I. Introduction
The first two sections were introduced Linux high availability clustering and load balancing cluster, you can also combine these two, that is, high-availability load balancing cluster Keepalived + LVS component, Keepalived added to the LVS in the following reasons:
1) LVS has a very crucial role in Dir (distributor), if the distributor is down, all the services and access all will be interrupted. Because the entrance all in the Dir, so it is necessary to do dispatcher highly available, use Keepalived to achieve high availability, load balancing Keepalived in fact, have a role.
2) When using LVS, if there is no other additional operations, to which a Real Server shutdown, go visit will be a problem, because the LVS still will forward the request to the server has downtime, Keepalived emergence can be resolved this problem, even if there is a Real Server is down, it can provide normal service, when the request came distribute it automatically detects Real Server back-end down, then it will not continue to forward the request to question Real Server on.
At least one full Keepalived + LVS architecture requires two scheduler (role Dir), were installed Keepalived to achieve high availability. Keepalived built ipvsadm function, so no need to install ipvsadm package, do not write and execute lvs_dir script.
Second, create Keepalived + LVS high-availability load balancing cluster
Load balancing to achieve high availability cluster using the following Keepalived + LVS.
2.1 Prepare the cluster nodes
Preparation servers 4, wherein two distributor as standby (also called scheduler, referred to the Dir), Keepalived services are installed, the other two are Real Server, nginx services are installed, as the user requests the server process.
Hostname masternode: master scheduler, network card is 192.168.93.140 IP within the network (VMWare Network Address Translation NAT).
Hostname backupnode: backup scheduler, the card is 192.168.93.139 IP within the network (VMWare Network Address Translation NAT).
Hostname datanode1: Real Server 1, IP is 192.168.93.141.
Hostname datanode2: Real Server 2, IP is 192.168.93.142.
VIP：192.168.93.200
4 servers are set up IP address and host name, and turn off the firewall.
2.2 Set Keepalived master primary server
Edit the configuration file /etc/keepalived/keepalived.conf, set the following:
[@ masternode the root keepalived] # Vim / etc / keepalived / keepalived.conf ! the Configuration File for Master Node keepalived global_defs { notification_email { [email protected] } notification_email_from [email protected] smtp_server 127.0 . 0.1 smtp_connect_timeout 30 the router_id of the LVS_DEVEL } vrrp_instance VI_1 { State the mASTER # is bound VIP card ens33 interface ens33 virtual_router_id 51 is # priority of the primary server 100 priority 100 advert_int . 1 {authentication AUTH_TYPE the PASS AUTH_PASS moonxy } virtual_ipaddress { 192.168 . 93.200 } } virtual_server 192.168 . 93.200 80 { # once every 10 seconds to check the state of Real Server delay_loop . 6 #LVS load balancing algorithm lb_algo RR #LVS the DR mode lb_kind the DR # same IP the connection is assigned to the same Real Server how many seconds, set here to 0 persistence_timeout 0 # TCP protocol status check Real Server protocol TCP real_server 192.168 . 93.141 80 { # Weight weight 100 TCP_CHECK { # 10 seconds without response timeout connect_timeout The 10 nb_get_retry . 3 delay_before_retry . 3 connect_port 80 } } real_server 192.168 . 93.142 80 { # weights weight 100 TCP_CHECK { # 10 seconds without response timeout connect_timeout The 10 nb_get_retry . 3 delay_before_retry . 3 connect_port 80 } } }
Good configuration settings master configuration file and backup server are substantially the same, the state is set to the BACKUP, the priority is set to 90
2.3 Setting backup backup server Keepalived
Edit the configuration file /etc/keepalived/keepalived.conf, set the following:
[@ backupnode the root keepalived] # Vim / etc / keepalived / keepalived.conf ! the Configuration File for Master Node keepalived global_defs { notification_email { [email protected] } notification_email_from [email protected] smtp_server 127.0 . 0.1 smtp_connect_timeout 30 the router_id of the LVS_DEVEL } vrrp_instance VI_1 { State the bACKUP # binding to VIP card ens33 interface ens33 virtual_router_id 51 is # priority setting backup server 90 priority 90 advert_int . 1 {authentication AUTH_TYPE the PASS AUTH_PASS moonxy } virtual_ipaddress { 192.168 . 93.200 } } virtual_server 192.168 . 93.200 80 { # once every 10 seconds to check the state of Real Server delay_loop . 6 #LVS load balancing algorithm lb_algo RR #LVS the DR mode lb_kind the DR # same IP the connection is assigned to the same Real Server how many seconds, set here to 0 persistence_timeout 0 # TCP protocol status check Real Server protocol TCP real_server 192.168 . 93.141 80{ # Weight weight 100 TCP_CHECK { # 10 seconds without response timeout connect_timeout The 10 nb_get_retry . 3 delay_before_retry . 3 connect_port 80 } } real_server 192.168 . 93.142 80 { # weights weight 100 TCP_CHECK { # 10 seconds without response timeout connect_timeout The 10 nb_get_retry . 3 delay_before_retry . 3 connect_port 80 } } }
If the LVS script performed before, you need to do something in advance:
# ipvsadm -C
# systemctl restart network
Or may restart the scheduling server.
