CN116401060A

CN116401060A - Resource isolation method and device under super fusion, electronic equipment and storage medium

Info

Publication number: CN116401060A
Application number: CN202310387679.3A
Authority: CN
Inventors: 请求不公布姓名
Original assignee: Anchao Cloud Software Co Ltd
Current assignee: Anchao Cloud Software Co Ltd
Priority date: 2023-04-12
Filing date: 2023-04-12
Publication date: 2023-07-07

Abstract

The invention discloses a resource isolation method under super fusion, and belongs to the technical field of super fusion. The method for isolating the resources under the super fusion comprises the steps of judging whether the resources of the super fusion service group are sufficient and whether the resources of the virtual machine group are sufficient, configuring the resources of the virtual machine group to be maximum through a scheduling service of an operating system when the resources of the super fusion service group are insufficient and the resources of the virtual machine group are insufficient, killing the virtual machine which consumes the most memory through a memory overflow mechanism, and further distributing the preparation resources to the super fusion service through the scheduling service of the operating system. The invention is suitable for the physical machine with the super-fusion deployed, which can reasonably allocate resources under the condition of limited CPU and memory resources, and ensure that the super-fusion service runs with optimal performance and stability.

Description

Resource isolation method and device under super fusion, electronic equipment and storage medium

Technical Field

The present invention relates to the field of super fusion technologies, and in particular, to a method for isolating resources under super fusion, and an apparatus, an electronic device, and a storage medium for implementing the method for isolating resources under super fusion.

Background

Super-fusion refers to the fusion of virtualized computing and distributed storage. Compared with a longitudinal expansion (scaling) mode of adding computing resources or storage resources, the super-fusion architecture realizes transverse expansion by adding nodes, and all the nodes are aggregated through a network to form a uniform resource pool. The super fusion comprises a super fusion Cloud Management Platform (CMP), a Software Defined Storage (SDS), a Software Defined Computation (SDC), a Software Defined Network (SDN) and a software defined security Service (SEC). These services unify the computing, storage and network resources to pooled resources by which users create virtual machines to provide services to users. However, the virtual machine created by the user cannot be used without limitation, as shown in fig. 1, if the 48 cores or 256G memory of the whole physical machine are used by the virtual machine, the management service on the super fusion cannot fight for resources at this time, so that the virtual machine is completely crashed and cannot be used.

In order to avoid the above problem, the method adopted at present is that a part of resources are used by a virtual machine, and the rest of resources are reserved for the super fusion service, as shown in fig. 1, 3 cores and 32G memory are reserved in each CPU for the super fusion service. However, in this method, since the resources reserved for the super fusion server are not in one NUMA node, the scheduling performance of the system CPU and the memory resources is not high, and the super fusion service cannot be operated with the best performance. In addition, if the super fusion service integrally uses more resources, a corresponding recovery mechanism does not exist when the system crashes, so that the super fusion service can stably run.

The information disclosed in this background section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person of ordinary skill in the art.

Disclosure of Invention

The invention aims to provide a resource isolation method under super fusion, which is suitable for a physical machine with super fusion service deployed, and can reasonably allocate resources under the condition that CPU and memory resources of the physical machine are limited, so as to ensure that the super fusion service runs with optimal performance and stability.

Another object of the present invention is to provide a device for isolating resources under super fusion, an electronic device, and a storage medium, which are suitable for a physical machine where a super fusion service is deployed, and which are capable of reasonably allocating resources under the condition that the CPU and memory resources of the physical machine are limited, so as to ensure that the super fusion service operates with optimal performance and stability.

In order to achieve the above object, an embodiment of the present invention provides a method for isolating resources under super fusion, where the method for isolating resources under super fusion includes:

judging whether the resources of the super fusion service group are sufficient, whether the resources of the virtual machine group are sufficient and whether the prepared resources are sufficient;

and in response to the insufficient resources of the super fusion service group, the insufficient resources of the virtual machine group and the insufficient resources of the preparation resources, configuring the resources of the virtual machine group to be the maximum value based on the scheduling service, killing the virtual machine which consumes the most memory based on the memory overflow mechanism, and further distributing the resources to the super fusion service based on the scheduling service.

In one or more embodiments of the present invention, the method for isolating resources under super fusion further includes:

and in response to the insufficient resources of the super fusion service group, the sufficient resources of the virtual machine group and the insufficient preparation resources, the process with the largest memory consumption in the super fusion service group is killed based on a memory overflow mechanism.

judging whether the resources of the super fusion service group are unevenly used or not;

and in response to uneven resource use of the super fusion service group, performing migration of the super fusion service among NUMA nodes.

