HA dual-machine cluster

HA is the abbreviation of Highly Available (highly available) , which is the abbreviation of the dual-machine cluster system. It improves the availability of the cluster and is an effective solution to ensure business continuity. Generally, there are two or more nodes, and they are divided into active nodes and standby nodes. .

Usually, the one that is executing the business is called the active node, and the one that is a backup of the active node is called the standby node . When a problem occurs on the active node and the running business (task) cannot run normally, the standby node will detect it and immediately take over the active node to execute the business. This enables non-interruption or brief interruption of business.

application

LATCH HA is a highly reliable server cluster system solution of the Linux Latch series.

The smooth functioning of every industry today relies on computer systems. Product ordering, distribution, customer service, financial settlement, insurance and reservations all require highly reliable information service systems. This naturally requires the computer downtime to be as short as possible.

LATCH HA can automatically detect system faults and has an intelligent detection and processing system for system crashes . It uses various means to detect failures in hardware, storage devices, networks and system software. Once a computer failure is detected, LATCH HA will transfer all processes on the computer to the backup computer to provide customers with uninterrupted services. LATCH HA uses these methods to provide the best service reliability and reduce computer downtime to the lowest level.

Once the computer goes down, LATCH HA can shorten the downtime to minutes or even seconds. And it can restore the system to the state before the downtime, thereby ensuring the stable operation of the computer. Under the protection of LATCH HA, computer systems no longer need to be interrupted due to hardware failures or routine maintenance. You know, one minute of downtime may bring immeasurable losses to the operations of banks, railways, aviation and other industries.

Another key to achieving high availability is the redundancy strategy, which in short means multiple backups of hosts, network devices, and backup devices (of course not just simple backup machines).

LATCH HA has multiple redundancy modes, including dual-machine online standby mode, dual-machine ready mode and three-machine mode. All modes provide support for various backup resources such as file systems, databases, network IP addresses, system processes, SCSI devices, and NFS.

In addition, the realization of high availability should also have rich application program support: LATCH HA can be applied in various centralized, client-server mode or OLTP systems.

At the same time, it is compatible with various mainstream database systems and OLTP software on the market - such as Oracle, SYBASE, Informix and Tuxedo. LATCH HA also provides various application program interfaces.

Therefore, customers can integrate various functions into their proprietary software to ensure high system reliability, thus also protecting customers' investments.

heartbeat Linux cluster system

There are mainly two types of cluster systems we use:

High Availability (High Availability) HA cluster is implemented using Heartbeat; it is also called "dual-machine hot backup", "dual-machine mutual backup", and "dual-machine".

Load Balance Cluster , implemented using Linux Virtual Server (LVS);

The working principle of heartbeat (Linux-HA): The core of heartbeat includes two parts, the heartbeat monitoring part and the resource takeover part . Heartbeat monitoring can be performed through network links and serial ports, and it supports redundant links and sends messages to each other. message to tell the other party its current status. If no message is received from the other party within the specified time, the other party is considered invalid. At this time, the resource takeover module needs to be started to take over the resources or services running on the other party's host.

Guess you like

Origin blog.csdn.net/fuhanghang/article/details/132759160