High Availability cluster in Linux

High Availability cluster in Linux

1. Highly available clusters are simple

650) this.width=650; "Src=" Http://s4.51cto.com/wyfs02/M00/85/03/wKioL1eZrQvgMrjlAAB9HRK7wVo857.jpg-wh_500x0-wm_3 -wmp_4-s_221954575.jpg "title=" 1.jpg "alt=" Wkiol1ezrqvgmrjlaab9hrk7wvo857.jpg-wh_50 "/>

1.1.Messaging Layer: The main collection node of the transaction resources Heartbeat and other information, respectively, there are the following:

HeartbeatV1 heartbeatV2 heartbeatV3

Corosync

Cman

Keepalived

Ultramokey

1.2.crm:cluster Resourse Manager manages the resources collected by the messaging layer, respectively, in the following ways:

Heartbeat v1 comes with a resource manager: Haresources

Heartbeat v2 comes with Explorer: Haresources and CRM

Heartbeat V3: Resource Manager CRM developed as a standalone project, pacemaker

1.3.ra:resource Agent, Resource agent, by LSB compliant script (start,stop,restart,status)

The main feature is that the CRM Resource manager allocates the resources to be executed by RA

RA Classes:ra Type

Legacy Heartbeat v1 RA

LSB (/etc/rc.d/init.d/)

OCF (Open Cluster Framework): With Pacemaker and Linbit (DRBD)

STONITH Isolation

Resource Type: Resource type

Primitive: Basic type

Clone: Clone

Group: Groups

MASTER/SLAVE:DRBD, master-slave relationship running on two nodes

Resource stickiness: The degree to which a resource depends on a point, defined by score

Resource constraint: Constraint

Position constraint location: The degree of resource-to-node inclination

Permutation constraint coloation: inter-resource dependency

Sequential constraint order: The order in which the resource takes action

2. Highly Available clusters

2.1. Three basic requirements

Recommendations for name resolution between nodes are written under the/etc/hosts file

SSH communication between nodes, it is recommended to use key authentication

Time synchronization between nodes recommended using NTP server

2.2. Three configuration files:

Authkeys: Key file, 600,

Configuration of the Ha.cf:heartbeat service

Haresources: Resource Management configuration file

3. The background of this experiment: the dual cluster node is node1.willow.com,ip 1.1.1.18 node2.willow.com,ip for 1.1.1.19

The node1.willow.com host is configured as follows: (all the same as the node2.willow.com cluster configuration)

3.1. Installing the HA cluster package

#yum--nogpgcheck Localinstall perl-mailtools-1.77-1.el5.noarch.rpm

#yum--nogpgcheck Localinstall heartbeat-2.1.4-9.el5.i386.rpm heartbeat-gui-2.1.4-9.el5.i386.rpm heartbeat-pils-2.1.4-10.el5.i386.rpm heartbeat-stonith-2.1.4-10.el5.i386.rpm libnet-1.1.4-3.el5.i386.rpm

#cd/usr/share/doc/heartbeat-2.1.4/

#cp-P Authkeys ha.cf haresources/etc/ha.d/

#chmod Authkeys

#cd/etc/ha.d/

3.2.vim Authkeys #新增如下內容:

Auth 1

1 MD5 willow! #其中willow is the salts that MD5 needs.

3.3.vim ha.cf #修改成如下內容:

Logfacilitylocal0

KeepAlive 1

Bcasteth0# Linux

Auto_failback on

Nodenode1.willow.com

Nodenode2.willow.com

Ping 1.1.1.144

3.3.vim haresources #修改成如下內容:

Node1.willow.com Ipaddr::1.1.1.100/24/eth0 httpd

#yum-y Install httpd

#echo "1.willow.com

#chkconfig httpd off

#setenforce 0

#iptalbes-F

The 3.4.node2.willow.com host modifies only the following sections (others are identical to the node1.willow.com cluster configuration, omitted)

#echo "2.willow.com

3.5. Simple Cluster Test:

Under the node1.willow.com host:

#serivce Heartbeat Start

#ssh Node2 ' serivce heartbeat start '

#tail-F/var/log/messages #看日志

At this point, you can see the effect through Web Access VIP1.1.1.100

#cd/usr/lib/heartbeat

#./ha_stanby #模拟node1节点掉线, the web automatically goes to the Node2 node for service

At this point the Web Access to the VIP1.1.1.100 and look at the effect

3.6. Share Web content via NFS server 1.1.1.20 the third Linux host, then cluster test

To turn on NFS sharing on a 1.1.1.20 host:

#mkdir-pv/web/ha

#echo "

#vim/etc/exports

/web/ha 1.1.1.0/24 (RO)

#service NFS Start

#service Portmap Start

On the Node1 host: Stop the Heartbeat service

#ssh Node2 ' serivce heartbeat Stop '

#serivce Heartbeat Stop

On Node1 and Node2 two hosts: same-in-the-Go modification haresources file

#vim haresources #修改如下内容

node1.willow.com Ipaddr::1.1.1.100/24/eth0 Filesystem::1.1.1.20:/web/ha::/var/www/html::nfs httpd

Again on the Node1 host: Start the heartbeat service

#serivce Heartbeat Start

#ssh Node2 ' serivce heartbeat start '

#tail-F/var/log/messages #看日志

At this point, you can see the effect through Web Access VIP1.1.1.100

#cd/usr/lib/heartbeat

#./ha_stanby #模拟node1节点掉线, the web automatically goes to the Node2 node for service

At this point the Web Access to the VIP1.1.1.100 and look at the effect

will continue to update this blog .....

This article is from the "Xavier Willow" blog, please be sure to keep this source http://willow.blog.51cto.com/6574604/1831812

High Availability cluster in Linux