In one or more embodiments of the present invention, the migration of the super-fusion service between NUMA nodes includes the steps of:

the load of a CPU and a memory for the super fusion service in each NUMA node is obtained;

judging whether the load difference value of the CPU or the memory among different NUMA nodes is larger than a preset threshold value or not;

in response to the load difference being greater than a preset threshold, further determining whether there is a service having a load approaching and less than the difference;

in response to there being a service with a load and a difference being close and less than the difference, the service is migrated to another NUMA node.

In one or more embodiments of the present invention, the migration of the super fusion service between the NUMA nodes further includes:

and in response to the load difference value being less than a preset threshold, migrating the service with the lowest load in the NUMA nodes with the highest load to another NUMA node.

In one or more embodiments of the invention, the hyper-converged service comprises a software defined networking service and a software defined storage service configured to use resources in the same NUMA node.

In one or more embodiments of the present invention, the determining whether the resources of the super fusion service group are sufficient includes:

judging whether the resources of the super fusion service group are larger than a super fusion resource threshold value or not;

responsive to being greater than the super-converged resource threshold, determining that the super-converged service group is resource deficient.

In one or more embodiments of the present invention, the determining whether the resources of the virtual machine set are sufficient includes:

judging whether the resources of the virtual machine set are larger than a virtual machine resource threshold value or not;

responsive to being greater than the virtual machine resource threshold, determining that resources of the virtual machine group are insufficient.

In one or more embodiments of the present invention, the preparation resources are composed of resources reserved by all NUMA nodes, and each NUMA node is configured with resources for a super fusion service and resources for a virtual machine, the resources for the super fusion service in all NUMA nodes are configured as the super fusion service group, and the resources for the virtual machine in all NUMA nodes are configured as the virtual machine group.

The embodiment of the invention provides a resource isolation device under super fusion, which comprises:

the judging module is used for judging whether the resources of the super fusion service group are sufficient, whether the resources of the virtual machine group are sufficient and whether the prepared resources are sufficient;

and the resource allocation module is used for responding to insufficient resources of the super fusion service group, insufficient resources of the virtual machine group and insufficient preparation resources, configuring the resources of the virtual machine group to be maximum based on the scheduling service, killing the virtual machine which consumes the most memory based on the memory overflow mechanism, and further allocating the resources to the super fusion service based on the scheduling service.

An embodiment of the present invention provides an electronic device including:

at least one processor; and

at least one memory coupled to the at least one processor and storing a computer program for execution by the at least one processor, which when executed by the at least one processor, causes the electronic device to perform the method described above.

Embodiments of the present invention provide a computer readable storage medium having stored thereon a computer program which when executed by a machine implements the method described above.

Compared with the prior art, the invention configures the super-fusion service to belong to the same NUMA node, can improve the performance of accessing the memory across NUMA, and ensures that the super-fusion service operates in an optimal state. Meanwhile, when the resources used by the super fusion service are insufficient and the resources used by the virtual machine are insufficient, the virtual machine which consumes the most memory resources is killed through a memory overflow mechanism and the resources are distributed to the super fusion service, so that system breakdown can be avoided, and the super fusion service can be stably operated.

Drawings

FIG. 1 is a schematic diagram of a physical machine in the prior art;

FIG. 2 is a schematic diagram of a physical machine according to an embodiment of the present invention;

FIG. 3 is a flow chart of a method for resource isolation under super fusion according to an embodiment of the present invention;

FIG. 4 is a flow diagram of an NUMA inter-node migration service according to one embodiment of the invention;

FIG. 5 is a block diagram of a super-fusion resource isolation device according to an embodiment of the present invention.

Detailed Description

The following detailed description of embodiments of the invention is, therefore, to be taken in conjunction with the accompanying drawings, and it is to be understood that the scope of the invention is not limited to the specific embodiments.

Throughout the specification and claims, unless explicitly stated otherwise, the term "comprise" or variations thereof such as "comprises" or "comprising", etc. will be understood to include the stated element or component without excluding other elements or components.

According to the resource isolation method of the preferred embodiment of the invention, the method is suitable for the physical machine with the super fusion service deployed, and the physical machine can reasonably allocate resources under the condition of limited CPU and memory resources, so that the super fusion service can be ensured to run with optimal performance and stability.

