LVS/DR＋keepalived負載均衡實現
一、keepalived簡介
keepalived是分散式部署解決系統高可用的軟體，結合lvs(LinuxVirtual Server)使用，解決單機宕機的問題。
keepalived是一個基於VRRP協議來實現IPVS的高可用的解決方案。對於LVS負載均衡來說，如果前端的調度器direct發生故障，則後端的realserver是無法接受請求並響應的。因此，保證前端direct的高可用性是非常關鍵的，否則後端的伺服器是無法進行服務的。而我們的keepalived就可以用來解決單點故障（如LVS的前端direct故障）問題。keepalived的主要工作原理是：運行keepalived的兩台伺服器，其中一台為MASTER，另一台為BACKUP，正常情況下，所有的數據轉換功能和ARP請求響應都是由MASTER完成的，一旦MASTER發生故障，則BACKUP會馬上接管MASTER的工作，這種切換時非常迅速的。
二、測試環境
下面拿4台虛擬機進行環境測試,實驗環境為centos6.6 x86_64,具體用途和ip如下
三、軟體安裝
1、安裝lvs所需包ipvsadm
yum install -y ipvsadm
ln -s /usr/src/kernels/`uname -r` /usr/src/linux
lsmod |grep ip_vs
#注意Centos 6.X安裝lvs，使用1.26版本。並且需要先安裝yuminstall libnl* popt* -y
執行ipvsadm(modprobe ip_vs)把ip_vs模塊載入到內核
[root@test85 ~]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
#IP Virtual Server version 1.2.1 ---- ip_vs內核模塊版本
2、安裝keepalived
yum install -y keepalived
chkconfig keepalived on
註：在centos7系列系統中開機自動啟動使用systemctl enable keepalived
四、keepalived配置
先看下214.85keepalived主機上的配置
[root@test85 ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
#配置報警郵箱
notification_email {
charles@test.com
}
notification_email_from reportlog@test.com
smtp_server mail.test.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state MASTER #指定keepalived的角色，MASTER表示此主機是主伺服器，BACKUP表示此主機是備用伺服器
interface eth0 #指定HA監測網路的介面
lvs_sync_daemon_inteface eth0
virtual_router_id 55
#虛擬路由標識，這個標識是一個數字，同一個vrrp實例使用唯一的標識。即同一vrrp_instance下，MASTER和BACKUP必須是一致的
priority 100 #定義優先順序，數字越大，優先順序越高，在同一個vrrp_instance下，MASTER的優先順序必須大於BACKUP的優先順序
advert_int 1 #設定MASTER與BACKUP負載均衡器之間同步檢查的時間間隔，單位是秒
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress { #設置虛擬IP地址
192.168.214.89
}
}
virtual_server 192.168.214.89 80 {
delay_loop 6 #(每隔6秒查詢realserver狀態)
lb_algo rr #(lvs 演算法)
lb_kind DR #(使用lvs的DR模式)
#nat_mask 255.255.255.0
persistence_timeout 10 #(同一IP的連接10秒內被分配到同一台realserver)
protocol TCP #(用TCP協議檢查realserver狀態)
real_server 192.168.214.87 80 {
weight 100 #(權重)
TCP_CHECK {
connect_timeout 3 #(3秒無響應超時)
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.214.88 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
再看下214.86keepalived備機上的配置
! Configuration File for keepalived
global_defs {
notification_email {
charles@test.com
}
notification_email_from reportlog@test.com
smtp_server mail.test.com
smtp_connect_timeout 30
router_id LVS_DEVEL
}
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
lvs_sync_daemon_inteface eth0
virtual_router_id 55
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.214.89
}
}
virtual_server 192.168.214.89 80 {
delay_loop 6
lb_algo rr
lb_kind DR
#nat_mask 255.255.255.0
persistence_timeout 10
protocol TCP
real_server 192.168.214.87 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.214.88 80 {
weight 100
TCP_CHECK {
connect_timeout 3
connect_port 80
