WO2014082442A1

WO2014082442A1 - Virtual machine scheduling method and device

Info

Publication number: WO2014082442A1
Application number: PCT/CN2013/077775
Authority: WO
Inventors: 陈岩; 单卫华
Original assignee: 华为技术有限公司
Priority date: 2012-11-28
Filing date: 2013-06-24
Publication date: 2014-06-05
Also published as: CN102937912B; CN102937912A

Abstract

Provided is a virtual machine scheduling method and device. The method comprises the steps of: determining a reuse ratio of a source resource pool, the source resource pool being a heavy-load resource pool of a source physical calculation node, and the reuse ratio being a specific value of the quantity of VCPUs and the quantity of physical kernels; selecting a target resource pool, the target resource pool being a resource pool of a target physical calculation node, having the same reuse ratio as the source resource pool, and having enough set resources; and migrating a virtual machine of the source resource pool to the target resource pool. According to the embodiments of the present invention, the quality of service of the CPU during virtual machine scheduling can be guaranteed.

Description

The present invention claims the priority of a Chinese patent application filed on November 28, 2012 by the Chinese Patent Office, Application No. 201210494726.6, entitled "Virtual Machine Scheduling Method and Apparatus", the entire contents of which are incorporated by reference. Combined in this application.

TECHNICAL FIELD The present invention relates to communications technologies, and in particular, to a virtual machine scheduling method and device.

After the long-term operation of various services in the network, the physical computing nodes in the cloud computing resource pool are unevenly loaded, resulting in reduced service efficiency and waste of resources. The load of each physical computing node needs to be balanced through load balancing scheduling. At present, the most commonly used load balancing technology is load balancing scheduling of the central processing unit (CPU) occupancy of physical computing nodes. In this scheduling mode, the resource management system periodically collects the CPU usage of each physical computing node, and determines whether the physical computing node is overloaded by comparing the CPU usage of each physical computing node with a preset overloaded threshold. Virtual machine scheduling. When scheduling, the system first selects the light load computing node, and determines whether the physical computing node is lightly loaded by comparing the CPU usage of the physical computing node with a preset light load threshold. The system uses the light-loaded physical computing node as the target node and the overloaded physical computing node as the source node. When the scheduling is triggered, the virtual machine is migrated from the source node to the target node. When the CPU usage of the target node reaches a balanced threshold. When the CPU usage of the source node is higher than the reload threshold, the target node is reselected to continue to migrate the virtual machine. The virtual machine migration is stopped after the source node CPU usage is lower than the reload threshold. However, the scheduling scheme only pays attention to the CPU usage of the physical computing node, and cannot guarantee the CPU quality of the virtual machine after scheduling.

SUMMARY OF THE INVENTION In view of this, an embodiment of the present invention provides a virtual machine scheduling method and device, which are used to solve the problem that an existing load balancing scheduling solution cannot guarantee CPU service quality.

In a first aspect, a virtual machine scheduling method is provided, including:

Determining a multiplexing ratio of the source resource pool, where the source resource pool is a reload resource pool on the source physical computing node, where the multiplexing ratio is a ratio of the number of VCPUs to the number of physical cores;

Selecting a target resource pool, where the target resource pool is a resource pool on the target physical computing node, and The reuse ratio of the target resource pool is the same as the reuse ratio of the source resource pool, and the target resource pool has sufficient resources set;

The virtual machine on the source resource pool is migrated to the target resource pool.

In conjunction with the first aspect, in a first possible implementation manner of the first aspect, the target resource pool is configured with sufficient resources, including:

When (X+Y) <=N x ( M+K ), the target resource pool has sufficient resources set, where X is the number of VCPUs of the source resource pool, and Y is the original VCPU of the target resource pool. The number, N is the reuse ratio of the source resource pool and the target resource pool, M is the number of physical cores of the target resource pool, and K is the number of physical cores that the target resource pool can acquire from another resource pool.

With reference to the first aspect, in a second possible implementation manner of the first aspect, the determining a multiplexing ratio of the source resource pool includes:

Performing signaling interaction with the source physical computing node to obtain resource information of the source resource pool, where the resource information includes a quantity of physical cores included in the source resource pool and a virtual machine list included in the source resource pool; The multiplexing ratio is determined according to the number of the VCPUs and the number of the physical cores.

