Win Server 2008 R2 three-step simplified fault cluster

According to a survey conducted by a foreign authority, one of the most influential factors in business is service downtime. Therefore, how to ensure the continuous operation of the business has become a common problem for IT personnel. The failover clustering service in Windows Server 2008 R2 guarantees the availability of core business servers to meet the business's need to keep critical systems up and running. IT168 reporters have experienced the cluster configuration in Win Server 2008R2 in the previous hands-on service, and are very impressed with the cluster function in Win Server 2008R2. In this afternoon's course, Microsoft further explained the enhancement of cluster function in Win Server 2008R2.

The lecturer of this lecture is Liu Gang, a senior lecturer at Oriental Ruitong (Beijing) Consulting Services Co., Ltd., FY08 Microsoft Gold Instructor, with MCT, MCSE, MCDBA, MCTS, MCITP, etc. Microsoft certification, in-depth study of Microsoft system products, with 9 years of training instructors and project implementation experience.

Liu Gang: The function of server cluster has been around for a long time. In his impression, the earliest enterprise version of Windows NT4.0 has a cluster service. At that time, the configuration required for configuring a cluster is required. Still quite high, you need two servers with the same configuration and separate storage devices. When two servers run the same set of applications, one is required as the host, the main application is assumed by this server, and the other server is used as the standby machine. After the host fails, it can take over the services of the host at any time. At that time, it was called hot standby, which was actually a very common solution.

Prior to the Windows 2003 release, this clustering feature was not officially named, generally called the MS Cluster Service. Since the 2003 release, Microsoft has had clustering capabilities. A clear name is called: Failure over cluster. By the 2008 R2 release, Microsoft has further enhanced the fault clustering capabilities. Includes: Cluster Shared Volume (CSV); Virtual Machine Migration (Live Migrate); 64+ CPU Kernel Support; Branch Cache; Direct Access.

Among them, Liu Gang is highly acclaimed for cluster shared volumes and virtual machine migration (Live Migrate), especially Live Migrate, which is very beneficial to the virtual machine environment and is very convenient for virtual machine migration. In fact, in Liu Gang’s impression, when the earliest enterprises want to deploy faulty clusters, they often need extremely complicated application configuration processes. Many enterprises do not have the ability to complete this configuration work independently, and they need to provide on-site services, and each time Microsoft conducts various trainings, clustering. Service is also one of the most important. Since the 2003 release, the configuration of cluster services has been greatly simplified, and many enterprise administrators have been able to configure cluster services independently.

In the 2008 R2 release, the cluster service not only continued the simplification and user-friendly configuration of the 2003 version, but also provided best practice fault detection. Because the steps involved in the entire cluster configuration process are very cumbersome, as long as one step is slightly wrong, the entire cluster will be unsuccessfully configured. To this end, Microsoft summarizes the common errors in the previous cluster configuration process, and introduced fault detection for these common errors, so that in any step there is a slight error, the fault detection will give a very clear prompt, tell the configuration staff There was a configuration error. This feature has been provided in the 2003 version, but the error report is not clear enough, so the reference for the configurator is not so great. The 2008 R2 version will make the detection process more user-friendly, with built-in best practice analysis and detection tools. Helps analyze errors and make the cluster configuration process easy to use.

Liu Gang gave a detailed introduction to the role of fault clustering in virtual machine migration, since the cluster in Server 2008 r2 provides cluster shared volume (CSV); virtual machine implementation migration (Live Migrate) The ability to move a running virtual machine from one Hyper-V host to another without shutting down the virtual machine. The virtual machine will not be aware of the background migration and maintain the client operating system. TCP does not interrupt the connection. Applying this technology to virtual data centers can increase business agility, reduce costs, increase productivity, and keep clients connected to the server.

The difference from fast migration is: The virtual machine is saved and then restored on the target host. The application/workload running in the virtual machine is temporarily saved

and this function The implementation relies on the new cluster shared volume feature. In traditional cluster mode, a volume open can only be occupied by one service application, but the cluster shared volume (CSV) actually provides distributed file access resolution for Hyper-V. The solution enables multiple nodes to simultaneously access the same "real" shared volume. For a virtual machine, which node actually owns the LUN is completely transparent, the virtual machine is moved without modifying the ownership of the drive, and the virtual machine can Share CSV disks to reduce LUNs and enable live migration of virtual machines.