Corosyc + pcmake + drbd dual-web high availability solution

For websites that cannot be interrupted but have low access traffic, configuring two servers as one master and one slave is a commonly used method. This article is based on corosync + pacemaker + drbd to build a complete dual-LAMP high-availability cluster system

Create Basic Environment

First, check whether DRBD is supported by the kernel. My system is centos6.2. The default kernel version 6.2 Is 2.6.32-220. el6.i686. DRBD8.3 requires at least 2.6.33 Kernel support. Therefore, Kernel patches must be installed. I have a problem installing the kernel patch in centos6.2. There is no problem in the centos5.x series, so I actually compiled and installed it)

List of supported drbd versions

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44B006-0.jpg "style =" float: none; "title =" 1.jpg"/>

Sohu and 163 source do not have this 8.2 version downloaded from centos official source

Http://mirror.centos.org/centos/5/extras/i386/

Configure the hostname of both parties, write it to the hosts file, and establish ssh password-less communication between the two parties

The specific steps are not listed here

Hosts file

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4461463-1.jpg "title =" 2.jpg" style = "float: none;"/>

Step 1: DRBD

Create the disk partition to be synchronized

View Disk Partitions

[Root @ ha1 drbd. d] # fdisk/dev/sdb-l

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4463119-2.jpg "title =" 3.jpg" style = "float: none;"/>

The DRBD installation package is divided into the kernel part and the user space management part. Both packages need to be installed

[Root @ ha1 drbd] # yum localinstall kmod-drbd83-8.3.8-1.el5.centos.i686.rpm drbd83-8.3.8-1.el5.centos.i386.rpm

I actually install it by compiling. My configuration file is under/usr/local/etc. The path of the configuration file below takes the rpm package installation as an example. Centos6.2 system does not know why kernel patch installation is a problem)

Configure based on the sample configuration file

[Root @ ha1 drbd] # cp/usr/share/doc/drbd83-8.3.8/drbd. conf/etc/

Edit global configuration

[Root @ ha1 drbd] # cd/etc/drbd. d/

[Root @ ha1 drbd. d] # vi global_common.conf

Global {

Usage-count no;

}

Common {

Protocol C;

Handlers {

Pri-on-incon-degr "/usr/lib/drbd/notify-pri-on-incon-degr.sh;/usr/lib/drbd/notify-emergency-reboot.sh; echo B>/proc/sysrq-trigger; reboot-f ";

Pri-lost-after-sb "/usr/lib/drbd/notify-pri-lost-after-sb.sh;/usr/lib/drbd/notify-emergency-reboot.sh; echo B>/proc/sysrq-trigger; reboot-f ";

Local-io-error "/usr/lib/drbd/notify-io-error.sh;/usr/lib/drbd/notify-emergency-shutdown.sh; echo o>/proc/sysrq-trigger; halt-f ";

Fence-peer "/usr/lib/drbd/crm-fence-peer.sh ";

Split-brain "/usr/lib/drbd/notify-split-brain.sh root ";

Out-of-sync "/usr/lib/drbd/notify-out-of-sync.sh root ";

Before-resync-target "/usr/lib/drbd/snapshot-resync-target-lvm.sh-p 15 ---c 16 k ";

After-resync-target/usr/lib/drbd/unsnapshot-resync-target-lvm.sh;

}

Startup {

Wfc-timeout 120;

Degr-wfc-timeout 120;

}

Disk {

On-io-error detach;

Fencing resource-only;

}

Net {

Cram-hmac-alg "sha1 ";

Shared-secret "lustlost ";

}

Syncer {

Rate 100 M;

}

}

Define Resource Configuration

[Root @ ha1 drbd. d] # vi web. res

Resource web {

On ha1.lustlost.com {

Device/dev/drbd0;

Disk/dev/sdb1;

Address 172.16.93.148: 7789;

Meta-disk internal;

}

On ha2.lustlost.com {

Device/dev/drbd0;

Disk/dev/sdb1;

Address 172.16.93.149: 7789;

Meta-disk internal;

}

}

Copy the configuration file cp to ha2

[Root @ ha1 drbd. d] # scp/etc/drbd. conf ha2:/etc

[Root @ ha1 drbd. d] # scp/etc/drbd. d/* ha2:/etc/drbd. d/

Initialize Disk

Run on ha1 and ha2

Drbdadm create-md web

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44C502-3.jpg "title =" 4.jpg" style = "float: none;"/>

Start the service on both nodes

[Root @ ha2 drbd. d] # service drbd status

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4461L2-4.jpg "title =" 5.jpg" style = "float: none;"/>

After the startup, set one of the nodes as the master node, execute the following command, and the data will be automatically synchronized

[Root @ ha1 drbd. d] # drbdadm -- overwrite-data-of-peer primary web

View synchronization status

[Root @ ha1 drbd. d] # drbd-overview

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44625A-5.jpg "title =" 6.jpg" style = "float: none;"/>

After the synchronization is complete, check the status on the master node.