With reference to the first aspect, in a third possible implementation manner of the first aspect, the method further includes: acquiring a CPU usage of a CPU of each virtual machine in each resource pool, and obtaining all the CPU usage according to each virtual machine The sum of the CPU usage of the virtual machine and the number of physical cores that get each resource pool;

Determining, by the ratio of the sum of the CPU usage of the virtual machine and the number of the physical cores, the CPU usage of the resource pool;

Determining, by the source resource pool, a resource pool in which the CPU usage of the resource pool exceeds a set overload threshold;

The migrating the virtual machine on the source resource pool to the target resource pool includes: the sum of the CPU usage of the virtual machine to be migrated and the original CPU usage of the target resource pool is less than or equal to When the heavy load value is described, the virtual machine on the source resource pool is migrated to the target resource pool.

With reference to the first aspect, in a fourth possible implementation manner of the first aspect, the method further includes: interacting with each physical computing node, acquiring resource information of each resource pool, the resource information packet Includes a list of virtual machines and updates the list of virtual machines for each resource pool when the virtual machine is started, created, migrated, stopped, or deleted.

With reference to the first aspect, or any one of the first to fourth possible implementation manners of the first aspect, in the fifth possible implementation manner of the first aspect, the selecting the target resource pool includes:

When the resource pool of the source resource pool has the same multiplexing ratio as the source resource pool and has the set sufficient resources, the resource pool is selected as the target resource pool; or

If there is no resource pool with the same resource resource reuse ratio on the target physical computing node, if there is another resource pool with resource redundancy on the target physical computing node, the The resources in the resource pool of redundant resources allocate the reduced resources to the target resource pool.

In a second aspect, a virtual machine scheduling device is provided, including:

a determining module, configured to determine a multiplexing ratio of the source resource pool, where the source resource pool is a reloading resource pool on the source physical computing node, where the multiplexing ratio is a ratio of the number of VCPUs to the number of physical cores;

a selection module, configured to select a target resource pool, where the target resource pool is a resource pool on the target physical computing node, and the multiplexing ratio of the target resource pool is the same as the multiplexing ratio of the source resource pool, and the The target resource pool has sufficient resources set;

a migration module, configured to migrate a virtual machine on the source resource pool to the target resource pool. With reference to the second aspect, in a first possible implementation manner of the second aspect, the target resource pool selected by the selecting module is configured to have sufficient resources, including:

When (X+Y) <=N x ( Μ+Κ ), the target resource pool has sufficient resources set, where X is the number of VCPUs of the source resource pool, and Υ is the original VCPU of the target resource pool. The quantity, the reuse ratio of the source resource pool and the target resource pool, the number of physical cores of the target resource pool, and the number of physical cores that the target resource pool can acquire from another resource pool.

With reference to the second aspect, in a second possible implementation manner of the second aspect, the determining module is specifically configured to:

With reference to the second aspect, in a third possible implementation manner of the second aspect, the method further includes: a processing module, configured to acquire a central processor CPU usage of each virtual machine in each resource pool, according to each virtual machine The CPU usage is obtained as the sum of the CPU usage of all the virtual machines, and the number of physical cores of each resource pool; and the ratio of the sum of the CPU usage of the virtual machine to the number of the physical cores is determined as the resource. The CPU usage of the pool is determined as the resource pool of the resource pool whose CPU usage exceeds the set overload threshold;

The migrating module is configured to: when the sum of the CPU usage of the virtual machine to be migrated and the original CPU usage of the target resource pool is less than or equal to the heavy load threshold, the source resource pool is used. The virtual machine is migrated to the target resource pool.

With reference to the second aspect, in a fourth possible implementation of the second aspect, the method further includes: an update module, configured to interact with each physical computing node to obtain resource information of each resource pool, where the resource information includes a virtual A list of machines and a list of virtual machines for each resource pool when the virtual machine is started, created, migrated, stopped, or deleted.

