Server Cluster

Server clustering refers to the number of servers together with the same service, it seems as if the client is only a cluster of servers can take advantage of multiple computers for parallel computing to obtain very high computing speed, you can also use multiple computers to do backup, allowing any machine is broken or the entire system to normal operation. Server cluster Once installed on the server and run the Cluster service, the server can join the cluster. Clustering operation can reduce the number of single points of failure, and to achieve the high availability of clustered resources. The following sections briefly describe the operation of cluster creation and cluster node behavior. Note: For information on installing a cluster server, see the Windows server 2003 product family to help and Deployment Guide. Creating a cluster Product contains a server cluster to install the cluster software on the server and create a new cluster, cluster installation utility. Create a new cluster, the first choice as the first member of the cluster on a computer running the utility. The first step is to determine the cluster name and create the cluster database and initial cluster membership list to define a new cluster. Windows server 2003 Cluster Management Cluster has a new setup wizard and command line interface created using the cluster.exe (from remote to create) a cluster of features. The second step is to create a cluster, add all cluster members to use for shared data storage devices. This will create a new cluster with one node, its own local data storage devices and cluster shared resource - usually a disk or data storage and connection media resources. The final step is to create a cluster in another cluster member will be running on every computer setup utility. Whenever a new node to the cluster, the new node from the cluster are automatically members of the original copy of the database access to existing cluster. When a node joins or forms a cluster, the Cluster service updates the node's private copy of the configuration database. Form a cluster If the server is running the Cluster service and can not find the other nodes in the cluster, which they can form a cluster. To form a cluster, nodes must be able to obtain exclusive rights to the quorum resource. When the initial formation of a cluster, the cluster will include the first node cluster configuration database. Whenever a new node joins the cluster, the new node will obtain and maintain a local copy of cluster configuration database. Quorum resource recovery log with (which has nothing to do with the same node cluster configuration and state data) stored in the form of the latest version of the configuration database. The cluster is running the Cluster service uses the quorum recovery logs to do the following: to ensure that only one set of activities, the node can communicate with each other to form a cluster Only when a node can get right to the quorum resource control, it allows it to form a cluster Only when a node can be the same communication node controls the quorum resource, it allows it to join or remain in the existing cluster Other nodes from the cluster and the cluster service management interface point of view, when the formation of a cluster, each node in the cluster may be in three different states in one. Event handler will record the state, event log manager will be copied to the cluster of these states to other nodes. Cluster service states are: Offline. At this point the node is not completely valid cluster member. The node and its cluster server may be running, it may not run. Online. At this point the node is completely valid cluster member. It comply with the cluster database updates, the arbitration algorithm to exert their influence to maintain the heartbeat communication, and can own and run resource groups. Suspended. At this point the node is completely valid cluster member. It comply with the cluster database updates, the arbitration algorithm to exert their influence to maintain the heartbeat communication, but it can not accept resource groups. It can only support those it currently has a resource group. The reason why the state to provide suspended, in order to allow certain maintenance. Most of the server cluster components will be suspended as online and equivalent state. Superiority First, the cluster system to resolve all server hardware failures, when any failure of a server, such as: hard drives, memory, CPU, motherboard, I / O board and a power failure, running on this server application will switch to other servers. Second, the cluster system software solve the problem, we know that in the computer system, users are using applications and data, and applications running on the operating system, the operating system they run on the server. In this way, as long as applications, operating systems, any one of the three servers fails, the system actually stopped providing services to clients, such as our common software crashes, is one of this, even though the server hardware intact, but the server still can not provide services to clients. The biggest advantage of the cluster servers for fault monitoring is based on the application, that is, as long as the application server stops running, other associated server will take over the application, without regard to the application to stop running what are the reasons.
 Third, the cluster system can solve the human error caused by the application system stops working, for example, when an administrator to a server improper operation of the server goes down, so running on this server application will stop running. As the cluster is to monitor the application, so the server will take over the other related to this application. Shortcoming We know that the application of the cluster is only running on one server, if this application fails, the other a server will restart the application to take over the ark of the data in the shared disk area, thereby enabling the application of normal operation. We know that over the entire application process generally requires three steps: to detect and confirm the failure, restart the backup server applications over the shared data area. Therefore, in the process of switching will take some time, in principle, different sizes depending on the application of the switching time will be different, the larger the longer the application switch.