nb_get_retry 3
delay_before_retry 3
}
}
}
五、後端realserver操作
DR模式需要在後端真實機上運行以下腳本
#!/bin/bash
# description: Config realserver lo
#Written by :Charles
VIP1=192.168.214.89
. /etc/rc.d/init.d/functions
case 「$1」 in
start)
ifconfig lo:0 $VIP1 netmask 255.255.255.255 broadcast $VIP1
/sbin/route add –host $VIP1 dev lo:0
echo 「1」 >/proc/sys/net/ipv4/conf/lo/arp_ignore
echo 「2」 >/proc/sys/net/ipv4/conf/lo/arp_announce
echo 「1」 >/proc/sys/net/ipv4/conf/all/arp_ignore
echo 「2」 >/proc/sys/net/ipv4/conf/all/arp_announce
sysctl –p >/dev/null 2>&1
echo 「RealServer Start OK」
;;
stop)
ifconfig lo:0 down
route del $VIP1 >/dev/null 2>&1
echo 「0」 >/proc/sys/net/ipv4/conf/lo/arp_ignore
echo 「0」 >/proc/sys/net/ipv4/conf/lo/arp_announce
echo 「0」 >/proc/sys/net/ipv4/conf/all/arp_ignore
echo 「0」 >/proc/sys/net/ipv4/conf/all/arp_announce
echo 「RealServer Stoped」
;;
*)
echo 「Usage: $0 」
exit 1
esac
exit 0
#執行realserver.sh start開啟，stop關閉
#腳本設置成755許可權，並放入rc.local下讓其開機啟動運行
六、啟動keepalived服務及查看相關信息
在214.85和214.86上分別啟動keepalived服務
在214.85keepalived主機上查看信息
通過ip addr可以看到vip地址已經綁定在eth0網口上
[root@test85 ~]# ip addr
1: lo: mtu65536 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 state UP qlen1000
link/ether 00:0c:29:85:7a:67 brd ff:ff:ff:ff:ff:ff
inet 192.168.214.85/24 brd 192.168.214.255 scope global eth0
inet 192.168.214.89/32 scope global eth0
inet6 fe80::20c:29ff:fe85:7a67/64 scope link
valid_lft forever preferred_lft forever
在214.85上查看日誌信息，看到已成功進入keepalived主機模式
[root@test85 ~]# tail -f /var/log/messages
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1)Entering MASTER STATE
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:12:34 test85 Keepalived_healthcheckers[7975]:Netlink reflector reports IP 192.168.214.89 added
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:12:34 test85 Keepalived_vrrp[7977]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:12:36 test85 ntpd[1148]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:12:36 test85 ntpd[1148]: peers refreshed
May 4 14:12:39 test85 Keepalived_vrrp[7977]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:12:40 test85 root[7924] 192.168.5.80 53823 192.168.214.85 22: ipaddr
在214.86上查看日誌信息，看到已成功進入keepalived備機模式
May 4 14:12:37 web86 Keepalived_vrrp[31009]: Using LinkWatch kernel netlinkreflector...
May 4 14:12:37 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1)Entering BACKUP STATE
May 4 14:12:37 web86 Keepalived_vrrp[31009]: VRRP sockpool: [ifindex(2),proto(112), unicast(0), fd(10,11)]
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Opening file"/etc/keepalived/keepalived.conf".
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Configuration isusing : 14713 Bytes
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Using LinkWatchkernel netlink reflector...
