CN112527450B

CN112527450B - Super-fusion self-adaptive method, terminal and system based on different resources

Info

Publication number: CN112527450B
Application number: CN202011332356.7A
Authority: CN
Inventors: 张辉; 王猛; 吴瑞; 刘春�
Original assignee: Shandong Qianyun Qichuang Information Technology Co ltd; Shandong Trusted Cloud Information Technology Research Institute; Zhongan Trustworthy Qingdao Network Technology Co ltd
Current assignee: Shandong Qianyun Qichuang Information Technology Co ltd; Shandong Trusted Cloud Information Technology Research Institute; Zhongan Trustworthy Qingdao Network Technology Co ltd
Priority date: 2020-11-24
Filing date: 2020-11-24
Publication date: 2023-05-26
Anticipated expiration: 2040-11-24
Also published as: CN112527450A

Abstract

The disclosure provides a super-fusion self-adaptive method, a terminal and a system based on different resources, comprising the following steps: acquiring host parameters to be super-fused, and adjusting the resources of the host; after adjustment, the super fusion system is divided into different network small clusters, storage small clusters and calculation small clusters; creating a management platform virtual machine on a small cluster set with network, storage and computing performances meeting set conditions; and configuring the virtual machines so that the virtual machines migrate in the same network small cluster and the computing small cluster when migrating, and providing different hardware resources for the virtual machines with different requirements. The use of hardware resources by the virtual machine is optimized, and hardware with different resources is provided for virtual machines with different requirements. Meanwhile, the compatibility with the traditional super fusion system can be guaranteed to the greatest extent.

Description

Super-fusion self-adaptive method, terminal and system based on different resources

Technical Field

The disclosure belongs to the technical field of super fusion, and particularly relates to a super fusion self-adaptive method, a terminal and a system based on different resources.

Background

The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.

A conventional super fusion system consists of several hosts. Hosts are connected together by a network and build up a cluster. The disk on the host computer constructs a distributed storage system through the network. A corresponding virtual machine is created on the host and a distributed storage system is used. The whole system can be expanded continuously according to the requirements of users or contracted appropriately.

The inventor found in the research that in the actual use scenario, the user would prefer to use the existing device in order to facilitate old, rather than purchasing the same host and hard disk every time the system is created or expanded, to build a new network environment. However, the use of the existing devices will lead to a super-converged network environment with complex and variable environments, various hosts for users and complicated disk types. Meanwhile, as the super fusion cluster often has a barrel principle, once a certain performance is low, the performance of the whole cluster is seriously reduced. Resource homogenization is often required, which is contrary to the complex and extensive hardware resources in the old scenario.

Disclosure of Invention

In order to overcome the defects of the prior art, the present disclosure provides a super fusion adaptive method based on different resources, which realizes that the existing equipment is compatible with a common super fusion system.

To achieve the above object, one or more embodiments of the present disclosure provide the following technical solutions:

in a first aspect, a super-fusion adaptive method based on different resources is disclosed, comprising:

acquiring host parameters to be super-fused, and adjusting the resources of the host;

after adjustment, the super fusion system is divided into different network small clusters, storage small clusters and calculation small clusters;

creating a management platform virtual machine on a small cluster set with network, storage and computing performances meeting set conditions;

and configuring the virtual machines so that the virtual machines migrate in the same network small cluster and the computing small cluster when migrating, and providing different hardware resources for the virtual machines with different requirements.

According to the technical scheme, when the host parameters to be super-fused are obtained, aiming at each host for installing the operating systems, the super-fused installation management interface is arranged in each operating system, and the parameters of the super-fused host are transmitted to the installation interface through the installation process.

According to the further technical scheme, whether the host needs to be added into the super fusion system or not is determined through interaction between the management interface and the installation process.

Further technical solution, the parameters of the super fusion host include: the network connection status between the hosts, the storage on each host, the cpu core number on each host and the memory number on each host are determined based on the parameters, and the homogeneity between the hosts is determined based on the parameters, so that the super-fusion small clusters are formed by the parameters which are similar.

According to a further technical scheme, after judging the network connection condition among the hosts, sorting is carried out according to the network speed, and the hosts with different network speeds are divided into different small clusters.

Preferably, the network card is replaced to adjust the network for certain host network speeds below the set point.

