|
|
一、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，这些工作全部自动完成，不需要人工干涉，需要人工做的只是修复故障的web服务器。
工作原理
Layer3,4&5工作在IP/TCP协议栈的IP层，TCP层，及应用层,原理分别如下：
Layer3：Keepalived使用Layer3的方式工作式时，Keepalived会定期向服务器群中的服务器发送一个ICMP的数据包（既我们平时用的Ping程序）,如果发现某台服务的IP地址没有激活，Keepalived便报告这台服务器失效，并将它从服务器群中剔除，这种情况的典型例子是某台服务器被非法关机。Layer3的方式是以服务器的IP地址是否有效作为服务器工作正常与否的标准。
Layer4:如果您理解了Layer3的方式，Layer4就容易了。Layer4主要以TCP端口的状态来决定服务器工作正常与否。如web server的服务端口一般是80，如果Keepalived检测到80端口没有启动，则Keepalived将把这台服务器从服务器群中剔除。
Layer5：Layer5就是工作在具体的应用层了，比Layer3,Layer4要复杂一点，在网络上占用的带宽也要大一些。Keepalived将根据用户的设定检查服务器程序的运行是否正常，如果与用户的设定不相符，则Keepalived将把服务器从服务器群中剔除。
二、实验步骤：
1.创建管理节点在node1上，建立双机互信node1和node2，然后同步时间，安装keepalived
[iyunv@node1~]# ansible all -m yum -a 'name=keepalived state=present'
[iyunv@node1keepalived]# rpm -qc keepalived
/etc/keepalived/keepalived.conf//生成的主配置文件
/etc/sysconfig/keepalived
2.在node1上配置文件需要做一下修改
global_defs{
notification_email {
root@localhost //收邮件人，可以定义多个
}
notification_email_from kaadmin@localhost //发邮件人可以伪装
smtp_server 127.0.0.1 //发送邮件的服务器地址
smtp_connect_timeout 30 //连接超时时间
router_id LVS_DEVEL
}
vrrp_instanceVI_1 { //每一个vrrp_instance就是定义一个虚拟路由器的
state MASTER //由初始状态状态转换为master状态
interface eth0
virtual_router_id 51 //虚拟路由的id号，一般不能大于255的
priority 100 //初始化优先级
advert_int 1 //初始化通告
authentication { //认证机制
auth_type PASS
auth_pass 1111 //密码
}
virtual_ipaddress { //虚拟地址vip
172.16.2.8
}
}
3.把配置文件复制到node2上一份，并修改初始状态和优先级
[iyunv@node1keepalived]# scp keepalived.conf node2:/etc/keepalived/
[iyunv@node2~]# cd /etc/keepalived/
[iyunv@node2keepalived]# ls
keepalived.conf
[iyunv@node2keepalived]# vim keepalived.conf
vrrp_instanceVI_1 {
state BACKUP //初始化状态
interface eth0
virtual_router_id 51
priority 99 //优先级，一定要比master的优先级要低
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
172.16.2.8
}
}
在node1上开始启动服务[iyunv@node1 ~]# servicekeepalived start
然后检查ip地址
[iyunv@node1~]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:4e:22:fb brdff:ff:ff:ff:ff:ff
inet 172.16.2.1/16 brd 172.16.255.255 scopeglobal eth0
inet 172.16.2.8/32 scopeglobal eth0
inet 172.16.10.8/16 brd 172.16.255.255 scopeglobal secondary eth0:0
inet6 fe80::20c:29ff:fe4e:22fb/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether 2e:79:b3:b2:3e:31 brdff:ff:ff:ff:ff:ff
4.现在把node1的keepalived停掉
[iyunv@node1keepalived]# service keepalived stop
Stoppingkeepalived: [ OK ]
验证node2是否把virtual_ipaddress拿走
[iyunv@node2~]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:74:c7:7b brdff:ff:ff:ff:ff:ff
inet 172.16.2.16/16 brd172.16.255.255 scope global eth0
inet 172.16.2.8/32 scopeglobal eth0
inet6 fe80::20c:29ff:fe74:c77b/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether0a:b1:ef:7b:93:18 brd ff:ff:ff:ff:ff:ff
验证成功
可以在配置文件中手动通过vrrp_script定义一个外围的检测机制，并在vrrp_instance中通过定义track_script来追踪脚本执行过程，实现节点转移
实验测试在/etc/keepalived/keepalived.conf中做一下修改
global_defs{
notification_email {
root@localhost
}
notification_email_from kaadmin@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_script chk_maintainace { //检测机制的脚本名称为chk_maintainace
script "[[ -e/etc/keepalived/down ]] && exit 1 || exit 0" //可以是个脚本路径，也可以是脚本命令
