Heartbeat
System environment:
Firewall off,
Seliux status is disabled
Address resolution for groups between servers
node3:172.25.44.3
node4:172.25.44.4
Physical Real Machine: 172.25.44.250
vip:172.25.44.100
The following content is configured on both NODE3 and NOD4
Installing the Software
heartbeat-libs-3.0.4-2.el6.x86_64.rpm
heartbeat-3.0.4-2.el6.x86_64.rpm
ldirectord-3.9.5-3.1.x86_64.rpm
heartbeat-devel-3.0.4-2.el6.x86_64.rpm
Rpm-q heartbeat-d # #查看软件生成的配置文件
cd/usr/share/doc/heartbeat/
CP ha.cf Haresources AUTHKEYS/ETC/HA.D
1).
vim/etc/ha.d/ ha.cf # # # #需要添加或更改的内容, according to the specific circumstances and settings, parameters are not fixed
Logfile/var/log/ha-log
2 keepalive
30 Deadtime
10 Warntime
60 Initdead
738 Udpport
Bcast eth0
157 Auto_failback on
211 Node Server3.example.com
212 Node server4.example.com
Ping 172.25.44.250 #已存在的ip地址, non-node address,
253 respawn Hacluster/usr/lib64/heartbeat/ipfail
259 Apiauth Ipfail Gid=haclient uid=hacluster
#########################################
Master configuration file (/ETC/HA.D/HA.CF):
Debugfile/var/log/ha-debug
Debug Log file file, default value
Logfile/var/log/ha-log
system run log file , take default value
Logfacility local0
log level , default value
KeepAlive 2
heart rate , set yourself. 1:1 seconds; 200ms: 200 ms
Deadtime 30
node death time threshold, that is, from the node after 30 has not received the heartbeat that the master node died, set himself
Warntime 10
issue a warning time , set yourself
Initdead 120
after the daemon starts for the first time, it should wait 120 seconds before starting resources on the primary server
Udpport 694
heartbeat messaging UDP port, using port 694 for Bcast and Ucast communication, taking the default value
#baud 19200
serial baud rate , used in conjunction with serial.
#serial/DEV/TTYS0
Use serial port to transmit heartbeat information.
Bcast eth1
use UDP broadcast to notify Heartbeat
#ucast eth1 10.0.0.3 uses the UDP unicast of the NIC Eth1 to notify the heartbeat, eth1 IP
#mcast eth0 225.0.0.1 694 1 0
use UDP multicast to notify the heartbeat
Auto_failback on
when the primary node resumes , does it automatically cut back
#stonith Baytech/etc/ha.d/conf/stonith.baytech
Stonith to ensure data integrity in a shared storage environment
Watchdog/dev/watchdog
Watchdog can allow the system to restart the machine after 1 minutes of failure, this function can help the server really stop the heart
after jumping, you can recover your heartbeat again. If you use this feature , modify the/etc/modprobe.conf in the system to add the following
Options Softdog nowayout=0
This will load the "Softdog" kernel module in the kernel to generate the actual device files when the system is booted.
/dev/watchdog
Node station3.example.com
The primary node name , consistent with Uname–n. The first default is the primary node, so don't make the order
Node station4.example.com
Sub-node name , consistent with Uname–n
Ping 192.168.0.254
Respawn Hacluster/usr/lib/heartbeat/ipfail
Apiauth Ipfail gid=haclient Uid=hacluster
The default heartbeat does not detect any services other than itself, nor does it detect network conditions.
So when the network is interrupted , there is no switch between Load Balancer and Backup.
You can solve this problem by setting the ' ping nodes ' through the Ipfail plugin, but you cannot use a cluster node as the ping node.
###########################
2).
vim/etc/ha.d/ haresources
650) this.width=650; "title=" Picture 1.png "alt=" Wkiol1gmpfxizsb0aabfgyhk49o486.png "src=" http://s1.51cto.com/wyfs02/M01 /89/35/wkiol1gmpfxizsb0aabfgyhk49o486.png "/>
Server3.example.com Ipaddr::172.25.44.100/24/eth0 httpd
The virtual IP is 172.25.44.100, and the boot service is httpd
########################
resource File (/etc/ha.d/haresources):
station3.example.com Ipaddr::192.168.0.200/24/eth0 Ldirectord httpd
This file defines the startup scripts for each software required to implement the cluster , which must be placed in the/ETC/INIT.D or/ETC/HA.D/RESOURCE.D directory ipaddr the role of the Virutal IP, which is Heartbeart a script that comes with it; Ldirectord's role is to start the Ldirectord monitoring program, which will make the LVS defined in ldirectord.cf take effect.
and monitor its health ; HTTPD is the startup script for the Apache service.
3).
VIM/ETC/HA.D /authkeys
650) this.width=650; "title=" Picture 2.png "alt=" Wkiom1gmpgvbssgkaaael0odnei144.png "src=" http://s2.51cto.com/wyfs02/M02 /89/37/wkiom1gmpgvbssgkaaael0odnei144.png "/>
chmod 600/etc/ha.d/authkeys
transfer files to Node4 after configuration is complete
#scp ha.cf haresources Authkeys 172.25.44.4:/etc/ha.d/
######## #认证文件, this file permission must be 600#######
Test:
1.Writing test files (writing different content on Node3 and node4 for easy testing)
Vim/var/www/html/index.html
2.#/etc/init.d/heartbeat Start # # #l两边都开启hearbeat服务
Starting high-availability Services:INFO:Resource is stopped
Done.
#arp-an | grep 172.25.44.100 # #查看vip在哪个节点上运行 (executes this command on a physical real machine)
Viewing VIPs on the master node
#ip Addr Show
3. Browser access to vip:172.25.44.100, View Web content
4.NODE3 manually shut down the heartbeat service, to see if the VIP jumps to Node4, while httpd on the node4, Access 172.25.44.100, if the content and node Node4 test file content consistent, test success.
This article is from the "11897256" blog, please be sure to keep this source http://65147718.blog.51cto.com/11897256/1864791
Linux Heartbeat high-availability simple configuration