Preferably, the storage type and the storage size of each host disk are adjusted, so that the storage space, cpu performance and memory space of each host are kept consistent.

According to a further technical scheme, when the super fusion system is divided into different network sub-clusters, the super fusion system is divided according to the network speed of the network, wherein storage sub-clusters on the different network sub-clusters are different in computing sub-clusters.

According to the technical scheme, when the virtual machine is created, if the virtual machine is computationally intensive, the small cluster with good computation performance is preferentially used, and if the virtual machine is IO intensive, the small cluster with good storage performance is preferentially used.

Further technical scheme still includes: and expanding the super-converged host, wherein the resources on the expanded host are distributed into small clusters with similar performances, and if the performance difference between the expanded host and the original cluster resources exceeds a set value, the corresponding small clusters are independently established.

According to the technical scheme, network small clusters in the super-fusion clusters formed by all the hosts are finally fused to form a network with the same network speed, and due to the fusion of the network small clusters, storage small clusters on different network small clusters are fused, and the computing small clusters are fused.

According to the technical scheme, when the virtual machines are fused, the resources with the same performance are marked with the same labels and are transmitted to the management station, so that the virtual machines are selected when the virtual machines are built.

In a second aspect, a super-fusion adaptive terminal based on different resources is disclosed, comprising:

a processor and a storage medium;

the storage medium is used for storing a program executed by the processor and parameters of the super fusion host; the processor is used for executing the steps of the super fusion self-adaption method based on different resources.

In a third aspect, a super-fusion adaptive system based on different resources is disclosed, comprising:

and the host computers are connected together through a network and form a cluster, and the different host computers are subjected to super fusion by using the super fusion self-adaptive method based on different resources.

The one or more of the above technical solutions have the following beneficial effects:

according to the technical scheme, aiming at the super fusion system built by the product utilized old resources, super fusion is carried out on the basis of the existing host, targeted optimization and processing are carried out, so that the use of the hardware resources by the virtual machine is further optimized while the traditional super fusion characteristics of the product are maintained, and hardware with different resources is provided for the virtual machines with different requirements. Meanwhile, the compatibility with the traditional super fusion system can be guaranteed to the greatest extent.

The resource clusters in the technical scheme of the disclosure are further divided into small clusters according to the resource details, so that the resource clusters are multi-layered, and the resources are subdivided so as to be convenient to manage and use.

Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate and explain the exemplary embodiments of the disclosure and together with the description serve to explain the disclosure, and do not constitute an undue limitation on the disclosure.

FIG. 1 is a schematic diagram of a conventional super fusion system according to an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a multi-level super fusion system according to an embodiment of the present disclosure;

FIG. 3 is a schematic diagram of the installation of an adaptive super fusion system according to an embodiment of the present disclosure;

fig. 4 is a schematic diagram illustrating interaction between a management interface and an installation process according to an embodiment of the disclosure.

Detailed Description

It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments in accordance with the present disclosure. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.

Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.

Example 1

A conventional super fusion system is shown in fig. 1: a conventional super fusion system consists of several hosts. Hosts are connected together by a network and build up a cluster. The disk on the host computer constructs a distributed storage system through the network. A corresponding virtual machine is created on the host and a distributed storage system is used. The whole system can be continuously expanded along with the demands of users. Or to a suitable reduction in volume.

However, in an old super-converged system, the hardware resources of the user are relatively large, which can have serious influence on the performance of the system.

Based on the application requirements, the embodiment discloses a super-fusion self-adaptive method based on different resources, which forms a multi-level super-fusion system to more rationalize the utilization of hardware resources, and is shown in fig. 2.

Specifically, the method is to further subdivide the resources. Here, the resources are mainly divided into 3 parts, network resources, storage resources (hard disk used to construct distributed storage, etc.), and computing resources (refer to cpu and memory of a host, and for a computationally intensive virtual machine, such resources are important).

The scheme increases the concept of a small cluster, and comprises a network small cluster, a storage small cluster and a calculation small cluster aiming at different resources.

Specifically, the network small cluster: in some cases, there may be situations where network speeds in the cluster are inconsistent. Typically caused by a transitional phase of network modification. However, since network resources are critical to the clusters, the clusters need to be divided into different small clusters according to network speeds under different network environments. Such as the 1G network small cluster, and the 100M network small cluster, the latter is more suitable for backup and not suitable for a large number of io operations or computing operations. If the resources on 2 network clusters are mixed, the actual effect is the same as for a 100M cluster. The storage and computing clusters are further partitioned based on the same network cluster.