interval 1 //每隔1秒中检测一次
weight -2 //优先级减2
}
vrrp_instanceVI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
172.16.2.8
}
track_script { //调用外围脚本，追踪外围脚本执行过程
chk_maintainace
}
}
[iyunv@node1 keepalived]# touch down //在node1上创建down文件
[iyunv@node1 keepalived]# ls
down keepalived.conf keepalived.conf.bak
在node2上做同样的操作，但不创建down文件，之后一起重启服务
[iyunv@node1 keepalived]# ansible all -m shell -a 'service keepalivedrestart'
node2.magedu.com| success | rc=0 >>
Stoppingkeepalived: [FAILED]
Startingkeepalived: [ OK ]
node1.magedu.com| success | rc=0 >>
Stoppingkeepalived: [ OK ]
Startingkeepalived: [ OK ]
进行检测
[iyunv@node2keepalived]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:74:c7:7b brdff:ff:ff:ff:ff:ff
inet 172.16.2.16/16 brd172.16.255.255 scope global eth0
inet 172.16.2.8/32 scopeglobal eth0
inet6 fe80::20c:29ff:fe74:c77b/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether0a:b1:ef:7b:93:18 brd ff:ff:ff:ff:ff:ff
此时将node1中/etc/keepalived/下的down删除，进行查看
[iyunv@node1keepalived]# ls
down keepalived.conf keepalived.conf.bak
[iyunv@node1keepalived]# rm down
rm:remove regular empty file `down'? y
[iyunv@node1keepalived]# ls
keepalived.conf keepalived.conf.bak
[iyunv@node1 keepalived]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:4e:22:fb brdff:ff:ff:ff:ff:ff
inet 172.16.2.1/16 brd 172.16.255.255 scopeglobal eth0
inet 172.16.2.8/32 scopeglobal eth0
inet 172.16.10.8/16 brd 172.16.255.255scope global secondary eth0:0
inet6 fe80::20c:29ff:fe4e:22fb/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether0a:bd:4f:a9:ed:67 brd ff:ff:ff:ff:ff:ff
验证成功
三、详细介绍以下四个功能实现操作
1.如何在状态转换时进行通知？
2.如何配置Ipvs?
3.如何对某特定服务做高可用
4.如何实现基于多虚拟路由的master/master模型？
1.要在状态转换是进行通知，需要定义通知脚本可以在
vrrp_sync_group{
}中定义，也可以在
vrrp_instance{
｝中定义
通过man keepalived命令可以查看通知脚本定义的两种方法
第一种
# to MASTER transition
notify_master /path/to_master.sh
# to BACKUP transition
notify_backup /path/to_backup.sh
# FAULT transition
notify_fault "/path/fault.sh VG_1"
第二种
#arguments
# $1 ="GROUP"|"INSTANCE"
# $2 = name of group or instance
# $3 = target state of transition
# ("MASTER"|"BACKUP"|"FAULT")
notify /path/notify.sh
例如：
转换为MASTER的状态通知
#!/bin/bash
#
vip=172.16.2.8
contact='root@localhost'
thisip=`ifconfigeth0 | awk '/inet addr:/{print $2}' | awk -F: '{print $2}'`
notify(){
mailbody="vrrp transaction, $vipfloated to $thisip."
subject="$thisip is to be $vipmaster"
echo $mailbody | mail -s $subject $contact
}
notify
其他状态转换类似
下面用一个脚本notify.sh实现状态转换通知的简单示例：
#!/bin/bash
#Author: MageEdu <linuxedu@foxmail.com>
#description: An example of notify script
#
vip=172.16.2.8
contact='root@localhost'
notify(){
mailsubject="`hostname` to be $1: $vipfloating"
mailbody="`date '+%F %H:%M:%S'`: vrrptransition, `hostname` changed to be $1"
echo $mailbody | mail -s"$mailsubject" $contact
}
case"$1" in
master)
notify master
exit 0
;;
backup)
notify backup
exit 0
;;
fault)
notify fault
exit 0
;;
*)
echo 'Usage: `basename $0`{master|backup|fault}'
exit 1
;;
esac
进行测试
[iyunv@node1keepalived]# ./notify.sh backup
[iyunv@node1keepalived]# mail
HeirloomMail version 12.4 7/29/08. Type ? forhelp.
