CN101814121A - Zero downtime HA of predictive virtual machine - Google Patents

Abstract

The present invention provides a predictive virtual machine zero-downtime HA, which is based on regular detection of computer hardware, and then evaluates the health status of each hardware part. When the evaluation result shows that the computer is in a sub-healthy state, the virtual machine is actively implemented for smooth transfer. In order to achieve the purpose of uninterrupted business running on the virtual machine; the hardware includes the control computer, the main computer, the backup computer and the storage, and the control computer is connected with the main computer, the backup computer and the storage through a network cable to form a local area network. The virtual computer is also provided in the control computer system: a hardware state detection unit, an evaluation control unit, and a virtual service transfer unit.

Description

A predictive virtual machine zero downtime HA

technical field

The invention relates to computer virtualization technology, in particular to a virtual machine high-availability protection method or predictive virtual machine zero-downtime HA.

Background technique

l. Virtualization technology is adopted by more and more manufacturers. Virtualization can dramatically increase the efficiency and availability of resources and applications in an organization. Under the old "one application per computer" model, internal resources were underutilized and IT administrators spent too much time managing computers rather than innovating. With the help of virtualization technology, customers can maximize the use of physical resources.

2. Whether the virtualization technology has a certain degree of redundancy is an important inspection point for customers to choose a virtualization product. At present, most virtualization developers have their own set of redundancy mechanisms, which we call HA.

3. The so-called HA technology refers to a cluster environment composed of multiple physical nodes. When a node goes down, the virtual business running on this node will be transferred to another node. Under this mechanism, customers can continue to use virtual services.

4. However, there is a big defect in the traditional HA technology. When a node in the cluster goes down, the virtual machine is not transferred immediately. Since the data of the virtual machine is stored in the shared storage, the management side can restart the virtual machine on other computers to achieve the HA function. The restart will inevitably lead to a complete interruption of the virtual service. And this also leads to delays when customers use virtual services.

So, here we propose a new solution to prevent this from happening, we call it: a predictable virtual machine zero downtime HA.

Contents of the invention

The purpose of the present invention is to provide a predictive virtual machine zero downtime HA.

The purpose of the present invention is achieved in the following manner, using hardware detection technology to detect the underlying hardware of the server, the detection includes CPU temperature, fan speed, hard disk capacity, hard disk read and write speed, each item is checked, and a comprehensive Value, when the computer is in a sub-health state, actively migrate the virtual business, so that the virtual machine can actively implement the smooth migration of the virtual machine before the computer goes down, so as to achieve the purpose of uninterrupted business running on the virtual machine;

The hardware includes a control computer, a main computer, a backup computer and a memory. The control computer is connected with the main computer, the backup computer and the memory through a network cable to form a local area network. A virtual computer monitored by the control computer is set in the local area network. In the control computer system, there are also: A hardware status detection unit, an evaluation control unit, and a virtual service transfer unit, wherein:

The hardware status detection unit is responsible for regularly detecting the hardware health status of the main computer and the standby computer;

The evaluation control unit judges the comprehensive health status of the main computer and the backup computer hardware. When a computer is running in a sub-healthy state, it triggers the smooth transfer of the virtual machine to another computer to continue running, so as to ensure the continuity of customer business ;

The virtual service transfer implementation unit is responsible for handling virtual machine transfer implementation.

The beneficial effect of the present invention is that the continuous use of virtual services by customers is guaranteed. It has a good effect on customer work efficiency and economic benefits.

Detailed ways

A predictive virtual machine zero downtime HA of the present invention uses hardware detection technology to detect the underlying hardware of the server, including CPU temperature, fan speed, hard disk capacity, and hard disk read/write speed. Each item is checked and a composite value is given. If the computer is in a sub-healthy state, at this time, actively migrate the virtual service, so that the virtual machine can be transferred before the computer goes down.

The hardware includes a control computer, a main computer, a backup computer and a memory. The control computer is connected with the main computer, the backup computer and the memory through a network cable to form a local area network. A virtual computer monitored by the control computer is set in the local area network. In the control computer system, there are also: A hardware status detection unit, an evaluation control unit, and a virtual service transfer unit, wherein:

The hardware status detection unit is responsible for regularly detecting the hardware health status of the main computer and the standby computer;

The evaluation control unit judges the comprehensive health status of the main computer and the backup computer hardware. When a computer is running in a sub-healthy state, it triggers the smooth transfer of the virtual machine to another computer to continue running, so as to ensure the continuity of customer business ;

The virtual service transfer implementation unit is responsible for handling virtual machine transfer implementation.

Claims (1)

1. A predictive virtual machine zero downtime HA, characterized in that the hardware detection technology is used to detect the underlying hardware of the server, including CPU temperature, fan speed, hard disk capacity, hard disk read and write speed, and check each item , and give a comprehensive value, when the computer is in a sub-health state, actively migrate the virtual business, so that the virtual machine can actively implement the smooth migration of the virtual machine before the computer goes down, so as to achieve the purpose of uninterrupted business running on the virtual machine; The hardware includes a control computer, a main computer, a backup computer and a memory. The control computer is connected with the main computer, the backup computer and the memory through a network cable to form a local area network. A virtual computer monitored by the control computer is set in the local area network. In the control computer system, there are also: A hardware status detection unit, an evaluation control unit, and a virtual service transfer unit, wherein: The hardware status detection unit is responsible for regularly detecting the hardware health status of the main computer and the standby computer; The evaluation control unit judges the comprehensive health status of the main computer and the backup computer hardware. When a computer is running in a sub-healthy state, it triggers the smooth transfer of the virtual machine to another computer to continue running, so as to ensure the continuity of customer business ; The virtual service transfer implementation unit is responsible for handling virtual machine transfer implementation.