Distributed storage small clusters: distributed small clusters are generally divided into 3 types:

all-solid-state disks constitute a small cluster of distributed storage.

The mechanical disk and the solid-state disk are mixed to form a small storage cluster, wherein the solid-state disk needs to be used as a buffer disk of the mechanical disk.

The mechanical disks form a distributed small cluster. Smaller clusters may be further divided in the mechanical disc according to rotational speed, if desired.

The above is only a common scene example, and different small clusters can be divided according to different specific situations of storage performance. In this way, small clusters of solid state disks may be preferentially used when building IO-intensive virtual machines. While large-scale storage, when IO usage is less dense, a cluster of mechanical disks may be used.

Computing a small cluster: the CPU and memory resources on different hosts are different, and the computing resources are divided into different small clusters according to the performance of the CPU and the size of the memory. For virtual machines with relatively high computational resource requirements, small clusters with high CPU performance are prioritized.

It should be noted that the above-mentioned further division of all resources, i.e. the generation of small clusters, is only logical, with more emphasis on optimizing the performance than on isolation. In other words, when high-performance small clusters are starved of resources, the virtual machine may migrate to other poor-performance clusters as well. Small clusters with slightly poorer storage performance may also be used when high performance storage small cluster resources are scarce.

The technical scheme of the present disclosure is based on a super-fusion self-adaptive method of different resources, comprising:

According to the technical scheme, a virtual management layer of the resource is added, management and use of the resource are refined, non-homogeneous complex resources are managed, and the utilization of the resources is facilitated.

The self-adaptive super fusion system is installed by the following specific steps:

as shown in fig. 3, a host is selected and an operating system is installed. Each operating system is provided with a super-fusion installation management interface. After the installation, each host is operated with a management interface and an independent installation process for executing the command of the management interface.

Any host is selected, and the management interface can be opened. After the management interface is opened. Storage ips are set for each host according to the unified arrangement of the administrator.

In the technical scheme, all hosts have the same management program, and any one of the hosts can be started.

As shown in FIG. 4, the management interface automatically discovers all the super-fusion hosts that need to be installed. The discovery process is accomplished by interacting with an installation process on each machine.

And performing intelligent analysis on the super fusion host computer in the user interface. The parameters of the super fusion host are transmitted to the installation interface through the installation process, and the parameters are mainly carried out from 4 aspects:

the network connection status between hosts, the main parameters include: bandwidth between the individual hosts.

Storage on each host. The main parameters include:

storage type: mechanical storage, solid state storage, hybrid storage.

Storage size: the number of disks and the storage capacity of the disks.

The number and type of cpu cores on each host.

The amount of memory on each host.

The homogeneity of the super-fusion host is mainly examined here. Namely, the similarity among the super fusion hosts is ensured as much as possible. Hosts with the same bandwidth, hosts with the same storage type, hosts with smaller memory difference, and cpus are put together as much as possible to be used as super-fusion small clusters.

According to the result of the intelligent analysis, the performance condition of each resource option is given.

Network: the ordering will be according to the network speed. Hosts of different network speeds (100 m,1g,10g, etc.) will be divided into different small clusters. If the network speed of some machines is found to be remarkably low, the network card is replaced as much as possible, and the network is adjusted.

And (3) storing: the installation interface gives the storage type, storage size, and adjustment advice for each host disk. Storage is generally classified into 3 categories, one category being all-mechanical storage. One type is all solid state storage. The last category is hybrid storage, which uses a solid-state disk as a cache and a mechanical disk to store data, thereby ensuring the storage performance of the disk. Of course, the user may manually change the hybrid storage to full mechanical storage (while wasting the performance of solid state disks, allowing the user to modify). The user needs to adjust the hard disk condition of each machine, so that the storage space of each machine is kept consistent as much as possible.

Cpu and memory: the cpu performance and the memory space of each machine are kept as consistent as possible.

And the user adjusts the resources of the host according to the intelligent analysis result and the own requirements. After the adjustment is completed, intelligent analysis is performed again. Up to a state where a user deems reasonable.

