keepalived configuration Introduction

keepalived Configuration Introduction

A, keeplived Profile

The official reference documentation

use

Vrrp protocol software implementation, native ipvs service designed for high availability

Features

• Vrrp protocol flow based on completion address generating rule ipvs vip node address resides (pre-defined in the configuration file)
• Do health testing for the RS ipvs cluster
• Script-based call interface functions defined in the script completed by executing the script, thereby affecting the clusters affairs, in order to support nginx, haproxy services

protocol

​ vrrp协议：Virtual Router Redundancy Protocol

Way of working

Preemptive (default), non-preemptive

Operating mode

• Master / slave: single virtual router
• Active / active: master / slave (Virtual Router 1), prepared / main (Virtual Router 2)

the term

• Virtual Router: Virtual Router

• Virtual router ID: VRID (0-255), that uniquely identifies the virtual router

• Physical router:

  • master: master
  • backup: spare device
  • priority: Priority

• VIP：Virtual IP

• VMAc: virutal MAC (00-00-5e-00-01-VRID)

Two, keepalived installation

apt-get install keepalived

By default apt or yum install, keepalived -v can view the build parameters, which has been good enough, generally do not compile

Environment program

• The main configuration file: /etc/keepalived/keepalived.conf

  It is not apt to install the default configuration file, you can copy the example configuration file to change what you can

  Mar 22 19:48:50 director Keepalived[2124]: Starting Keepalived v1.3.9 (10/21,2017)
  Mar 22 19:48:50 director Keepalived[2124]: Unable to find configuration file /etc/keepalived/keepalived.conf (glob returned 3)
  Mar 22 19:48:50 director Keepalived[2124]: Stopped Keepalived v1.3.9 (10/21,2017)
  

  dpkg  -L keepalived |grep conf
  /usr/share/doc/keepalived/samples/keepalived.conf.sample
  

• The main program file: / usr / sbin / keepalived

• servcie : /lib/systemd/system/keepalived.service

• System parameter file: / etc / default / keepalived (service specified in the document)

Three, keepalive configuration

Part of keepalived.conf

1. global_defs global configuration

• {notification_email
  the root @ localhost object failover occurs the message transmitted by a plurality of rows can be written to distinguish, can be set to username @ hostname
  }
• Send mail notification messages notification_email_from [email protected]
• smtp_server 192.168.200.1 sending mail server address
• Timeout duration smtp_connect_timeout 30 connected to the mail server
• LVS_DEVEL physical node identifier router_id, generally provided with the same host name server
• checking all packets vrrp_skip_check_adv_addr consumption performance comparison, this configuration of the received packet and if the packet is a router with a source address of the skip check packets
• vrrp_strict strict compliance with VRRP protocol does not allow the situation: 1, no VIP address, unicast neighbor 2, 3 has an IPv6 address in VRRP version 2.
• vrrp_garp_interval 0 #ARP message transmission delay
• message transmission delay vrrp_gna_interval 0 #
• vrrp_mcast_group4 224.0.0.18 # default multicast IP address, 224.0.0.0 to 239.255.255.255

2. vrrp_instance VI_1 virtual routing instance

• state MASTER | BACKUP: the initial state of the current node on this virtual router status as MASTER or BACKUP

• interface IFACE_NAME: bindings for virtual router physical interface is currently used ens32, eth0, bond0, br0

• virtual_router_id VRID: this virtual router uniquely identified range is 0-255 (standby consistent, a plurality of nodes must be the same)

• priority 100: current physical node priority in this virtual router; range 1-254

• advert_int 1: vrrp announcement interval, default 1s

• authentication {# authentication mechanism

  auth_type AH | PASS encrypted form

  auth_pass <PASSWORD> authentication password valid only the first 8
  }

• virtual_ipaddress {

  #虚拟IP <IPADDR>/<MASK> brd <IPADDR> dev<STRING> scope <SCOPE> label <LABEL>

  192.168.200.17/24 dev eth1

  192.168.200.18/24 dev eth2 label eth2: 1 (when a plurality of cards have addresses, subnet interfaces needed)