"/var/spool/mail/root":6 messages 1 new 6 unread
U 1centos@stu2.magedu.c Sat Aug 1709:34 17/644 "*** SECURITY"
U 2Cron Daemon Tue Aug 2700:01 22/747 "Cron <root@s"
U 3Cron Daemon Fri Aug 3000:01 22/747 "Cron <root@s"
U 4Mail Delivery System Fri Aug 3017:42 91/2751 "Undelivered "
U 5Cron Daemon Tue Sep 3 00:01 22/747 "Cron<root@s"
>N 6 root Thu Sep 26 21:19 18/700 "node1.magedu"
&6
Message 6:
Fromroot@node1.magedu.com Thu Sep 2621:19:32 2013
Return-Path:<root@node1.magedu.com>
X-Original-To:root@localhost
Delivered-To:root@localhost.magedu.com
Date:Thu, 26 Sep 2013 21:19:32 +0800
To:root@localhost.magedu.com
Subject:node1.magedu.com to be backup: 172.16.2.8 floating
User-Agent:Heirloom mailx 12.4 7/29/08
Content-Type:text/plain; charset=us-ascii
From:root@node1.magedu.com (root)
Status:R
2013-09-26 21:19:32: vrrp transition, node1.magedu.com changed to bebackup
&quit
Held6 messages in /var/spool/mail/root
Youhave mail in /var/spool/mail/root
通过传参数master|backup|fault验证都可以成功
在配置文件keepalived.conf中进行脚本调用
vrrp_instanceVI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
172.16.2.8
}
track_script {
chk_maintainace
}
notify_master "/etc/keepalived/notify.shmaster"
notify_backup"/etc/keepalived/notify.sh backup"
notify_fault"/etc/keepalived/notify.sh fault"
}
为node2提供同样的配置然后进行测试
[iyunv@node1keepalived]# ls
down keepalived.conf keepalived.conf.bak notify.sh
[iyunv@node1keepalived]# rm -f down
[iyunv@node1keepalived]# mail
>N18 root Thu Sep 2621:57 18/700 "node1.magedu.comto be master: 172.16.2.8 floating"截取了一条
验证都可以成功
2、如何配置ipvs
virtual_server172.16.2.8 80{
delay_loop 6
lb_algo rr
lb_kind NAT
nat_mask 255.255.0.0
persistence_timeout 0
protocol TCP
#
real_server 172.16.2.1 80 {
weight 1
HTTP_GET {
url {
path /
state_code 200
}
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 172.16.2.16 80 {
weight 1
HTTP_GET {
url {
path /
state_code 200
}
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
在node2上做同样的修改，启动httpd服务,keepalived能自动生成规则，然后查看ipvsadm规则
[iyunv@node1keepalived]# ipvsadm -L -n
IPVirtual Server version 1.2.1 (size=4096)
ProtLocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 172.16.2.8:80 rr
-> 172.16.2.1:80 Local 1 0 0
-> 172.16.2.16:80 Masq 1 0 0
3、如何对某特定服务做高可用？以nginx为例进行讲解
在两个节点上安装nginx
[iyunv@node1~]# ansible all -m yum -a 'name=nginx state=present'
启动nginx服务，启动之前注意要停止httpd服务
[iyunv@node1~]# ansible all -m shell -a 'service nginx start'
node2.magedu.com| success | rc=0 >>
Startingnginx: [ OK ]
node1.magedu.com| success | rc=0 >>
Startingnginx: [ OK ]
对node1和node2中/etc/keepalived/下的notify.sh脚本进行修改
#!/bin/bash
#Author: MageEdu <linuxedu@foxmail.com>
#description: An example of notify script
#
vip=172.16.2.8
contact='root@localhost'
notify(){
mailsubject="`hostname` to be $1: $vipfloating"
mailbody="`date '+%F %H:%M:%S'`: vrrptransition, `hostname` changed to be $1"
echo $mailbody | mail -s"$mailsubject" $contact
}
case"$1" in
master)
notify master
/etc/rc.d/init.d/nginx start
exit 0
;;
backup)
notify backup
/etc/rc.d/init.d/nginx stop
exit 0
;;
fault)
notify fault
/etc/rc.d/init.d/nginx stop
exit 0
;;
*)
echo 'Usage: `basename $0`{master|backup|fault}'
exit 1
;;
esac
然后启动keepalived服务，可以看到在node1上80端口开始启用
[iyunv@node1keepalived]# ss -tanl | grep :80
LISTEN 0 128 *:80 *:*
然后在/etc/keepalive/下创建down文件，看nginx服务是否可以转移到node2上
