VMware VSphere 5.0 High availability and dual machine hot standby

High availability and dual-machine hot spares are the most important part of VMware vsphere 5.0, where high availability is not unique to vsphere, and enterprise use of high availability is for service continuity and data security. HA (High availability) is a group of ESXi server-based clustering features, the main purpose is when the virtual machine running host failure can be timely transfer of the host, to avoid prolonged downtime. While the FT (Fault tolerance) dual standby is to ensure that the maximum time for the virtual machine is not down, the virtual machine in a dual-machine hot-standby mode at the same time in two hosts, so that can greatly enhance the business continuity. This chapter explains Ha's high availability and FT dual-machine hot-standby capabilities.

One, VSphere ha high Availability

VSphere ha is a feature-rich product that continuously monitors all physical servers in a resource pool, restarts virtual machines affected by server failures, monitors and checks virtual machine "guest operating system" failures, and automatically launches virtual machines after a user-specified interval, and can take advantage of a "heartbeat" on the server : To automatically detect server failures, and to restart virtual machines on other physical servers in the same resource pool almost immediately without human intervention. By choosing the best physical server in a resource pool to restart virtual machines on VSphere ha (if used with VMware DRS), VSphere HA provides universally applicable and cost-effective failover protection for virtualized IT environments. Protect applications that do not have other failover options and make software applications that might otherwise not be protected from being placed in high availability. When an operating system failure is detected, the virtual machine is automatically restarted to protect the VM from failure (in the experimental phase). This protection of VSphere ha creates the first strong barrier to the entire IT infrastructure.

(i) Ha in vsphere

We know the benefits of HA through the above introduction. So what exactly is ha? HA is high availability by minimizing downtime due to routine maintenance operations (planning) and sudden system crashes (unplanned) to improve system and application availability. It differs from the fault-tolerant technology that is considered uninterrupted. Ha system is the most effective manual for enterprises to prevent the core computer system from downtime. VSphere ha allows ESXi host collections to work together as a group that provides virtual machines with a higher level of availability than an ESXI host provides separately. When you plan for creation and use in a new vsphere ha cluster, the options that you select affect how the cluster responds to host or virtual machine failures.

A vSphere ha cluster typically has a logical queue that includes two or more ESXi hosts. In an HA cluster, each VMware ESXi server is equipped with an HA agent that continuously detects heartbeat signals from other hosts in the cluster. The ESXi host sends a heartbeat signal every 5 seconds through the service host's network connection. If an ESXi host does not signal a heartbeat after 3 consecutive intervals, the host is defaulted to a failure or a connection to the network has occurred.

When an undetectable condition occurs, the virtual machines that were originally running on that host are automatically transferred to the other hosts in the cluster. Conversely, if a host fails to receive a heartbeat from one of the hosts in the cluster, the host initiates an internal process to detect the problem with its connection to other hosts in the cluster. If there is a problem, then all running virtual machines on this host will be interrupted and a pre-configured standby host can be started.

In fact, another notable feature of VSphere ha is the ability to fail over multiple ESXi servers in a cluster. For a vsphere ha failover, the client operating system considers a reboot due to a hardware crash and does not detect an orderly shutdown. Therefore, such a repair does not change the state of the operating system. In addition, any ongoing business in the virtual machine is not lost. The client operating system does not detect this difference even if the hardware devices of the standby ESXi server host are different from the hardware devices of the original ESXi server host. As a result, the failover of VSphere ha can be completely transparent to the customer, with little risk of any downtime.

(ii) how VSphere ha Works

Ha continuously monitors all ESXi server hosts in the cluster and detects failures. The agents placed on each host constantly issue a "heartbeat" to the other hosts in the cluster, and the termination of the heartbeat will cause the restart process on the other hosts for all affected virtual machines. How VSphere ha Works

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M01/54/2A/wKioL1R63N6zZeT_AAEma5eNreg694.jpg "title=" Qq20141130165723.png "alt=" Wkiol1r63n6zzet_aaema5enreg694.jpg "/>

When creating a vsphere ha cluster, you can choose to use a single host as the preferred host to communicate with the Vcenter server and monitor the status of other hosts, subordinate hosts, and their virtual machines. Different types of host failures may occur, the preferred host must detect and handle the corresponding failure, and the failed host must be distinguished from the host that is in the network partition or is isolated from the network. The preferred host uses the data store heartbeat to determine the type of failure.

If Vsphere HA is enabled for the cluster, all active hosts will participate in the election to select the preferred host for the cluster. Hosts that mount the largest number of data stores have an advantage in the election. There is only one preferred host per cluster and all other hosts are subordinate hosts. If the preferred host fails, shuts down, or is removed from the cluster, a new election is made.

The preferred host monitors the activity of the subordinate hosts in the cluster and is completed by exchanging a network heartbeat every second. When a preferred host stops receiving these heartbeats from a subordinate host, it checks host activity before declaring that the host has failed. First, the host performs an activity check to determine whether the secondary host is exchanging heartbeats between data stores. Also, first the host checks whether the host responds to ICMP pings that are sent to its administrative IP address. If the preferred host cannot communicate directly with the agent on the subordinate host, the subordinate host does not respond to the ICMP ping, and the agent does not issue a heartbeat that is considered to have failed and restarts the host's virtual machine on the standby host. If such a subordinate host exchanges a heartbeat with the datastore, the preferred master assumes that it is in a network partition or isolated network, and therefore continues to monitor the host and its virtual machines.

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M01/54/2C/wKiom1R64QORO5wgAAEzhOchD3s904.jpg "title=" Qq20141130171835.png "alt=" Wkiom1r64qoro5wgaaezhochd3s904.jpg "/>

When a preferred host in a vsphere ha cluster cannot communicate with a subordinate host through the management network, the preferred host uses the data store heartbeat to determine whether the subordinate host is faulted, is in a network partition, or is isolated from the network. If the secondary host has stopped the data store heartbeat, the dependent host is considered to be faulty and its virtual machine has been restarted elsewhere.

(iii) requirements for vSphere ha clusters

Vsphere ha requires that ha be turned on in the cluster, and all hosts in the cluster must have vsphere ha licensing, requiring at least two hosts to configure static IP addresses for all hosts. If you are using DHCP, you must ensure that the address of each host is retained during reboot, or the virtual machine network will be interrupted. To ensure that any virtual machine can run on any host within the cluster, all hosts should have access to the same virtual machine network and data store. Similarly, virtual machines must reside on shared storage instead of local storage, or they will fail to failover if the host fails.

Network redundancy is required for the management network, and the number of vsphere ha heartbeat data stores for the host is 2.

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M02/54/37/wKioL1R8TtnyGnFuAAFMVebpwcY432.jpg "title=" Qq20141201191447.png "alt=" Wkiol1r8ttnygnfuaafmvebpwcy432.jpg "/>

This article is from the "Zhong blog" blog, make sure to keep this source http://capfzgs.blog.51cto.com/7729146/1585188

VMware VSphere 5.0 high availability and dual machine hot standby