Corosync+pacemaker making a highly available Web cluster

Last Update:2014-09-16 Source: Internet

Author: User

Developer on Alibaba Coud: Build your first app with APIs, SDKs, and tutorials on the Alibaba Cloud. Read more ＞

Lab Environment:

Two hosts: centos6.5+httpd2.4+php5.5 the basic environment that makes up the web, and the Web page is accessed properly, and ensures that the HTTPD24 service does not boot up.

Node1.mylinux.com 10.204.80.79

Node2.mylinux.com 10.204.80.80

I am here to enable ansible to facilitate the management of two nodes, enable a host as a management node, ip:10.204.80.71, in the hosts of the three hosts to add the corresponding name resolution, management node to node1,node2 two nodes enable SSH trust mechanism, Add Node1 and Node2 to the Webservers group in the ansible configuration file.

Install the Software:

Installing Corosync and pacemaker on both hosts and installing the Yum module with the ansible command on the management host will take some time to install.

# ansible webservers-m yum-a ' Name=corosync,pacemaker state=installed '

650) this.width=650; "Style=" background-image:none; border-right-0px; margin:0px; padding-left:0px; padding-right:0px; border-top-width:0px; border-bottom-width:0px; border-left-width:0px; padding-top:0px "title=" image "border=" 0 "alt=" image "src=" http://img1.51cto.com/attachment/201409/16/287547_ 1410857765mxmo.png "width=" 635 "height=" 334 "/>

This shows that Corosync and pacemaker have already been installed, and here pacemaker is done as a plug-in for Corosync, and in the later versions of RHEL7, pacemaker will be run as a standalone daemon.

Configuration Corosync:

After installing Corosync, there will be a default corosync.config.example in the/etc/corosync directory, instead of Corosync.config

Add the Corosync.config configuration file back to pacemaker as a Corosync plugin. Add the following content

Service {
ver:0
Name:pacemaker
# Use_mgmtd:yes
}

aisexec {
User:root
Group:root
}

and set the IP address of bindnetaddr in this profile is my host's network address, node network segment in the 10.204.80 network segment, set to 10.204.80.0 segment bindnetaddr:10.204.80.0, the configuration file is as follows

# Please read the COROSYNC.CONF.5 manual Pagecompatibility:whitetanktotem {version:2secauth:onthreads:0interface {ring number:0bindnetaddr:10.204.80.0mcastaddr:226.94.13.10mcastport:5405ttl:1}}logging {fileline:offto_stderr:noto_ Logfile:yesto_syslog:yeslogfile:/var/log/cluster/corosync.logdebug:offtimestamp:onlogger_subsys {subsys: AMFDEBUG:OFF}}AMF {mode:disabled}service {      ver:0      name:pacemaker       # use_mgmtd:yes}aisexec {     User: Root     Group:root}

Note Here is BINDNETADDR address only set local network segment, otherwise there may be an error corosync [TOTEM] The network interface is down, so that the node heartbeat can not pass over. Refer to the official documentation for a very clear word.

Bindnetaddr
This specifies the network address the Corosync executive should bind to. For example, if the local interface are 192.168.5.92 with netmask
255.255.255.0, set Bindnetaddr to 192.168.5.0. If The local interface is 192.168.5.92 with netmask 255.255.255.192 and set Bindnetaddr to
192.168.5.64, and so forth.

The authentication key file when generating communication between nodes:

#corosync-kegen Press A few carriage returns, the Authkey key file is generated

Copy the corosync.conf and Authkey to the/etc/corosync/directory of the Node2 node.

To configure the interface:

In rhel6.3 before the commonly used version is CRMSH, is a Resource manager configuration interface, in the rhel6.4 after the use of PCs, here can be used in two, contrast. So Crmsh is not in centos6.5 's Yum library, so go to download install CRMSH or build your own installation Crmsh.

: http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-6/x86_64/

There are three files that need to be downloaded crmsh-2.1-1.6.x86_64.rpm, pssh-2.3.1-4.1.x86_64.rpm, python-pssh-2.3.1-4.1.x86_64.rpm; because CRMSH need to rely on PSSH, So also need to download

Install the Configuration interface: Install CRMSSH In addition to several files downloaded, there are other Python dependencies, so use Yum Localinstall to install to resolve dependencies. Put these three files in the/root directory and install

# yum--nogpgcheck localinstall crmsh-2.1-1.6.x86_64.rpm pssh-2.3.1-4.1.x86_64.rpm python-pssh-2.3.1-4.1.x86_64.rpm

Installation complete, enter CRM into CRM Shell

Start Corosync on the Node1

Start Node 2 on Node 1

Start complete, use the command # CRM status to view node status

Both node 1 and Node 2 have started normally, and the cluster is in a normal working state.