After the user finishes the adjustment, the super fusion system is built. The construction of the super fusion system is the same as that of the conventional super fusion construction. When only one network small cluster exists in the system, the small clusters are stored and calculated, the scheme is the same as the traditional super fusion.

In the operation process, different small clusters are in parallel relation, and no cross exists. But at the management level allows users to fuse different cells.

After the steps are finished, different network small clusters are divided into the super fusion system according to the network speed of the network. The storage clusters on different network clusters, the computation clusters are different. This is because the network has a decisive influence on the cluster performance.

The storage small clusters are divided according to the performance of the magnetic disk. The storage performance can be generally divided into a class A small cluster (all-solid-state disk), a class B small cluster (mechanical disk with cache) and a class C small cluster (all-mechanical disk). The partitioning of the small clusters is manually adjustable and manually forced identified.

Different small clusters are divided according to the performance of the CPU of the host and the size of the memory. Generally, the performance of a CPU depends on the main frequency and IPC (how many instructions are executed per clock cycle of the CPU). And the memory is mainly referred to as the memory size.

Hosts with similar computing performance are divided into different small clusters. This ensures that computationally intensive hosts prefer small clusters that are computationally efficient.

And creating a management platform virtual machine on the network, storing and the small cluster set with the best computing performance. And transmitting the small cluster division information to the management station. The management platform virtual machine distributes the good resources as much as possible.

There is a management station virtual machine (sometimes called a host-engine) in the super fusion system, and this virtual machine manages the resources of the whole super fusion system. This virtual machine is automatically created by commands on the host.

After the management platform virtual machine is created and the information of the small clusters is saved, the virtual machine can be created by using the information. When virtual machines are created, if computationally intensive, the small clusters with good computational performance are preferentially used. If the memory is IO intensive, a small cluster with good memory performance is preferentially used. The default virtual machine typically uses hybrid storage, thereby compromising performance and cost.

And when the virtual machine is migrated, the virtual machine is migrated in the same network small cluster and the computing small cluster. The virtual machine may be configured to decide whether it can migrate to other small clusters when the current small cluster resources are strained.

Virtual machine migration is a migration in the general sense of cloud computing virtualization. For example, the host where the virtual machine is located is dead or abnormal, and the virtual machine can be migrated to other hosts.

When a virtual machine creates a template and creates multiple virtual machines from the template, it is possible to select which small clusters to use, respectively.

Creating templates is the cloud computing virtualization section creating many virtual machines with the same template.

Maintaining the super fusion cluster:

the super fusion host expands similarly to the creation process. Except that the management station is not created anymore. After the host expansion is finished, the resources on the expanded host are also distributed into small clusters with similar performances. If the performance difference between the extended host and the original cluster resource is larger. The corresponding small cluster is created separately.

Typically, there are no multiple small clusters of networks among the clusters. If present, is due in large part to network complexity. In general, with the upgrading of the network, the network small clusters in the super-convergence cluster are converged finally, and are generally converged into a network with the same network speed. At this time, due to the fusion of the network small clusters, the storage small clusters on different network small clusters are fused with the computing small clusters. The fusion is essentially that resources with the same performance are labeled with the same label and transmitted to the management platform, so that a user can select when building a virtual machine. Virtual machines that have been created on small clusters are not affected.

When the difference between the network small clusters and the computing small clusters is larger and larger, for example, networks with different network speeds appear in the network, the newly added hosts cause larger difference of computing resources, and at the moment, the network small clusters and the computing small clusters can be further divided.

The scheme disclosed by the disclosure is a system which is convenient to use and compatible with common super fusion. A compromise between utilization and performance is achieved, and system updating hardware upgrade is not affected.

Example two

The purpose of this embodiment is to provide a super-fusion adaptive terminal based on different resources, including:

a processor and a storage medium;

Example III

It is an object of the present embodiment to provide a super-fusion adaptive system based on different resources, comprising:

The steps involved in the apparatus of the above embodiment correspond to those of the first embodiment of the method, and the detailed description of the embodiment refers to the relevant description of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present disclosure.

It will be appreciated by those skilled in the art that the modules or steps of the disclosure described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, so that they may be stored in storage means and executed by computing means, or they may be fabricated separately as individual integrated circuit modules, or a plurality of modules or steps in them may be fabricated as a single integrated circuit module. The present disclosure is not limited to any specific combination of hardware and software.

