Foreword
High Availability Cluster, or HA Cluster for short, refers to server cluster technology for the purpose of reducing service interruption time. It can be seen from the above that the LVS cluster itself cannot achieve high availability. For example, the Director Server cannot detect the health of the Real Server. Once one or all of them are Real Server is down, Director Server will continue to forward requests, causing the site to be inaccessible. Similarly, if Director Server downtime site is even more unlikely to function normally. This article will explain how to implement LVS cluster high availability based on heartbeat v1.
Heartbeat
Introduction
Heartbeat is a Linux-HA project One of the components, since 1999, has released many versions, which is the most successful example of the current open source Linux-HA project, and has been widely used in the industry.
Working principle
Heartbeat’s core consists of two parts, the heartbeat monitoring part and the resource takeover part. Heartbeat monitoring can be carried out through the network link and the serial port, and supports redundant links. Send messages to each other to tell the other party about their current status. If the message sent by the other party is not received within the specified time, then the other party is considered to be invalid. At this time, the resource takeover module needs to be activated to take over the resources running on the other party’s host. Or service.
Achieve high availability of LVS cluster based on Heartbeat v1
< span style="font-family:'Microsoft YaHei','Microsoft YaHei';color:rgb(0,112,192);">Experimental topology
Experimental environment
node1:node1.scholar.com 172.16. 10.123 CentOS6.6
node2:node2.scholar.com 172.16.10.124 CentOS6.6
Real Server1:192.168.1.10 (VIP) 172.16.10.125 (RIP) CentOS6.6
Real Server2: 192.168.1.10 (VIP) 172.16.10.126 (RIP) CentOS6.6
Notes
The premise of configuring a high-availability cluster: ( Take the two-node heartbeat as an example)
①Time must be synchronized
Use ntp server
②Nodes must communicate with each other by name
Resolve node names
/etc/host
The host name used in the cluster is the host name represented by `uname -n`
③ping node
Required only for even nodes
④ssh Key authentication for communication
Configuration process
Time synchronization
span>
Please ensure that the time of the two nodes is synchronized, here No more detailed description
Analysis name configuration
[root@node1 ~]# vim/etc/hosts172.16.10. 123 node1.scholar.com node1172.16.10.124 node2.scholar.com node2[root@node1~]#vim/etc/sysconfig/networkHOSTNAME=node1.scholar.com[root@node1~]#uname-nnode1.scholar. com# Both nodes need to be operated as above
ssh key Configuration
[root@node1~]# ssh-keygen-trsa-P''[root@node1~]#ssh-copy -id -i.ssh/id_rsa.pub root@node2[root@node2~]# ssh-keygen-trsa-P''[root@node2~]#ssh-copy-id-i.ssh/id_rsa.pub root@node1[root@node1~]# date; ssh node2'date' #Test Sun Jun 7 17:46:03 CST 2015 Sun Jun 7 17:46:03 CST 2015
Install the required software package
#Resolve dependencies[root@node1 ~]#yuminstallperl-TimeDatenet-snmp- libs libnet PyXML -y #Need epel source support[root@node1~]#cd heartbeat2[root@node1heartbeat2]#lsheartbeat-2.1.4-12.el6.x86_64.rpm heartbeat-pils-2.1.4-12.el6 .x86_64.rpmheartbeat-gui-2.1.4-12.el6.x86_64.rpm heartbeat-stonith-2.1.4-12.el6.x86_64.rpmheartbeat-ldirectord- 2.1.4-12.el6.x86_64.rpm[root@node1 heartbeat2]# rpm-ivh heartbeat-2.1.4-12.el6.x86_64.rpm heartbeat-pils-2.1.4-12.el6.x86_64.rpm heartbeat -stonith-2.1.4-12.el6.x86_64.rpm[root@node1 heartbeat2]# yum install heartbeat-ldirectord-2.1.4-12.el6.x86_64.rpm -y# Both nodes perform the above operationsConfigure heartbeat
Prepare the configuration file
Configure algorithm key
[root@node1 ~]#openssl rand-hex 84d8fd6cb49d2047b[root@node1~]#vim/etc/ha.d/authkeysauth22sha14d8fd6cb49d2047bConfigure the main configuration file
[root@ node1~]#grep-v"#"/etc/ha.d/ha.cf|grep-v"^$"logfile /var/log/ha-log#Log storage location keepalive 2#Specify the heartbeat use interval deadtime 30 Broadcast #Designated backup node to take over the service resource timeout time of the master node warntime 10 #Specify the time of the heartbeat delay initdead 120 #Resolve network start delay udpport 694 #694 0cast 0 m #1 694 broadcast port # 694 0 m # 694 Define whether the service will be automatically switched back to node node1.scholar.com when the master node is restored #Master node node node2.scholar.com #Standby node ping 172.16.0.1 #arbitration equipmentConfiguration Explorer
[root@node1 ~ ]# vim /etc/ha.d/haresources node1.scholar.com 192.168.1.10/32/eth0/192.168.1.10 ldirectord::/etc/ha.d/ldirectord.cfpass the configuration file to the backup node
Configure ldirectord
Prepare the configuration file and configure
[root@node1 ~]# cp /usr/share/doc/heartbeat-ldirectord-2.1.4/ldirectord.cf/etc/ha.d/[root@node1~]#grep- v ^# /etc/ha.d/ldirectord.cfche cktimeout=3 #Detection timeout time checkinterval=1 #Detection interval time autoreload=yes #Modify the configuration file, and reload without restarting the service quick=yes #real server Deleted from the lvs list after downtime, and automatically added to the list after recovery virtual=192.168.1.10:80 #VIP real=172.16.10.125:80 gate #real server real=172.16.10.126:80 gate #real server fallback=127.0.0.1:80 gate #If the RS nodes are all down, enable local Loopback port address service=http protocol detection request=".health.html" #detection file receive="ok" #Detection content, judge whether RS is alive scheduler=rr=255 #Scheduling algorithm #persist=255 .255.255Transfer the configuration file to the standby node, and prohibit each node ldirectord from enabling self-starting
Prepare the fallback file
RS settings
#The two RSs are configured as followsConfigure kernel parameters
Prepare health detection files and site files
Test page
< img onload="if(this.width>650) this.width=650;" src="/wp-content/uploadshttp:/img.voidcn.com/vcimg/static/loading.png" title="7.jpg "alt="wKioL1V0Q-bSXEvkAABeL0Q43g0048.jpg" d="4196266" s="b7d_2fd" t="jpg">
Start heartbeat
Check if the resource is effective
High availability test
Refresh the page
Simulate the downtime of RS2
#service httpd stop
No matter how you refresh it, it is the RS1 page
Next we will also stop RS1
Switch back tofallback page
We stop node1 and see if node2 can take over resources
[root@node1 ~]# service heartbeat stop
The resource is successfully taken over, and the above access will not be affected in any way. Of course, if node1 is started this time, the resource will be grabbed again In the past, there will be no demonstration here
The high availability of LVS based on heartbeat v1 is now complete
The end
Well, This is the end of the implementation of LVS high availability based on heartbeat v1. The whole process is quite fun. You can leave a message if you encounter problems during the configuration process. The next article will explain it based on Heartbeat v2 high-availability cluster, you can continue to pay attention if you are interested. The above is only for personal study, if there are mistakes or omissions, please don’t spray~~~