In conjunction with the second aspect or the first to fourth possible implementation manners of the second aspect, in a fifth possible implementation manner of the second aspect, the selecting module is specifically configured to:

A resource pool having the same multiplexing ratio as the source resource pool is not present on the target physical computing node, or there is a resource pool having the same multiplexing ratio as the source resource pool, but the resource pool does not have a set. When there are sufficient resources, if there is another resource redundant resource pool on the target physical computing node, the resources in the resource pool of the additional resource redundancy are reduced, and the reduced resources are allocated to the target resource pool. Through the above technical solution, when the virtual machine is scheduled, the resource pool having the same multiplexing ratio is used as the target resource pool of the migration virtual machine. Since the resource pools of the same multiplexing ratio have the same CPU service quality, after the virtual machine is scheduled, Can guarantee the quality of CPU service.

BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly described below. It is obvious that the drawings in the following description are the present invention. For some embodiments, other drawings may be obtained from those skilled in the art without departing from the drawings.

1 is a schematic flowchart of an embodiment of a virtual machine scheduling method according to the present invention;

2 is a schematic diagram of dividing a resource pool by a physical computing node according to an embodiment of the present invention;

3 is a schematic diagram of an update situation of a resource pool in an embodiment of the present invention;

4 is a schematic diagram of another update of a resource pool in an embodiment of the present invention;

FIG. 5 is a schematic diagram of scheduling a virtual machine in an embodiment of the present invention; FIG.

6 is a schematic flowchart of another embodiment of a virtual machine scheduling method according to the present invention;

FIG. 7 is another schematic diagram of scheduling a virtual machine in an embodiment of the present invention;

8 is a schematic structural diagram of an embodiment of a virtual machine scheduling device according to the present invention;

FIG. 9 is a schematic structural diagram of another embodiment of a virtual machine scheduling device according to the present invention.

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. The embodiments are a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

FIG. 1 is a schematic flowchart of a virtual machine scheduling method according to an embodiment of the present invention, including:

11: The scheduling node determines a multiplexing ratio of the source resource pool, where the source resource pool is a reloading resource pool on the source physical computing node, where the multiplexing ratio is a ratio of the number of VCPUs to the number of physical cores;

12: The scheduling node selects a target resource pool, where the target resource pool is a resource pool on the target physical computing node, and the multiplexing ratio of the target resource pool is the same as the multiplexing ratio of the source resource pool, and the target The resource pool has sufficient resources for the virtual machine hosting the source resource pool;

13: The scheduling node migrates the virtual machine on the source resource pool to the target resource pool.

Each physical compute node's CPU has multiple physical cores. The underlying operating system process and kernel binding technology can control the virtual machine's virtual CPU (VCPU) process to run on different physical cores, thereby isolating different virtual machines. Machine conflicts on CPU resource usage.

The CPU of the physical computing node can be divided into multiple resource pools, and different resource pools have different CPU Quality of Service (QoS). Specifically, resources can be divided according to the multiplexing ratio Pool, reuse ratio refers to the ratio of the number of VCPUs in a resource pool to the number of physical cores in the resource pool. For a virtual machine (VM) in the same resource pool, you need to ensure that the virtual machine's CPU service quality is met when all the virtual machine services in the resource pool are the busiest. According to the algorithm of the underlying CPU fair scheduling, the CPU resources that each virtual machine can acquire are related to the number of VCPUs in the resource pool and the number of physical cores. When all virtual machines in the resource pool are busy, the physical CPU resources that each VCPU can acquire are physical cores. Number / VCPU total.

Based on the above principles, the CPU service quality of the virtual machine can be guaranteed by controlling the number of physical cores and the number of VCPUs in the resource pool.

As shown in FIG. 2, resource pools may be divided according to a multiplexing ratio, and each resource pool has a respective multiplexing ratio. For example, physical compute node A can be divided into a high-performance resource pool, a normal resource pool, and a default resource pool. The reuse ratio of the high-performance resource pool is 1:1, the reuse ratio of the common resource pool is 2:1, and the reuse ratio of the default resource pool is 4:1. That is to say, each VCPU in the high-performance resource pool can monopolize one physical core, and each VCPU can obtain 100% computing power when it is the busiest; one VCPU in the common resource pool can co-locate one physical core with another VCPU. Each VCPU can get 50% of computing power when it is the busiest; the VCPU in the default resource pool can run 4 VCPUs on a physical core. When each VCPU is busy, each VCPU can get 25%. The guarantee of computing power.

