Apsara Stack: high-availability cluster in Hyper-V Environment (1)

According to relevant statistics, server integration is the main reason for enterprises to deploy virtualization. Although server integration helps users improve resource utilization, it also has advantages such as energy saving and centralized management, but it also brings great risks. In the physical environment, for stability and security considerations, a server usually runs only one application. Once the server fails, only the services provided by the application will be affected; after virtualization, multiple application servers run on one physical server in the form of virtual machines. Although they are independent of each other, they are like "putting all the eggs in the same basket ", in the event of a hardware failure, all virtual machines will stop running. In other words, a server failure may cause large-scale service paralysis.

Server virtualization vendors are aware of the risks after virtualization, and focus on shortening the downtime and ensuring the high availability of servers, such as the HA of VMware Virtual Infrastructure 3, and LiveRecovery of Virtual Iron V4 XEE, Microsoft's Hyper-V is no exception, especially, hyper-V's high availability feature can be integrated with Windows Server 2008's built-in Failover Cluster, when the server fails, the service will be quickly switched to another normal server, reducing the unplanned downtime caused by the accident to a minimum.

Compared with the high availability provided by other Server Virtualization Technologies, Windows Server 2008 Hyper-V failover clusters have the following advantages:

1. It is a standard feature of Windows Server 2008 Enterprise/Datacenter. Hyper-V is regarded as a service supported by the cluster and does not require additional purchase authorization.
2. A maximum of 16 nodes are supported. during testing, you can manually specify the transferred node. when a disaster occurs, the cluster system will automatically find and move the node to a node with sufficient memory capacity.
3. Compared with the MSCSMicrosoft Cluster Service earlier than Windows Server 2003), Failover Cluster greatly simplifies the setup and management steps, allowing new users to quickly familiarize themselves with how to operate

Hyper-V failover cluster architecture

Before deploying a Hyper-V failover cluster, you must first understand the architecture of the Failover cluster. The basic failover cluster components must contain two physical servers, a shared storage device, and a SAN Environment. This architecture is called a 2-node failover cluster such as 0 ). The physical Server must be fully compatible with the 64-bit processor of Windows Server 2008. We strongly recommend that you use two servers of the same model and specification. Shared storage devices generally refer to disk arrays on the FC or iSCSI interfaces. However, these devices are generally inexpensive and have limited functions and scalability. A shared Storage device, called Storage Server, has emerged in recent years. It fully integrates industry-standard servers and Storage virtualization software because of its complete functions and reasonable prices, with high scalability, it has become the first choice for many enterprises to build virtual environment storage systems.

Taking Apsara stack NSSFalconStor Network Storage Server as an example, it combines the powerful computing capabilities of quad-core, dual-processor x64 servers, as well as large-capacity, low-cost SATA, and SAS hard disks as Storage resources, it provides the same or even higher I/O performance as a medium/high disk array. In addition, advanced features such as snapshots, images, and automatically streamlined Thin Provisioning are available for intermediate and advanced disk arrays. Apsara NSS also supports remote replication, extends the local failover Cluster to a remote disaster recovery data center and becomes a Multi-site Cluster of multiple platform clusters, that is, Hyper-V Remote Disaster Recovery.