The foregoing description of the preferred embodiments of the present disclosure is provided only and not intended to limit the disclosure so that various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

While the specific embodiments of the present disclosure have been described above with reference to the drawings, it should be understood that the present disclosure is not limited to the embodiments, and that various modifications and changes can be made by one skilled in the art without inventive effort on the basis of the technical solutions of the present disclosure while remaining within the scope of the present disclosure.

Claims

1. The super fusion self-adaptive method based on different resources is characterized by comprising the following steps:

creating a management platform virtual machine on a small cluster set with network, storage and computing performances meeting set conditions; configuring the virtual machine so as to enable the virtual machine to migrate in the same network small cluster and the computing small cluster when migrating, and providing different hardware resources for the virtual machines with different requirements;

automatically creating a management platform virtual machine in the super fusion system on a small cluster set with the best computing performance in a network, wherein the management platform virtual machine is used for managing resources of the whole super fusion system and sending small cluster division information to the management platform; creating a virtual machine by using the information after the management station virtual machine is created and the information of the small clusters is stored; and carrying out intelligent analysis on the super-fusion host computer in the user interface, transmitting parameters of the super-fusion host computer to the installation interface through the installation process, giving out performance conditions of each resource option according to the intelligent analysis result, and adjusting the resources of the host computer by a user according to the intelligent analysis result and own requirements, and carrying out intelligent analysis again after the adjustment is finished until a user considers a reasonable state.

2. The method of claim 1, wherein when obtaining the host parameters to be super-fused, for each host installed with an operating system, the host is provided with a super-fusion installation management interface in each operating system, and the parameters of the super-fusion host are transmitted to the installation interface through an installation process.

3. The method of claim 2, wherein the determination of whether the host needs to join the super-fusion system is performed by interaction between the management interface and the installation process.

4. The method of claim 1, wherein the parameters of the super-fusion host include: the network connection state between the hosts, the storage on each host, the cpu core number on each host and the memory number on each host, and determining the homogeneity between the hosts based on the parameters, and forming the super-fusion small cluster by the parameters similar to each other;

after judging the network connection condition among the hosts, sorting according to the network speed, and dividing the hosts with different network speeds into different small clusters;

when the network speed of some hosts is lower than a set value, the network card is replaced to adjust the network;

and adjusting the storage type and the storage size of each host disk to ensure that the storage space, the cpu performance and the memory space of each host are consistent.

5. The super-fusion adaptive method based on different resources according to claim 1, wherein when the super-fusion system is divided into different network sub-clusters, the super-fusion system is divided according to the network speed of the network, wherein the storage sub-clusters on the different network sub-clusters are different, and the computing sub-clusters are different.

6. The super-fusion adaptive method based on different resources according to claim 1, wherein when the virtual machine is created, if it is computationally intensive, the small cluster with good computation performance is preferentially used, and if it is IO intensive, the small cluster with good storage performance is preferentially used.

7. The super-fusion adaptive method based on different resources according to claim 1, further comprising: and expanding the super-converged host, wherein the resources on the expanded host are distributed into small clusters with similar performances, and if the performance difference between the expanded host and the original cluster resources exceeds a set value, the corresponding small clusters are independently established.

8. The super-fusion self-adaptive method based on different resources according to claim 1, wherein network small clusters in the super-fusion clusters formed by all hosts are finally fused to form a network with the same network speed, and the storage small clusters on different network small clusters are fused due to the fusion of the network small clusters, and the calculation small clusters are fused; when the virtual machine is built, the resources with the same performance are marked with the same labels and transmitted to the management platform, so that the virtual machine is selected.

9. The super-fusion self-adaptive terminal based on different resources is characterized by comprising:

a processor and a storage medium;

the storage medium is used for storing a program executed by the processor and parameters of the super fusion host; the processor is configured to perform the steps of the different resource based super fusion adaptation method according to any of the preceding claims 1-8.

10. The super-fusion self-adaptive system based on different resources is characterized by comprising the following components:

several different host computers connected together by a network and forming a cluster, the different host computers being super-fused by means of the super-fusion adaptive method based on different resources as claimed in any one of the preceding claims 1-8.