  }

• track_interface {# Configure monitor the network interface, once failure occurs, the address transition into FAULT condition is realized eth0 eth1 ...}

3. virtual_server lvs related settings

The default configuration examples

virtual_server 10.10.10.2 1358  {   ip 端口
    delay_loop 6   检查后端服务器的时间间隔 6s
    lb_algo rr  定义调度方法  rr 
    lb_kind NAT 集群的类型
    persistence_timeout 50  持久连接时长(长链接设置，用于短时间内保存session ,一般不用这个，用redis)  
    protocol TCP  指定服务协议 

    sorry_server 192.168.200.200 1358

    real_server 192.168.200.2 1358 {
        weight 1   RS权重 
        HTTP_GET {  应用层检测 
            url {
              path /testurl3/test.jsp  定义要监控的URL 
              digest 640205b7b0fc66c1ea91c463fac6334d  对请求的页面进行hash运算，然后和这个hash码进行比对，如果hash码一样就表示状态正常
            }
            connect_timeout 3  连接超时时间
            retry 3  重试次数 
            delay_before_retry 3  重试之前的延迟时长     
        }
    }
}

lvs settings is divided into three parts

• vs-related settings

  • The definition of virtual server (virtual services) (virtual_server can set multiple)
  • virtual_server IP port # define virtual host IP address and its port
  • virtual_server fwmark int #ipvs firewall marking, to achieve load balancing cluster-based firewall
  • virtual_server group string # define multiple virtual servers into groups, the groups defined as virtual services
  • delay_loop <INT>: Check backend server interval
  • lb_algo rr | wrr | lc | wlc | lblc | sh | dh: defined scheduling method
  • lb_kind NAT | TUN | DR: type cluster
  • persistence_timeout <INT>: long persistent connection
  • protocol TCP | UDP | SCTP: Specifies the service agreement
  • sorry_server <IPADDR> <PORT>: All RS fails, standby server address

• rs-related settings

  ​ real_server <IPADDR><PORT> {

  ​ weight<INT> RS权重

  notify_up <STRING> | <QUOTED-STRING> on-line notification script RS

  notify_down <STRING> | <QUOTED-STRING> RS offline notification script

  HTTP_GET | SSL_GET | TCP_CHECK | SMTP_CHECK | MISC_CHEC K {...} defining the current state of health of the host detection method

  }

• Back-end server health checks

  To check the service application layer determines whether there is a case of suspended animation, java application using the associated code quality related

  HTTP_GET|SSL_GET：应用层检测 
  HTTP_GET|SSL_GET { 
  
  	url { path <URL_PATH>：定义要监控的URL 
  		status_code <INT>：判断上述检测机制为健康状态的响应码 
  		} 
  	connect_timeout <INTEGER>：连接请求的超时时长
  	nb_get_retry <INT>：重试次数 
  	delay_before_retry <INT>：重试之前的延迟时长
  	connect_ip <IP ADDRESS>：向当前RS哪个IP地址发起健康状态检测请求 
  	connect_port <PORT>：向当前RS的哪个PORT发起健康状态检测请求 
  	bindto <IP ADDRESS>：发出健康状态检测请求时使用的源地址 
  	bind_port <PORT>：发出健康状态检测请求时使用的源端口 
  }
  

  Check-off layer 4 mainly through the port status code determination is alive php, nginx like using tcp

  TCP_CHECK { 
  	connect_ip <IP ADDRESS>：向当前RS的哪个IP地址发起健康状态检测请求
  	connect_port <PORT>：向当前RS的哪个PORT发起健康状态检测请求
  	bindto <IP ADDRESS>：发出健康状态检测请求时使用的源地址
  	bind_port <PORT>：发出健康状态检测请求时使用的源端口
  	connect_timeout <INTEGER>：连接请求的超时时长 
  }