Red Hat RHCS Cluster System Configuration

Red Hat RHCS Cluster System Configuration

In short, a cluster is a group of computers that provide users with a set of network resources as a whole. These individual computer systems are cluster nodes ). An ideal cluster is that users never realize the underlying node of the cluster system. In his/her view, the cluster is a system, not multiple computer systems. In addition, administrators of the cluster system can add or delete nodes of the cluster system at will.

A high-availability cluster is an entity that connects several servers and provides failover functions through special software. Availability refers to the system's uptime. In the 7x24x365 working environment, 99% of the Availability refers to the downtime of 87 hours and 36 minutes in a year, generally, in key services, such a one-day failure time is unacceptable. Therefore, the concept of restoring errors mentioned above is proposed to meet 99.999% of high availability requirements.

Here are some concepts:

1. Services are resources provided in the HA cluster, including Float IP, shared storage, and apache.

2. A Member Server (Member Server) is also called a Node. It is the Server that actually runs services in HA to provide resources.

3. The Failover Domain is a collection of servers that provide resources In HA. When a member in HA fails, the service can be switched to another normal Member Server. In HA, an invalid domain generally contains two member servers (no virtual technology is applied ).

4. HeartBeat is a method for monitoring the status of member servers in HA. Generally, the HeartBeat is transmitted through a network cable and a serial line.

5. A Single Point Of failure (SPOF) is a part Of a system. When it fails or stops running, the entire system cannot work. In HA, dual-power supply, multiple NICs, and dual switches are usually used to avoid SPOF.

6. Quorum is a method used by HA to store Member Server Information in a shared disk to accurately determine whether the server and the services it provides are normal. The shared status information includes whether the cluster is active or not. The service status information includes whether the service is running and which member is running the service. Each member checks the information to keep other members updated. In a group with only two members, each member periodically writes a timestamp and cluster status information to the two shared cluster partitions in the shared disk storage area. To ensure proper cluster operations, if a member cannot write data to the primary shared cluster partition or block shared cluster partition at startup, it will not be allowed to join the cluster. In addition, if a cluster member fails to update its timestamp or heartbeats, the member will be deleted from the cluster.

7. When a fault occurs on one node, the Fence device restarts the faulty node through the Fence device, in this way, non-manual intervention is performed to prevent problematic nodes from accessing shared storage, resulting in file system conflicts. For Fence devices, there is an external power manager such as APC. many servers are built in, but the names of different manufacturers are different. For example, HP is called iLo, IBM is called BMC, and Dell is called DRAC.

LUCI--RHCS (RedHat Cluster Suite) provides a variety of cluster configuration and management tools, commonly used GUI-based system-config-Cluster, Conga and so on, also provides a command line-based management tools. System-config-cluster is a graphical management tool used to create clusters and configure cluster nodes. It consists of cluster node configuration and cluster management, create a cluster node configuration file and maintain the node running status. It is generally used in earlier versions of RHCS. Conga is a new network-based cluster configuration tool. Unlike system-config-cluster, Conga configures and manages cluster nodes on the web. Conga consists of two parts: luci and ricci. luci is installed on a stand-alone computer (that is, a server) for configuring and managing clusters. ricci is installed on each cluster node, luci communicates with each node in the cluster through ricci.

I. Construction of HA

Lab environment: RHEL6.5 iptables and selinux disabled

Three hosts: IP hostname

192.168.2.60 (server) server60.example.com

192.168.20.7 (node1) server67.example.com

192.168.20.8 (node2) server68.example.com

Modify the/etc/hosts of two nodes respectively.

192.168.20.7 server67.example.com

192.168.20.8 server68.example.com

Close fire wall # iptables-F

# Service iptables save

Disable selinux # setenforce 0
 

Synchronize system time of two hosts

# Yum install ntpdate

# Ntpdate 192.168.2.60

Install ricci on two node hosts

# Yum install ricci

Start ricci Service

# Service ricci start
 

Set to boot automatically

# Chkconfig ricci on

Set a password for ricci

# Passwd ricci

Restart the luci of the server

# Service luci restart

Stop luci... [OK]
Start luci... [OK]
Point your web browser to https://server60.example.com: 8084 (or equivalent) to access luci
Web Access https://server60.example.com: 8084/cluster/

Go to the webpage to edit

Log On with root

Log on with a local user

 

 

When logging on with root, click Admin in the upper right corner to authorize normal user lee.

In this case, you can create a cluster.

 

At this time, the host will be restarted at two nodes. This is why the ricci service is set to enable automatic startup and disable iptables.

After the creation is complete, the file/etc/cluster. conf is automatically written to the node.

<? Xml version = "2.0"?>
<Cluster config_version = "1" name = "leeypp">
<Clusternodes>
<Clusternode name = "server67.example.com" nodeid = "1"/>
<Clusternode name = "server68.example.com" nodeid = "2"/>
</Clusternodes>
<Cman expected_votes = "1" two_node = "1"/>
<Fencedevices/>
<Rm/>
</Cluster>

For more details, please continue to read the highlights on the next page:

1 2 3 Next Page