The scheduling node may collect resource information of a resource pool of each physical computing node, and the resource information includes the number of physical cores in each resource pool and a list of virtual machines. For example, the scheduling node can know that the number of physical cores in the high performance resource pool is 6 through signaling interaction, and the virtual machine list includes VM1-1, VM1-2, VM2-1, VM2-2, VM2-3, and VM2-4. The resource information can be used to determine the reuse ratio of the resource pools so as to migrate to the resource pool of the same multiplexing ratio when overloaded.

In addition, each resource pool can expand or shrink as virtual machines are started, created, migrated, stopped, or deleted.

As shown in Figure 3, when a virtual machine is started, created, or moved in a normal resource pool, the default resource pool can be compressed and the common resource pool can be expanded. For example, add a VM to a normal resource pool. Or,

As shown in Figure 4, when a virtual machine is stopped, deleted, or migrated in a normal resource pool, the normal resource pool can be contracted and the default resource pool can be expanded. For example, one VM is reduced in a normal resource pool.

In this embodiment, when the virtual machine is migrated, not only the CPU usage rate but also the resource pool having the same multiplexing ratio is migrated. For example, referring to Figure 5, there are two physical computing nodes, namely physical computing node A and physical computing node B, and the left part is the VM of the first two physical computing nodes. The right part is the case of the VMs of the two physical compute nodes after migration. Each physical computing node is divided into three resource pools, and the two resource pools in the same plane have the same multiplexing ratio, that is, the resource pools at the top of the two physical computing nodes have the same multiplexing ratio, and the resource pool in the middle With the same multiplexing ratio, the resource pool at the bottom has the same multiplexing ratio.

In load balancing, the virtual machine migrates in a resource pool with the same multiplexing ratio. For example, one VM of the top resource pool of physical computing node A migrates to the top resource pool of physical computing node B, and the central resource of physical computing node A One VM of the pool migrates to the central resource pool of the physical computing node B, and one VM of the bottom resource pool of the physical computing node B migrates to the local resource pool of the physical computing node A. CPU QoS can be guaranteed due to migration within a resource pool of the same multiplexing ratio.

In this embodiment, by migrating a virtual machine to a resource pool having the same multiplexing ratio as the source resource pool, the same CPU service quality can be ensured because the multiplexing ratio is the same. Therefore, the virtual machine scheduling in this embodiment can implement load balancing. Guarantee the quality of CPU service.

FIG. 6 is a schematic flowchart of another embodiment of a virtual machine scheduling method according to the present invention, including:

61: The scheduling node determines whether there is an overloaded resource pool, and if so, executes 62, otherwise executes 67. The scheduling node may obtain information about each resource pool by interacting with the physical computing node, where the information may include the CPU usage of each virtual machine in each resource pool and the number of physical cores included in each resource pool.

Based on the information, you can determine the CPU usage of each resource pool. The calculation formula can be: CPU usage per resource pool = CPU usage of all VMs in the resource pool / Number of physical cores included in the resource pool.

In addition, the scheduling node can set the overloaded threshold. When the CPU usage of a resource pool exceeds the overloaded threshold, the resource pool is the overloaded resource pool.

62: The scheduling node determines the reuse ratio of the reloaded resource pool.

The information exchanged between the foregoing scheduling node and the resource pool may further include virtual machine information, and the number of VCPUs included in one resource pool may be determined according to the virtual machine information, and then the multiplexing ratio of one resource pool may be calculated: the resource pool includes The number of VCPUs / the number of physical cores included in the resource pool.

63: The scheduling node determines whether the virtual machine to be migrated is selected in the reload resource pool, and if so, executes 64, otherwise executes 67.

Among them, some virtual machines may be bound during setup, and migration is not possible. Therefore, the selected candidate is A migrated virtual machine is a virtual machine that can be migrated to another resource pool.

64: The scheduling node determines whether the target resource pool is selected, and if so, executes 65, otherwise executes 67. The target resource pool satisfies the following conditions:

It has the same multiplexing ratio as the overloaded resource pool, and the resources are sufficient.