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Activatinghealthchecker for service [192.168.214.87]:80
May 4 14:12:37 web86 Keepalived_healthcheckers[31007]: Activatinghealthchecker for service [192.168.214.88]:80
後端真實機啟動腳本後，查看網卡信息，看到vip已成功綁定在迴環口上。
[root@web87 ~]#ipaddr
1: lo: mtu65536 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
inet 192.168.214.89/32 brd 192.168.214.89 scope global lo:0
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0:
mtu 1500 qdisc pfifo_fast state UP qlen1000
link/ether 00:0c:29:38:31:ad brd ff:ff:ff:ff:ff:ff
inet 192.168.214.87/24 brd 192.168.214.255 scope global eth0
inet6 fe80::20c:29ff:fe38:31ad/64 scope link
valid_lft forever preferred_lft forever
通過ipvsadm –L –n查看相應lvs連接信息
[root@test85 ~]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.214.89:80 rr persistent 10
-> 192.168.214.87:80 Route 100 2 2
-> 192.168.214.88:80 Route 100 0 0
七、keepalived測試
使用vip地址192.168.214.89訪問後端192.168.214.87和192.168.214.88的頁面
正常訪問沒問題後，我們來模擬lvs集群故障
首先把keepalived master主機214.85宕機，看備機能否接管過來，vip地址是否會漂移過來
在214.86上查看日誌，發現備機成功切換成了主機狀態，vip地址成功漂移了過來
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1)Transition to MASTER STATE
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) EnteringMASTER STATE
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for192.168.214.89
May 4 14:35:35 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:35:36 web86 ntpd[1230]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:35:36 web86 ntpd[1230]: peers refreshed
May 4 14:35:40 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
然後，再把214.85主機恢復，由於214.85擁有較高的優先順序，會從214.86搶回MASTER狀態，相應的214.86會回歸到原來的Backup狀態
214.85日誌記錄，重新回到了MASTER狀態
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) TransitiontoMASTER STATE
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Receivedlower prio advert, forcing new election
May 4 14:41:55 test85 Keepalived_vrrp[8066]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) EnteringMASTER STATE
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:41:56 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:41:56 test85 Keepalived_healthcheckers[8064]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:41:58 test85 ntpd[1148]: Listen normally on 8 eth0 192.168.214.89UDP 123
May 4 14:41:58 test85 ntpd[1148]: peers refreshed
May 4 14:42:01 test85 Keepalived_vrrp[8066]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
218.86日誌記錄,接收到了高優先順序請求，從之前的MASTER狀態變回了BACKUP狀態
May 4 14:35:34 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to MASTER state
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1)Entering MASTER STATE
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) settingprotocol VIPs.
May 4 14:35:35 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:35:35 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 added
May 4 14:35:36 web86 ntpd[1230]: Listen normally on 7 eth0 192.168.214.89UDP 123
May 4 14:35:36 web86 ntpd[1230]: peers refreshed
May 4 14:35:40 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) Sendinggratuitous ARPs on eth0 for 192.168.214.89
May 4 14:36:41 web86 root[30963] 192.168.5.80 53824 192.168.214.86 22: ipaddr
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1)Received higher prio advert
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) EnteringBACKUP STATE
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Instance(VI_1) removingprotocol VIPs.
May 4 14:41:55 web86 Keepalived_vrrp[31009]: VRRP_Group(VG1) Syncinginstances to BACKUP state
May 4 14:41:55 web86 Keepalived_healthcheckers[31007]: Netlink reflectorreports IP 192.168.214.89 removed
May 4 14:41:56 web86 ntpd[1230]: Deleting interface #7 eth0,192.168.214.89#123, interface stats: received=0, sent=0, dropped=0,active_time=380 secs
最後，再模擬後端真實機214.87服務宕掉，看是否vip只請求214.88
通過日誌查看得知，keepalived集群探測到後端真實機214.87的80埠不通，把它從vip請求列表中移除了
May 414:48:00 test85 Keepalived_healthcheckers[8064]: TCP connection to[192.168.214.87]:80 failed !!!
May 4 14:48:00 test85Keepalived_healthcheckers[8064]: Removing service [192.168.214.87]:80 from VS[192.168.214.89]:80
當重新探測到後端真實機214.87服務恢復後，又把它加入了請求列表中
May 4 14:52:55 test85 Keepalived_healthcheckers[8064]: TCP connection to[192.168.214.87]:80 success.
May 4 14:52:55 test85 Keepalived_healthcheckers[8064]: Adding service[192.168.214.87]:80 to VS [192.168.214.89]:80
TAG:運維講堂 |