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4462225-6.jpg "title =" 7.jpg" style = "float: none;"/>

View the status of the slave Node

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44A943-7.jpg "title =" 8.jpg" style = "float: none;"/>

Format a drbd Device

[Root @ ha1 drbd. d] # mkfs. ext3/dev/drbd0

Mount to directory

[Root @ ha1 drbd. d] # mount/dev/drbd0/web/

Step 2: LAMP Platform

Yum is directly installed for convenience.

[Root @ ha1 ~] # Yum install httpd mysql-server php-mysql-y

Edit apache configuration. The directory points to the directory of the drbd mount point.

[Root @ ha1 ~] # Vi/etc/httpd/conf/httpd. conf

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4463C3-8.jpg "title =" 10.jpg" style = "float: none;"/>

Edit the mysql configuration. The data directory also points to the drbd mount point.

[Root @ ha1 ~] # Vi/etc/my. cnf

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z4464P9-9.jpg "title =" 11.jpg" style = "float: none;"/>

Drbd, Mysql, and httpd are configured as cluster services. If mysql and php are configured to be disabled after startup, the cluster service control is enabled and disabled.

[Root @ ha1 data] # chkconfig mysqld off

[Root @ ha1 data] # chkconfig httpd off

[Root @ ha1 data] # chkconfig drbd off

Configure the ha2 node as well.

Step 3: Corosync + pacemaker

Install corosync and pacemaker

[Root @ ha1 ~] # Yum install corosync pacemaker-y

Edit corosync global configuration

[Root @ ha1 ~] # Cd/etc/corosync/

[Root @ ha1 corosync] # vi corosync. conf

The configuration is as follows:

# Please read the corosync. conf.5 manual page

Compatibility: whitetank

Totem {

Version: 2

Secauth: off

Threads: 0

Interface {

Ringnumber: 0

Bindnetaddr: 172.16.93.148

Mcastaddr: 226.94.1.1

Mcastport: 5405

Ttl: 1

}

}

Logging {

Fileline: off

To_stderr: no

To_logfile: yes

To_syslog: yes

Logfile:/var/log/cluster/corosync. log

Debug: off

Timestamp: on

Logger_subsys {

Subsys: AMF

Debug: off

}

}

Amf {

Mode: disabled

}

Service {

Ver: 0

Name: pacemaker

}

Aisexec {

User: root

Group: root

}

Generate a key file

[Root @ ha2 corosync] # corosync-keygen

The system load pool may not have enough random numbers for a newly installed system or a system with short boot duration. You cannot use this command to generate a key file, you can use the find/command to make the system perform random I/O operations.

Copy the configuration file to ha2

[Root @ ha1 corosync] # scp-p corosync. conf authkey ha2:/etc/corosync/

Then go to the crm configuration page and say crm is really easy to use. All command options can be automatically completed, just like configuring an exchange route)

[Root @ ha1 corosync] # crm

Crm (live) # configure

Configure three global attributes

Crm (live) configure # property no-quorum-policy = ignore because it is a dual-node, the number of votes is ignored.

Crm (live) configure # property stonith-enabled = false disable the stonith Device

Crm (live) configure # rsc_defaults resource-stickiness = 100 improve resource stickiness and prevent resource transfer during server fault recovery

Plan the resources to be configured.

In dual-LAMP nodes, the following resources are available:

IP address, mysql, httpd, drbd, mount file system. All these resources must be on the same node. Therefore, a group is defined first. In addition, the sequence of starting these five resources must be clarified first.

First, define these resources.

Cluster virtual IP Address

Crm (live) configure # primitive webip ocf: heartbeat: IPaddr params ip = 172.16.93.150 broadcast = 255.255.255.0

Database

Crm (live) configure # primitive webmysql lsb: mysqld

Web Server

Crm (live) configure # primitive webserver lsb: httpd

Drbd Device

Crm (live) configure # primitive webdrbd ocf: heartbeat: drbd params drbd_resource = web

Drbd master-slave Definition

Crm (live) configure # MS ms_drbd webdrbd meta master-mak = "1" master-node-max = "1" clone-max = "2" clone-node-max = "1" notify = "true" mount Definition

Crm (live) configure # primitive webfs ocf: heartbeat: Filesystem params device =/dev/drbd0 directory =/web fstype = ext3 op start timeout = 60 s op stop timeout = 60 s

Specify constraints to constrain all resources together

Crm (live) configure # colocation myweb inf: webip webserver webmysql webfs ms_drbd: Master

Define order

Crm (live) configure # order webfs_after_drbd inf: ms_drbd: promote webfs: start # mount the file system only when drbd is promoted to the primary node

Crm (live) configure # order webmysql_after_webfs inf: webfs: start webmysql: start # mysql can be started only after the file system is mounted

Crm (live) configure # order webserver_after_webip inf: webip: start webserver: start # apache is started only after the IP address is configured.

Crm (live) configure # order webserver_after_webfs inf: webfs: start webserver: start

# Apache can be started only after the file system is mounted

View all the defined resources

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44A619-10.jpg "title =" 12.jpg" style = "float: none;"/>

View resource startup status

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131227/1Z44613A-11.png "title =" 13.png" style = "float: none;"/>

By disabling the ha1 Nic, all resources can be successfully transferred to ha2.

This article from the "lustlost-lost in desire" blog, please be sure to keep this source http://lustlost.blog.51cto.com/2600869/1177524