatusoutput("ip addr show %s|egrep -o ‘inet [0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}‘|egrep -o ‘[^a-z ].*‘"%(interface))
ip = res[1]
ip = ip.replace(‘\n‘, ‘ ‘)

res = subprocess.getstatusoutput("hostname")
hostname = res[1]

res = subprocess.getstatusoutput("uptime")
cpu = res[1]

res = subprocess.getstatusoutput("free -h")
memory = res[1]

res = subprocess.getstatusoutput("df -h")
disk = res[1]

notify_message = {‘master‘:‘Change to master‘, ‘backup‘:‘Change to backup‘, ‘fault‘:‘Server error‘}
message = notify_message.get(notify,‘{%s} no this notify‘%(notify))

info = ‘‘‘
-------- info of %s  -----
Hostname:%s

Interface:%s

IP:%s

CPU:%s

Memory:
%s

Disk:
%s

Message: %s
‘‘‘ % (hostname, hostname, interface, ip, cpu, memory,disk,  message)

client = Client(‘https://key@sentry.io/1219450‘)
client.captureMessage(info)

启动

[root@node1 keepalived]# systemctl  start keepalived

主节点通告

[root@node1 keepalived]# tcpdump -i ens34 -nn host 224.1.101.33
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ens34, link-type EN10MB (Ethernet), capture size 262144 bytes
13:04:17.240395 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

备用节点收通告

[root@node2 keepalived]# tcpdump -i ens34 -nn host 224.1.101.33
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on ens34, link-type EN10MB (Ethernet), capture size 262144 bytes
13:05:36.437749 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

主节点查看

[root@node1 keepalived]# ip a
3: ens34: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:46:a6:40 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.200/24 brd 192.168.1.255 scope global ens34
       valid_lft forever preferred_lft forever
    inet 192.168.1.198/32 scope global ens34:0
       valid_lft forever preferred_lft forever
    inet6 fe80::20c:29ff:fe46:a640/64 scope link
       valid_lft forever preferred_lft forever

从节点

此时从节点没有 ens34:0

主节点故障

#ip到从节点上
[root@node2 keepalived]# ifconfig
ens34:0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.1.199  netmask 255.255.255.255  broadcast 0.0.0.0
        ether 00:0c:29:34:f5:84  txqueuelen 1000  (Ethernet)

主节点恢复，因为优先级比从节点高，所以ip又到主节点上

邮箱读取
主服务开启：
技术分享图片
技术分享图片

主节点停掉服务
技术分享图片

日志配置

[root@node1 keepalived]# vim /etc/sysconfig/keepalived
KEEPALIVED_OPTIONS="-D -S 3"

[root@node1 keepalived]# vim /etc/rsyslog.conf
local3.*                                                /var/log/keepalive.log

[root@node1 keepalived]# systemctl restart rsyslog
[root@node1 keepalived]# systemctl restart keepalived

lvs双主模型

双主配置单节点配置,另外节点只需改动主从、routerid、优先级即可

global_defs {
   notification_email {
        root@localhost
   }
   notification_email_from Alexandre.Cassen@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node1
   vrrp_mcast_group4 224.1.101.33
}

vrrp_instance VI_1 {
    state MASTER
    interface ens34
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.1.198 dev ens34 label ens34:0
    }

    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

vrrp_instance VI_2 {
    state BACKUP
    interface ens34
    virtual_router_id 52
    priority 98
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 2222
    }
    virtual_ipaddress {
        192.168.1.197
    }

    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

virtual_server 192.168.1.198 80 {
    delay_loop 1                     
    lb_algo wrr                   
    lb_kind DR                   
    protocol TCP               
    sorry_server 192.168.1.200 80        

    real_server 192.168.1.202 80{
        weight 1
        HTTP_GET {
                url {
                        path /index.html
                        status_code 200
                }
                nb_get_retry 3
                delay_before_retry 2
                connect_timeout 3
        }
    }
    real_server 192.168.1.203 80{
        weight 1
        TCP_CHECK {
                nb_get_retry 3
                delay_before_retry 2
                connect_timeout 3
        }
    }
}

virtual_server 192.168.1.197 80 {
    delay_loop 1
    lb_algo wrr
    lb_kind DR
    protocol TCP
    sorry_server 192.168.1.200 80