2.4 Start-related services
In masternode start respectively above and backupnode keepalived service, as follows:
masternode：
[root@masternode log]# systemctl start keepalived [root@masternode log]# ps aux |grep keepalived root 7204 0.0 0.1 122876 1368 ? Ss 00:46 0:00 /usr/sbin/keepalived -D root 7205 0.0 0.3 133840 3360 ? S 00:46 0:00 /usr/sbin/keepalived -D root 7206 0.0 0.2 133776 2636 ? S 00:46 0:00 /usr/sbin/keepalived -D root 7215 0.0 0.0 112708 980 pts/0 R+ 00:58 0:00 grep --color=auto keepalived [root@masternode log]# ip addr 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:3b:90:07 brd ff:ff:ff:ff:ff:ff inet 192.168.93.140/24 brd 192.168.93.255 scope global noprefixroute ens33 valid_lft forever preferred_lft forever inet 192.168.93.200/32 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::1bb9:5b87:893c:e112/64 scope link noprefixroute valid_lft forever preferred_lft forever 3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:3b:90:11 brd ff:ff:ff:ff:ff:ff inet 192.168.150.140/24 brd 192.168.150.255 scope global noprefixroute ens37 valid_lft forever preferred_lft forever inet6 fe80::20c:29ff:fe3b:9011/64 scope link valid_lft forever preferred_lft forever [root@masternode log]# ipvsadm -Ln IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.93.200:80 rr -> 192.168.93.141:80 Route 100 0 0 -> 192.168.93.142:80 Route 100 0 0
backupnode：
[root@backupnode keepalived]# systemctl start keepalived [root@backupnode keepalived]# ps aux |grep keepalived root 7185 0.0 0.1 122876 1376 ? Ss 00:47 0:00 /usr/sbin/keepalived -D root 7186 0.0 0.3 133840 3364 ? S 00:47 0:00 /usr/sbin/keepalived -D root 7187 0.0 0.2 133776 2640 ? S 00:47 0:00 /usr/sbin/keepalived -D root 7189 0.0 0.0 112708 980 pts/0 S+ 00:47 0:00 grep --color=auto keepalived [root@backupnode keepalived]# ip addr 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:88:29:9b brd ff:ff:ff:ff:ff:ff inet 192.168.93.139/24 brd 192.168.93.255 scope global noprefixroute ens33 valid_lft forever preferred_lft forever inet6 fe80::8055:62bc:4a07:d345/64 scope link tentative noprefixroute dadfailed valid_lft forever preferred_lft forever inet6 fe80::1bb9:5b87:893c:e112/64 scope link tentative noprefixroute dadfailed valid_lft forever preferred_lft forever inet6 fe80::8cac:8f3b:14b2:47ae/64 scope link noprefixroute valid_lft forever preferred_lft forever 3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:88:29:a5 brd ff:ff:ff:ff:ff:ff inet 192.168.150.139/24 brd 192.168.150.255 scope global noprefixroute ens37 valid_lft forever preferred_lft forever inet6 fe80::20c:29ff:fe88:29a5/64 scope link valid_lft forever preferred_lft forever [root@backupnode keepalived]# ipvsadm -Ln IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.93.200:80 rr -> 192.168.93.141:80 Route 100 0 0 -> 192.168.93.142:80 Route 100 0 0
At this point you can see VIP: 192.168.93.200 bound to the network card ens33 masternode of.
Since Keepalived LVS pattern defined in the configuration file for DR, so it needs to be performed on two RS lvs_rs.sh script respectively, which is the same as a script, as follows:
[root@datanode1 sbin]# service nginx start Redirecting to /bin/systemctl start nginx.service [root@datanode1 sbin]# sh /usr/local/sbin/lvs_rs.sh ...... [root@datanode2 sbin]# service nginx start Redirecting to /bin/systemctl start nginx.service [root@datanode2 sbin]# sh /usr/local/sbin/lvs_rs.sh
After the completion of the relevant service starts, executable test.
2.5 Test high availability load balancing cluster
VIP address request on the host: 192.168.93.200, as follows:
We can see realizes load balancing.
1) Test off both schedulers maternode master scheduler
Let's look at the distributor of high availability, close masternode main distributor, as follows:
You can see masternode has unbound the VIP address, backupnode automatically bound by the VIP, as follows:
VIP address request again, as shown below normal:
2）测试关闭 Real Server 中的 datanode2 的 nginx 服务
现在我们再关闭 RS 中的 datanode2 的 nginx，如下：
访问 VIP 地址，看到了 Keepalived 已经自动将所有请求都转发到了 datanode1 上面了，如下：
ipvsadm 转发规则会自动将 "删除" datanode2，如下：
3）测试再次启动 Real Server 中的 datanode2 的 nginx
如果再次启动 datanode2 上面的 nginx 服务，此时 ipvsadm 又会自动添加其转发规则，如下：
访问 VIP，如下：
可以发现 Keepalived 功能确实很强大。