[iyunv@node1keepalived]# ls
keepalived.conf keepalived.conf.bak notify.sh
[iyunv@node1keepalived]# touch down
[iyunv@node1keepalived]# ss -tanl | grep :80
[iyunv@node1keepalived]#
在node2上进行查看
[iyunv@node2keepalived]# ss -tanl | grep :80
LISTEN 0 128 *:80 *:*
验证成功，说明实现了nginx的高可用服务
总结：要对某特定服务做高可用有两个要点
一是：要提供监控服务脚本
二是：在vrrp实例中追踪服务
修改配置文件keepalived.conf
vrrp_script chk_nginx {
script "killall -0 nginx"
interval 1
weight -2
}
vrrp_instanceVI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
172.16.2.8
}
track_script {
chk_maintainace
chk_nginx
}
在node2上做同样的修改
测试：
[iyunv@node2keepalived]# killall nginx
Youhave new mail in /var/spool/mail/root
[iyunv@node2keepalived]# ss -tanl | grep :80
[iyunv@node2keepalived]#
在node1上
[iyunv@node1keepalived]# ss -tanl | grep :80
LISTEN 0 128 *:80 *:*
验证成功
4、如何实现基于多虚拟路由的master/master模型？
要实现双主模型需要定义两个vrrp_instance，在node1的配置文件中要一下修改：
vrrp_instanceVI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
172.16.2.8
}
track_script {
chk_maintainace
chk_nginx
}
notify_master"/etc/keepalived/notify.sh master"
notify_backup "/etc/keepalived/notify.shbackup"
notify_fault "/etc/keepalived/notify.shfault"
}
vrrp_instance VI_2 {
state BACKUP
interface eth0
virtual_router_id 55
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 2111
}
virtual_ipaddress {
172.16.2.18
}
track_script {
chk_maintainace
chk_nginx
}
notify_master"/etc/keepalived/notify.sh master"
notify_backup"/etc/keepalived/notify.sh backup"
notify_fault "/etc/keepalived/notify.shfault"
}
在node2上做同样的修改，重启keepalived，进行测试
[iyunv@node1keepalived]# service nginx status
nginx(pid 28688) is running...
[iyunv@node1keepalived]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:4e:22:fb brdff:ff:ff:ff:ff:ff
inet 172.16.2.1/16 brd 172.16.255.255 scopeglobal eth0
inet 172.16.2.8/32 scopeglobal eth0
inet 172.16.10.8/16 brd 172.16.255.255scope global secondary eth0:0
inet6 fe80::20c:29ff:fe4e:22fb/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether 6a:7a:4f:e0:c1:8a brdff:ff:ff:ff:ff:ff
Youhave new mail in /var/spool/mail/root
在node2上
[iyunv@node2keepalived]# service nginx start
Startingnginx: [ OK ]
[iyunv@node2keepalived]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:74:c7:7b brd ff:ff:ff:ff:ff:ff
inet 172.16.2.16/16 brd172.16.255.255 scope global eth0
inet 172.16.2.18/32 scopeglobal eth0
inet6 fe80::20c:29ff:fe74:c77b/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether3a:4e:e8:4c:57:04 brd ff:ff:ff:ff:ff:ff
让node2的keepalived停掉，查看地址是否发生转移
[iyunv@node1keepalived]# ip addr show
1:lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2:eth0:
mtu 1500 qdisc pfifo_fast stateUP qlen 1000
link/ether 00:0c:29:4e:22:fb brd ff:ff:ff:ff:ff:ff
inet 172.16.2.1/16 brd172.16.255.255 scope global eth0
inet 172.16.2.8/32 scopeglobal eth0
inet 172.16.2.18/32 scopeglobal eth0
inet 172.16.10.8/16 brd 172.16.255.255scope global secondary eth0:0
inet6 fe80::20c:29ff:fe4e:22fb/64 scopelink
valid_lft forever preferred_lft forever
3:pan0:
mtu 1500 qdisc noop state DOWN
link/ether 6a:7a:4f:e0:c1:8a brdff:ff:ff:ff:ff:ff
Youhave new mail in /var/spool/mail/root
验证成功
总结：以上是我操作的过程，不足之处多多指点！！
</root@s"
</root@s"
</root@s"
|
|
|
QQ群⑧：
窥视卡
雷达卡
发表于 2013-9-29 08:43:58
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
千斤顶
显身卡