    real_server 192.168.1.202 80{
        weight 1
        HTTP_GET {
                url {
                        path /index.html
                        status_code 200
                }
                nb_get_retry 3
                delay_before_retry 2
                connect_timeout 3
        }
    }
    real_server 192.168.1.203 80{
        weight 1
        TCP_CHECK {
                nb_get_retry 3
                delay_before_retry 2
                connect_timeout 3
        }
    }
}
[root@node1 keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.1.197:80 wrr
  -> 192.168.1.202:80             Route   1      0          0
  -> 192.168.1.203:80             Route   1      0          0
TCP  192.168.1.198:80 wrr
  -> 192.168.1.202:80             Route   1      0          0
  -> 192.168.1.203:80             Route   1      0          0
[root@node2 keepalived]# ipvsadm -ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.1.197:80 wrr
  -> 192.168.1.202:80             Route   1      0          0
  -> 192.168.1.203:80             Route   1      0          0
TCP  192.168.1.198:80 wrr
  -> 192.168.1.202:80             Route   1      0          0
  -> 192.168.1.203:80             Route   1      0          0

由此可见：ipvs规则两台机器配置是一样的，只是vip是飘动的，keepalive针对lvs还实现了健康状态监测

marvindeMacBook-Pro:~ marvin$ curl http://192.168.1.197/
node4
marvindeMacBook-Pro:~ marvin$ curl http://192.168.1.197/
node3

#停掉node3节点
marvindeMacBook-Pro:~ marvin$ curl http://192.168.1.197/
node4
marvindeMacBook-Pro:~ marvin$ curl http://192.168.1.197/
node4

自定义监测模型-nginx

配置手动下线文件脚本

[root@node1 keepalived]# vim down.sh
 [[ -f /etc/keepalived/down ]] && exit 1 || exit 0

配置文件

[root@node1 keepalived]# vim keepalived.conf

 global_defs {
   notification_email {
        root@localhost
   }
   notification_email_from Alexandre.Cassen@localhost
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id node1
   vrrp_mcast_group4 224.1.101.33
}

vrrp_script chk_down {
        script "/etc/keepalived/down.sh"
        interval 1
        weight -20
        fall 1
        rise 1
}

vrrp_script chk_nginx {
    script "pidof nginx && exit 0 || exit 1"
    interval 1
    weight -3
    fall 2
    rise 2      #有可能重启后又down 所以2次比较合理
}

vrrp_instance VI_1 {
    state MASTER
    interface ens34
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    track_script {
        chk_down
        chk_nginx
    }
    virtual_ipaddress {
        192.168.1.198/24
    }

    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

/etc/keepalived/notify.sh主要在里面加入

res = subprocess.getstatusoutput("systemctl start mynginx")
res = subprocess.getstatusoutput("pidof nginx")
nginxpid = res[1]

测试

[root@node2 keepalived]# tcpdump -i ens34 -nn host 224.1.101.33

[root@node1 keepalived]# touch down
15:29:27.073406 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
15:29:28.074393 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 80, authtype simple, intvl 1s, length 20
15:29:28.074710 IP 192.168.1.201 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 99, authtype simple, intvl 1s, length 20

[root@node1 keepalived]#  rm -f down
15:30:17.179289 IP 192.168.1.201 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 99, authtype simple, intvl 1s, length 20
15:30:17.180088 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

[root@node1 keepalived]# systemctl stop mynginx
15:31:01.251469 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
15:31:02.253183 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
15:31:03.254806 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 97, authtype simple, intvl 1s, length 20
15:31:03.255137 IP 192.168.1.201 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 99, authtype simple, intvl 1s, length 20
15:30:17.180088 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20
15:30:18.181628 IP 192.168.1.200 > 224.1.101.33: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s,

随后自动启动 是因为    notify_backup "/etc/keepalived/notify.sh backup"
这个脚本写了nginx 启动

邮件查看
技术分享图片
技术分享图片

Linux 高可用keepalive

标签：linux keepalive

原文地址：http://blog.51cto.com/marvin89/2125140
踩
(0)
赞
(0)
   
举报
评论 一句话评论（0）
登录后才能评论！
分享档案
更多>
2021年07月29日 (22)
2021年07月28日 (40)
2021年07月27日 (32)
2021年07月26日 (79)
2021年07月23日 (29)
2021年07月22日 (30)
2021年07月21日 (42)
2021年07月20日 (16)
2021年07月19日 (90)
2021年07月16日 (35)
周排行
mamicode.com排行更多图片
更多
友情链接
兰亭集智   国之画   百度统计   站长统计   阿里云   chrome插件   新版天听网
关于我们 - 联系我们 - 留言反馈
© 2014 mamicode.com 版权所有  联系我们:gaon5@hotmail.com
迷上了代码！