Sufficient resources means: Suppose the number of VCPUs of the virtual machine to be migrated is X, the number of existing VCPUs in the target resource pool is Y, and the number of physical cores in the target resource pool is M, which can be obtained by squeezing the low-level resource pool. The number of physical cores is K, and the reuse ratio of the target resource pool and the overloaded resource pool is N: l. Then, when (Χ+Υ) <=Ν χ ( Μ+Κ ), the resources are sufficient, otherwise the resources are insufficient. The number of physical cores that can be obtained refers to the number of physical cores of the resource pool that can be squeezed. For example, the default default resource pool exists on each physical compute node, and the remaining resource pools, such as common resource pools and high-performance resources. The pool is squeezed on the basis of the default resource pool, that is, the physical core of the default resource pool is allocated to the rest of the resource pool. The number of physical cores in the squeezable resource pool is then the number of physical cores included in the current default resource pool.

Optionally, the target resource pool may be the original physical computing node, that is, the original resource pool with the same multiplexing ratio as the source resource pool and sufficient resources, or the target resource pool may also be the target physical computing node. Squeezed from the remaining resource pools. For example, the original resource pool on the target resource pool is only the default resource pool, and the source resource pool is the normal resource pool. When there are redundant resources in the default resource pool, you can use the default resource pool. Extrusion of some resources, these part of the resources constitute a common resource pool, used as a target resource pool; or, the source resource pool is a common resource pool, and the resources of the original common resource pool on the target physical compute node are not enough, and the default When the resources of the resource pool are redundant, some resources may be squeezed out from the default resource pool, and the resources are added to the original common resource pool to form a target resource pool together with the original resources.

65: The scheduling node migrates the virtual machine to be migrated to the target resource pool.

Optionally, before the migration, the load of the target resource pool may be determined. After the target resource pool is added to the virtual machine to be migrated, the virtual machine is not overloaded, that is, the CPU of the virtual machine to be migrated. The virtual machine on the source resource pool is migrated to the target resource pool when the sum of the usage and the original CPU usage of the target resource pool is less than or equal to the set heavy load threshold of the target resource pool.

66: The scheduling node determines whether the overloaded resource pool migrated out of the virtual machine is still overloaded. If yes, repeat step 63 and subsequent steps, otherwise execute 67. 67: End.

In this embodiment, for example, referring to FIG. 7, 4, there is a physical computing node A and a physical computing node B, the physical computing node A has a multiplexing ratio of 1:1 resource pool reload, and the physical computing node B is reused. The resource pools of 2:1 and 4:1 are overloaded. After the above process, the multiplexing ratio of the physical computing node A to the virtual pool of the 1:1 resource pool can be migrated to the physical computing node B. :1 resource pool, the physical computing node B is more multiplexed than the 2:1 resource pool virtual machine to the physical computing node A reuse ratio 2:1 resource pool, the physical computing node B multiplexing ratio The virtual machine of the 4:1 resource pool is migrated to the physical computing node A with a reuse ratio of 4:1.

It can be understood that, in this embodiment, the load balancing scheduling between the physical computing node A and the physical computing node B is taken as an example, or the heavy load of the physical computing node A is migrated to the physical computing node B, and the physical computing node B is physically The virtual machine that reuses the resource pool of 2:1 is migrated to the resource pool of the physical computing node C with a multiplexing ratio of 2:1.

In this embodiment, by migrating a virtual machine to a resource pool having the same multiplexing ratio, the service quality of the virtual machine CPU can be ensured, and the user experience can be ensured without being affected, and the server load balancing effect can be achieved at the same time.

FIG. 8 is a schematic structural diagram of an embodiment of a virtual machine scheduling device according to the present invention. The device includes a determining module 81, a selecting module 82, and a migration module 83. The determining module 81 is configured to determine a multiplexing ratio of a source resource pool, and the source resource pool. For the source physical computing node, the reuse ratio is the ratio of the number of VCPUs to the number of physical cores; the selection module 82 is configured to select a target resource pool, where the target resource pool is the target physical computing node. Resource pool, and the reuse ratio of the target resource pool is the same as the reuse ratio of the source resource pool, and the target resource pool has sufficient resources set; the migration module 83 is configured to use the source resource pool The virtual machine is migrated to the target resource pool.

Optionally, the target resource pool selected by the selecting module has sufficient resources to be set to include: when (Χ+Υ) <=Ν χ ( Μ+Κ ), the target resource pool has sufficient resources set, X is the number of VCPUs in the source resource pool, Υ is the number of original VCPUs in the target resource pool, 复用 is the reuse ratio of the source resource pool and the target resource pool, and Μ is the number of physical cores in the target resource pool. The number of physical cores that the target resource pool can acquire from another resource pool.