Perform the PS auxf command to see the related processes Corosync initiated

650) this.width=650; "Style=" background-image:none; border-right-0px; padding-left:0px; padding-right:0px; border-top-width:0px; border-bottom-width:0px; border-left-width:0px; padding-top:0px "title=" image "border=" 0 "alt=" image "src=" http://img1.51cto.com/attachment/201409/16/287547_ 1410857767a3s5.png "width=" 613 "height=" 103 "/>

To configure the working properties of a cluster:

Corosync enabled by default Stonith, the current cluster does not have a corresponding stonith device, so the first to disable Stonith

#crm Configure show Displays the configuration information for the current cluster, you can see that Stonith has been disabled

To add resources to a cluster:

Corosync supported resource agents are LSB and OCF, and LSB is the startup script class for some services, such as some scripts under the/etc/init.d/directory. To view the types supported by the current cluster system:

#crm RA Classes

Stonith is specifically designed for the configuration of Stonith devices

To view a list of all resource proxies under a type, you can use the #crm RA list LSB | OCF (Heartbeat | pacmemaker) | Stonith to view.

The resources required for a Web cluster are vip,web services,

The VIP of my two nodes is: 10.204.80.88 Web service is HTTP24

The syntax of the configuration command for CRM:

Primitive <rsc> [<class>:[<provider>:]]<type>

[Params attr_list]

[Operations Id_spec]

[Op op_type [<attribute>=<value> ...]

Op_type:: Start | Stop | Monitor

Primitive Master Resources

Format Primitive ID (resource name) resource Agent

Parametric params

Meta META attribute information

Utilization usage Information

Operations Operation information operation type start |stop | Monitor (monitoring, define if the resource stops if it will be discovered and passed to the cluster and whether it is re-shipped)

Add a ip=10.204.80.88 Master resource

Make the HTTPD24 resource a Web service resource. Detection service verify, and commit commit

View resource Status

The display has two nodes, two resource settings,

Because the resource is not defined as a group and no constraints are defined, the default two resources are run on two nodes respectively.

Define these two resources to the same group

In #crm then enter configure to enter configuration mode, then enter Help Group

You can view the Help documentation with a detailed description of the group command and an example

Group syntax

Group <name> <rsc> [<rsc> ...]
[Description=<description>]
[Meta Attr_list]
[Params attr_list]

Attr_list:: [$id =<id>] <attr>=<val> [<attr>=<val> ...] | $id-ref=<id>

Define WEBIP and Webserverf in the same group Webgroup, then view the cluster status

Two resources are running on the same node Node1, accessing the test

After the Node1 is turned off, the discovery resource is not transferred to the Node2 node, and the cluster status is checked #crm status

Found in the cluster status is already "without quorum" has lost the legal votes, is not in the cluster state, but I have only two nodes here, this is unreasonable, you can ignore the following command quorum the cluster status check cannot be satisfied;

# CRM Configure Property No-quorum-policy=ignore

View cluster status

The cluster status is already running on Node 2.

Web page Access OK

Start the Node1 node.

Node1 and Node2 are both online, but the resources are not transferred back to the Node1 node because the resource stickiness is not defined here

There are three constraints to defining resources:

Location constraint: Defines which nodes a resource can, cannot, or should run on as much as possible

Order order constraint: Constraints define the order in which cluster resources are started on nodes

colocation arrangement constraints: Arranging constraints to define cluster resources or to run concurrently on a node

I define here webgroup default run on Node1 node

#crm Configure Prefer_node1 webserver 200:node1.mylinux.com

Note Here is: the back with the node must have the exact interval, or will error,

View configuration Files

650) this.width=650; "Style=" background-image:none; border-bottom:0px; border-left:0px; margin:0px; padding-left:0px; padding-right:0px; border-top:0px; border-right:0px; padding-top:0px "title=" image "border=" 0 "alt=" image "src=" http://img1.51cto.com/attachment/201409/16/287547_ 1410857773b0pp.png "" 536 "height=" 293 "/>

Order constraints, to define the mandatory requirement to start WEBIP before starting the HTTP service, the command is as follows

#crm Configure Order Httpd-after-ip Mandatory:webip webserver

Corosync and pacemaker have other powerful features, if health checks, how often do you check, and so on, here comes the first

Corosync+pacemaker making a highly available Web cluster

This article is an English version of an article which is originally in the Chinese language on aliyun.com and is provided for information purposes only. This website makes no representation or warranty of any kind, either expressed or implied, as to the accuracy, completeness ownership or reliability of the article or any translations thereof. If you have any concerns or complaints relating to the article, please send an email, providing a detailed description of the concern or complaint, to info-contact@alibabacloud.com. A staff member will contact you within 5 working days. Once verified, infringing content will be removed immediately.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

Get Started for Free

Sales Support

1 on 1 presale consultation

Chat Contact Sales
After-Sales Support

24/7 Technical Support 6 Free Tickets per Quarter Faster Response

Open a Ticket
Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.

Learn More

Corosync+pacemaker making a highly available Web cluster

Contact Us

What's Trending

Top 10 Tags

Top 10 Keywords

A Free Trial That Lets You Build Big!

Sales Support

After-Sales Support

Corosync+pacemaker making a highly available Web cluster

Contact Us

What's Trending

Top 10 Tags

Top 10 Keywords

Trending Topic

A Free Trial That Lets You Build Big!

Sales Support

After-Sales Support