Introduction to Linux cluster principle and installation configuration

Linux cluster principle

Linux cluster system includes cluster node and cluster Manager. Cluster nodes, sometimes referred to as nodes, servers or server nodes, are the systems that provide processing resources, and they do the actual work of the cluster. In general, it must be configured to become part of the cluster, and it must run the cluster's application software. Applications can be software that is dedicated to clustering or standard software designed for distributed systems. The Linux cluster Manager is a logical structure that bundles nodes together to form the appearance of a single system, which is used to decompose tasks into all nodes.

Clusters have different types for a variety of reasons, the most direct cause of the Linux cluster is the sharing of CPU resources, balancing the load between multiple machines, obtaining more system reliability, or providing a backup system (toggle) when the host fails.

Through the analysis of the related software, the function of the cluster load is realized through the traffic management, and there are several implementation methods: direct routing (directly forwarding), network address translation (NAT) and tunneling Technology (tunneling).

Direct forwarding is a method that can be used when the computer participating in the cluster is on the same network segment as the computer that is managing the control. When a control-managed computer receives a request package, it is sent directly to the node participating in the cluster. Its advantage is to return to the customer flow without the control of the host, fast, less cost.

Perhaps you are more familiar with this method of network address translation (NAT). The address converter has a legitimate IP address that can be accessed by the outside world, and it modifies the address of the outgoing packet from the proprietary network, which appears to be from the address translator itself. When the outside package is sent to the converter, it can determine which node should send the packet to the intranet. Its advantage is to save the IP address, can camouflage the internal, the disadvantage is inefficient, because the flow back to the requester to go through the converter.

Tunneling Technology (Tunneling) This method is the forwarding mechanism used by Linux cluster nodes when they are not in the same network segment, and is to encapsulate IP packets in other network traffic. From a security perspective, you should use VPN in tunneling technology, or you can use leased lines.

The services that a Linux cluster can provide are TCP/ip-based Web services, mail services, news services, DNS services, and proxy servers. I'm going to use a specific product Turbolinux Cluster Server to implement a load-balanced Linux cluster system that provides Web and FTP services.

Linux cluster planning

1. Services provided: Web, FTP.

2. Do a more complete load balancing system so that more functions can be used.

3. Use 4 servers, of which 3 install Turbolinux Cluster server,1 to install Windows Sever.

Linux cluster installation

1. Installing Turbo Linux on 3 servers also requires Apache and WU-FTPD to be installed because the Linux cluster provides this service. After the installation is complete, reboot the machine, and connect the optical drive under the directory/mnt/cdrom, execute. /tlcs-install, and then follow the prompts to complete the installation.

2. Install Windows Server on 1 servers, and install Internet Information Server 5.0.

Configuring the Linux Cluster Manager

1. Set the IP address, subnet mask, routing, etc. of each server, and set up 1 turbolinux servers as DNS servers so that they can forward parsing and reverse parsing. This example server is named PC1 and the domain is test.com.

2. Configure the cluster Server. Perform turbolinuxclusteradmin, set as follows (hyphen is connected to the menu option or its subordinate menu, after the colon is set up):

(1) Clusterserver configuration-cluster services-application stability Agents:

HTTP is the default service and is not set.

Ftp:/usr/lib/ftpagent

(2) Clusterserver configuration--cluster services--service Settings:

Http,80:tcp,sticky.

Ftp,21:tcp,ftp.

(3) Clusterserver configuration--servers Configuration:

PC1 (pc1.test.com), direct,ping

PC2 (pc2.test.com), direct,ping

PC3 (pc3.test.com), direct,ping

PC4 (pc4.test.com), direct,ping

(4) Clusterserver Configuration--advance Traffic Managers:

Advance Traffic Manager System:pc1.test.com

Advance traffic Manager Setting: Default value

(5) Clusterserver configurationàvirtual severs:

Host as: pc1.test.com

SENDMAIL:MASTER@PC1. test.com

Server Pool Name:server Group1

(6) Clusterserver configurationàgloble Settings:

Network settings: netmask 255.255.255.0

Configuring nodes in the cluster

1. Configure Windows Server node

Because Turbolinux Cluster server itself can be automatically synchronized by tools, you can configure Windows Server only.

(1) Start → set → control Panel → add New hardware → next → add/remove equipment failure → add new device → No, then select hardware → other equipment from the list →microsoft:microsoft loopback adapter→ complete.

(2) Right-click on the desktop to select the Network Neighborhood → attribute →tcp/ip, set IP address, default gateway, subnet mask (first set to 255.255.255.0).

(3) Start → run →regedit→ find entries in the registry that are related to Microsoft Loopback Adapter and change the subnet mask to 255.255.255.255.

(4) Configure the system to run the appropriate service and configure the configuration appropriate for Cluster Manager management so that it can be used in Control Manager.

2. Configure Turbo Linux Cluster server node

(1) Select Tlcs_content _sync to perform content synchronization in the Admin menu, enter the node machine password to be configured, and copy the service content from the cluster Manager to the node.

(2) Select Tlcs_ Config _sync in the admin menu to perform settings synchronization, enter the node machine password to be configured, and copy the settings from the cluster Manager to the node.

By this, we can see the running state on the cluster manager PC1, connect the client to the switch on the server, and the client can request the Web and FTP service.