Optionally, the determining module is specifically configured to:

Performing signaling interaction with the source physical computing node to obtain resource information of the source resource pool, where the resource information includes the number of physical cores included in the source resource pool, and the source resource pool includes a list of virtual machines; determining the multiplexing ratio according to the number of VCPUs and the number of physical cores.

Optionally, the device further includes:

The processing module is configured to obtain the CPU usage of each virtual machine, obtain the sum of the CPU usage of all the virtual machines in each resource pool according to the CPU usage of each virtual machine, and obtain the number of physical cores of each resource pool; The ratio of the sum of the CPU usage of the virtual machine to the number of the physical cores is determined as the CPU usage of the resource pool; and the CPU usage of the resource pool exceeds the set heavy load threshold. The pool is determined to be the source resource pool; the migration module is specifically configured to: the sum of the CPU usage of the virtual machine to be migrated and the original CPU usage of the target resource pool is less than or equal to the heavy load When the value is, the virtual machine on the source resource pool is migrated to the target resource pool.

Optionally, the device further includes:

An update module, configured to interact with each physical computing node, obtain resource information of each resource pool, where the resource information includes a virtual machine list, and update each resource when the virtual machine is started, created, migrated, stopped, or deleted. A list of virtual machines in the pool.

Optionally, the selecting module is specifically configured to:

A resource pool having the same multiplexing ratio as the source resource pool is not present on the target physical computing node, or there is a resource pool having the same multiplexing ratio as the source resource pool, but the resource pool does not have a set. When there are sufficient resources, if there is another resource redundant resource pool on the target physical computing node, the resources in the resource pool of the additional resource redundancy are reduced, and the reduced resources are allocated to the target resource pool.

In addition, as shown in FIG. 9, the embodiment of the present invention further provides a communication device, which may be a virtual machine scheduling server, where the communication device 90 includes a transmitting circuit 902, a receiving circuit 903, a power controller 904, and a processor 905. , memory 906 and antenna 901. The processor 905 controls the operation of the device 90, which may also be referred to as a CPU. Memory 906 can include read only memory and random access memory and provides instructions and data to processor 905. A portion of the memory 906 may also include non-volatile line random access memory (NVRAM). The transmitting circuit 902 and the receiving circuit 903 may be coupled to the antenna 901, and the antenna may be in a wireless form, or may be a wired network cable, an optical fiber, or the like. Each of the devices 90 The components are coupled together by a bus system 9100, wherein the bus system 9100 includes a power bus, a control bus, and a status signal bus in addition to the data bus. However, for clarity of description, various buses are labeled as bus system 9100 in the figure.

The method disclosed in the foregoing embodiments of the present invention may be applied to the processor 905 or implemented by the processor 905. Processor 905 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method may be completed by an integrated logic circuit of hardware in the processor 905 or an instruction in the form of software. For performing the method disclosed in the embodiments of the present invention, the processor 905 may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic. Device, discrete gate or transistor logic and logic block diagram. The general purpose processor may be a microprocessor or the processor or any conventional processor, decoder or the like. The steps of the method disclosed in the embodiments of the present invention may be directly implemented as hardware processor execution completion, or performed by a combination of hardware and software modules in the processor. The software module can be located in a conventional storage medium such as random access memory, flash memory, read only memory, programmable read only memory or electrically erasable programmable memory, registers, and the like. The storage medium is located in memory 906, and processor 905 reads the information in memory 906 and, in conjunction with its hardware, performs the steps of the above method. The message in the above method can be transmitted or received through the transmitting circuit 902 and the receiving circuit 903.

Specifically, the device may include a processor, where the processor is specifically configured to: determine a multiplexing ratio of the source resource pool, where the source resource pool is a reload resource pool on the source physical computing node, and the multiplexing ratio is a VCPU Ratio of the number of physical cores to the number of physical cores; selecting a target resource pool, the target resource pool being a resource pool on the target physical computing node, and multiplexing ratio of the target resource pool to the source resource pool The same, and the target resource pool has sufficient resources set; the virtual machine on the source resource pool is migrated to the target resource pool.

