Server Dictionary: What is a cluster server?

Cluster, the English name is Cluster. In layman's terms, clustering is a technology that connects multiple systems together so that multiple servers can work like a machine or look like a machine. The use of cluster systems is usually to improve the stability of the system and the data processing capabilities and service capabilities of the network center.

Take, for example, we have erected a WWW server, built on top of an e-commerce site, but as time goes on, more and more fame, then it hits more High, the load of the WWW server is getting higher and higher. In this case, we must upgrade the capabilities of the WWW server to meet the growing demand for service. At this time, we are faced with two options:

1) upgrade WWW server with a faster CPU, adding more memory, it has a stronger performance; but growing service requests will The server is overloaded again and needs to be upgraded again, thus falling into the strange circle of upgrade. Also, when upgrading, you have to consider how the service can be connected, can you stop!

2) increase WWW server, allowing multiple servers to accomplish the same service.

The second method is a cluster of servers, in short, is a group of independent servers behave as a single system on the network, and be managed in a single mode system. This single system provides highly reliable service to customer workstations. A server cluster consists of multiple servers with shared data storage space. Each server communicates with each other through an internal LAN. When one of the servers fails, the applications it runs will be automatically taken over by other servers; In this case, all computers in the cluster have a common name, and any server in the cluster system can be used by all network users.

server running in a cluster system is not necessarily high-end products, but the server cluster it can provide quite a high performance non-stop service; each server can bear part of computing tasks, and because of the cluster The performance of multiple servers, therefore, the computing power of the overall system will be improved; at the same time, each server can also undertake certain fault-tolerant tasks. When one of the servers fails, the system can use the special software to support this. The server is isolated from the system and achieves new load balancing through a load transfer mechanism between the servers, while alerting the system administrator.

common server clustering method


Currently, there are two common methods server cluster, one is connected to the backup server to the master server when the primary server fails When the backup server is put into operation, all the tasks on the primary server are taken over. Another approach is to connect multiple servers that share the same application and database computing tasks and improve the response time of critical large applications. At the same time, each server also undertakes some fault-tolerant tasks. Once a server fails, the system can isolate the server from the system with the support of the system software, and complete the new load distribution through the load transfer mechanism of each server. A common cluster of PC servers is a cluster of two servers. A UNIX system can support a cluster system of eight servers. Compaq's dedicated system OpenVMS can support a cluster system of up to 96 servers.

In a clustered system, all computers have a common name, and services running on any system in the cluster can be used by all network customers. The cluster must be able to coordinate the management of errors and failures of separate components and transparently add components to the cluster. The user's public data is placed in a shared disk enclosure, and the application is installed on all servers, that is, applications running on the cluster need to be installed on all servers. When the cluster system is running normally, the application runs on only one server, and only this server can manipulate the data area of ​​the application on the shared disk cabinet. Other servers monitor the server as long as the application on the server stops. Operation (whether it is hardware damage, operating system crash, application software failure, or application failure caused by human error), other servers will take over the application running on this server and take over the corresponding data area on the shared disk cabinet. come. The specific takeover process is performed in three parts, the system takes over, loads the application, and connects the client.

advantage


a cluster system, the cluster system addresses all of the server hardware failure, when one of the servers any failure, such as: hard drives, memory, CPU , motherboard, I /O board and power failure, the application running on this server will switch to other servers.

Second, the cluster system software system can solve the problem, we know that, in a computer system, is used by the user application and data, and applications running on top of Caozuojitong, Caozuojitong and running on the server on. In this way, as long as any one of the application system, the operating system, and the server fails, the system actually stops providing services to the client. For example, one of the common software crashes is one of the cases, although the server hardware is intact. However, the server still cannot provide services to the client. The biggest advantage of the cluster is that the monitoring of the faulty server is application-based. That is, as long as the application of the server stops running, other related servers will take over the application, regardless of the reason why the application stops running.

Third, the cluster system can address the situation stopped working application system caused by human error, for example, when a server administrator improper operation of the server is down, so running applications on this server is also It stopped running. Since the cluster is monitoring the application, other related servers will take over the application. Insufficient




cluster system we know of only Application Clusters running on a single server, if the application fails, the other will be a server restart This application takes over the data area on the shared disk enclosure to get the application up and running again. We know that the takeover process for the entire application generally requires three steps: detecting and confirming the failure, the backup server restarting the application, and taking over the shared data area. Therefore, it takes a certain amount of time in the process of switching. In principle, the switching time varies according to the size of the application, and the longer the application switching time is.