Optionally, the target resource pool has sufficient resources set, including:

When (X+Y) <=N x ( Μ+Κ ), the target resource pool has sufficient resources set, where X is the number of VCPUs of the source resource pool, and Y is the original VCPU of the target resource pool. The number, N is the reuse ratio of the source resource pool and the target resource pool, M is the number of physical cores of the target resource pool, and K is the number of physical cores that the target resource pool can acquire from another resource pool.

Optionally, the processor is specifically configured to: perform signaling interaction with the source physical computing node, and acquire The resource information of the source resource pool, the resource information includes a number of physical cores included in the source resource pool, and a virtual machine list included in the source resource pool; determining, according to the virtual machine list, the source resource pool The number of VCPUs; the multiplexing ratio is determined according to the number of VCPUs and the number of physical cores.

Optionally, the processor is configured to obtain the CPU usage of each virtual machine in each resource pool, obtain the sum of the CPU usage of all the virtual machines according to the CPU usage of each virtual machine, and obtain the physical of each resource pool. The number of cores; the ratio of the sum of the CPU usage of the virtual machine to the number of the physical cores is determined as the CPU usage of the resource pool; the CPU usage of the resource pool exceeds the set overload The resource pool of the source resource pool is determined to be the source resource pool; and the virtual machine on the source resource pool is migrated to the target resource pool when the virtual load reload threshold is to be migrated.

Optionally, the processor is further configured to interact with each physical computing node to obtain resource information of each resource pool, where the resource information includes a virtual machine list, and when the virtual machine is started, created, migrated, stopped, or deleted, Update the list of virtual machines for each resource pool.

Optionally, the processor is specifically configured to: when there is a resource pool with the same multiplexing ratio as the source resource pool and having sufficient resources set, the resource pool is selected as the resource pool. Or a resource pool having the same multiplexing ratio as the source resource pool, or a resource pool having the same multiplexing ratio as the source resource pool but having the same resource pool If the resource pool does not have enough resources to be set, if there is another resource redundancy resource pool on the target physical computing node, the resources in the resource pool of the additional resource redundancy are reduced, and the resource allocation is reduced. Give the target resource pool.

It will be clearly understood by those skilled in the art that for the convenience and brevity of the description, only the division of each functional module described above is exemplified. In practical applications, the above function assignment can be completed by different functional modules as needed. The internal structure of the device is divided into different functional modules to perform all or part of the functions described above. For the specific working process of the system, the device and the unit described above, reference may be made to the corresponding process in the foregoing method embodiments, and details are not described herein again.

In the several embodiments provided by the present application, it should be understood that the disclosed system, device and square The law can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be used. Combined or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form. The components displayed by the unit may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. The instructions include a plurality of instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor to perform all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like. The medium of the code.

The above embodiments are only used to illustrate the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that The technical solutions described in the embodiments are modified, or some of the technical features are equivalently replaced; and the modifications or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

Rights request

1. A virtual machine scheduling method, characterized by including:

Determine the reuse ratio of the source resource pool, which is the overloaded resource pool on the source physical computing node, and the reuse ratio is the ratio of the number of virtual central processors VCPU to the number of physical cores; select the target resource Pool, the target resource pool is a resource pool on the target physical computing node, and the reuse ratio of the target resource pool is the same as the reuse ratio of the source resource pool, and the target resource pool has sufficient set resource;

Migrate virtual machines on the source resource pool to the target resource pool.

2. The method according to claim 1, characterized in that the target resource pool has set sufficient resources, including:

When (X+Y) <=N x (M+K), the target resource pool has sufficient set resources, where X is the number of VCPUs in the source resource pool, and Y is the number of original VCPUs in the target resource pool. Quantity, N is the reuse ratio of the source resource pool and the target resource pool, M is the number of physical cores in the target resource pool, and K is the number of physical cores that the target resource pool can obtain from another resource pool.

3. The method according to claim 1, characterized in that determining the reuse ratio of the source resource pool includes:

Perform signaling interaction with the source physical computing node to obtain resource information of the source resource pool, where the resource information includes the number of physical cores included in the source resource pool and a list of virtual machines included in the source resource pool; The multiplexing ratio is determined according to the number of VCPUs and the number of physical cores.

4. The method according to claim 1, further comprising:

Obtain the central processor CPU usage of each virtual machine in each resource pool, obtain the sum of the CPU usage of all virtual machines based on the CPU usage of each virtual machine, and obtain the number of physical cores in each resource pool;

The ratio of the sum of the CPU occupancy rates of the virtual machines to the number of physical cores is determined as the CPU occupancy rate of the resource pool;

Determine the resource pool whose CPU usage exceeds the set reload threshold as the source resource pool; Migrating the virtual machine on the source resource pool to the target resource pool includes: when the sum of the CPU occupancy rate of the virtual machine to be migrated and the original CPU occupancy rate of the target resource pool is less than or equal to When the reload threshold is reached, migrate the virtual machine on the source resource pool to the target resource pool.

5. The method according to claim 1, further comprising:

Interact with each physical computing node to obtain resource information of each resource pool. The resource information includes a virtual machine list, and update the virtual machine list of each resource pool when a virtual machine is started, created, migrated, stopped, or deleted. .

6. The method according to any one of claims 1-5, characterized in that the selection of the target resource pool includes:

When a resource pool with the same reuse ratio as the source resource pool and sufficient resources exists on the target physical computing node, select the resource pool as the target resource pool; or,

When there is no resource pool with the same reuse ratio as the source resource pool on the target physical computing node and sufficient resources are available, if there is another resource pool with redundant resources on the target physical computing node, then the For other resources in the redundant resource pool, the reduced resources are allocated to the target resource pool.

7. A virtual machine scheduling device, characterized by including:

Determining module, used to determine the reuse ratio of the source resource pool, the source resource pool is the overloaded resource pool on the source physical computing node, the reuse ratio is the number of virtual central processors VCPU and the number of physical cores ratio;

The selection module is used to select a target resource pool, the target resource pool is a resource pool on the target physical computing node, and the reuse ratio of the target resource pool is the same as the reuse ratio of the source resource pool, and the The target resource pool has sufficient resources;

A migration module, configured to migrate virtual machines on the source resource pool to the target resource pool.

8. The device according to claim 7, wherein the target resource pool selected by the selection module has set sufficient resources including:

When (X+Y) <=N x (M+K), the target resource pool has sufficient set resources, where X is the number of VCPUs in the source resource pool, and Y is the number of original VCPUs in the target resource pool. quantity, N is the reuse ratio of the source resource pool and the target resource pool, M is the number of physical cores in the target resource pool, and K is the target resource pool. The number of physical cores that the target resource pool can obtain from another resource pool.

9. The device according to claim 7, wherein the determination module is specifically configured to: perform signaling interaction with the source physical computing node, and obtain resource information of the source resource pool, where the resource information includes The number of physical cores included in the source resource pool and the list of virtual machines included in the source resource pool; The multiplexing ratio is determined based on the number of VCPUs and the number of physical cores.

10. The device according to claim 7, further comprising:

The processing module is used to obtain the central processor CPU usage of each virtual machine in each resource pool, obtain the sum of the CPU usage of all virtual machines based on the CPU usage of each virtual machine, and obtain the physical core of each resource pool. number; determine the ratio of the sum of the CPU occupancy rates of the virtual machines to the number of physical cores as the CPU occupancy rate of the resource pool; set the CPU occupancy rate of the resource pool to exceed the set overload width. The resource pool of the value is determined to be the source resource pool;

The migration module is specifically configured to: when the sum of the CPU occupancy rate of the virtual machine to be migrated and the original CPU occupancy rate of the target resource pool is less than or equal to the reload threshold, transfer the source resource pool to the source resource pool. Migrate the virtual machines to the target resource pool.

11. The device according to claim 7, further comprising:

Update module, used to interact with each physical computing node, obtain resource information of each resource pool, the resource information includes a virtual machine list, and update each resource when the virtual machine is started, created, migrated, stopped or deleted List of virtual machines for the pool.

12. The device according to any one of claims 7-11, characterized in that the selection module is specifically used for:

There is no resource pool with the same reuse ratio as the source resource pool on the target physical computing node,

When resources are sufficient, if there is another resource pool with redundant resources on the target physical computing node, the resources in the other resource pool with redundant resources are reduced, and the reduced resources are allocated to the target resource pool.