As shown in fig. 2, the resource isolation method will be described in detail below by taking a physical machine having 2 CPUs and 8 memory banks, and 2 NUMA nodes (NUMA 0 and NUMA1, respectively) as an example. Wherein, each CPU corresponds to 4 memory banks, and each CPU comprises 24 cores (Die), 48 cores in total; the capacity of each memory bank is 16G, and the total capacity is 128G; each NUMA node includes 24 cores and 64G memory.

Of course, the present invention is not limited to the physical machine, but is also applicable to a physical machine including other number of CPUs and memory banks, and the number of cores of the CPUs and the capacity of the memory banks are not limited.

In order to realize the resource isolation method of the invention, so that the physical machine reasonably distributes resources under the condition of limited CPU and memory resources, the super fusion service is ensured to run with optimal performance and stability, and the physical machine is required to be correspondingly configured to meet the actual demands.

Specifically, first, which resources are used by the hyper-fusion service is configured, i.e., which cores and which memories of the CPU are used by the hyper-fusion service is configured. In particular implementations, a first core of a first CPU is configured for use by a Software Defined Network (SDN) service, and a memory of a NUMA0 node is configured for use by the SDN service; the second core and the third core of the first CPU are configured for use with a Software Defined Storage (SDS) service, and the memory of the NUMA0 node is configured for use with the Software Defined Storage (SDS) service. Configuring a first core of a second CPU for use by a super fusion Cloud Management Platform (CMP) service, and configuring a memory of the NUMA1 node for use by the super fusion Cloud Management Platform (CMP) service; the second core and the third core of the second CPU are configured for use by a Software Defined Computing (SDC) service, and the memory of the NUMA1 node is configured for use by the Software Defined Computing (SDC) service. In practice, for example, mysql service of one of the super fusion management services may be bound to a specific NUMA node by the following instructions.

#numactl-physcopubnd=0, 3-locallatamomysqld_multi-defaults-extra-file=/etc/mysqld_multi.cnfstart, wherein,

-physiccpubnd specifies the bound cpu node;

-locallloc means using memory means, without interleaving, to avoid degrading performance;

mysqld_multi is a mysql instance launch command.

By establishing NUMA nodes and enabling some super-converged services to use resources in the NUMA nodes, namely configuring some super-converged services to belong to the same NUMA node, the problem that in the prior art, the performance of accessing memory across NUMA is low due to the fact that some super-converged services do not belong to the same NUMA node is avoided, namely configuring some super-converged services to belong to the same NUMA node, the performance of accessing memory across NUMA can be improved, and the super-converged services can operate in an optimal state.

Further, to avoid the impact of network latency on distributed storage, software defined network (SD N) services and Software Defined Storage (SDs) services need to be configured in the same NUMA node. Meanwhile, the physical machine network card also needs CPU resources when forwarding the traffic, so the resources used by the network card for forwarding the traffic need to be the same as the resources used by the Software Defined Network (SDN) service, as shown in fig. 2, so that the network card forwards the traffic by using the first core of the first CPU. In practice, the resources used by the network to forward traffic may be configured in a manner:

first, the system interrupt service is turned off by the following instructions:

serviceirqbalancestop

finally, binding a specific nuclear forwarding message through an instruction:

further, for each NUMA node, a certain resource is reserved, and the resources form a spare resource (buffer), that is, for each NUMA node, a certain number of CPU cores and a certain capacity of memory are reserved as spare resources, so that when the resources are insufficient, the spare resources are dynamically allocated to the service with insufficient resources, such as the spare resources are allocated to the super-fusion service or the virtual machine. In the above physical machine, each NUMA node further reserves 2 cores and 16G memory as reserved resources, that is, in the above physical machine, 4 cores and 32G memory are reserved as reserved resources, so that the reserved resources are dynamically allocated to the service with insufficient resources when the resources are insufficient, and the reserved cores and the reserved memory capacity can be set according to actual requirements.

Further, after a certain preparation resource is reserved, resources for the super fusion service in all NUMA nodes are configured as a Group and marked as a super fusion service Group, as shown in fig. 2, a first core of the first CPU, a second core of the first CPU, a third core of the first CPU, a memory corresponding to the NUMA0 node, a first core of the second CPU, a second core of the second CPU, a third core of the second CPU and a memory corresponding to the NUMA1 node are configured as a Group and marked as a super fusion service Group1. Meanwhile, resources for virtual machines in all NUMA nodes are configured as a Group and marked as a virtual machine Group, and as shown in FIG. 2, a sixth core to a twenty-third core of the first CPU and a sixth core to a twenty-third core of the second CPU are configured as a virtual machine Group2.

After the super fusion service group and the virtual machine group are configured, a total memory limit (i.e. a resource threshold) is further set for each group, so as to limit the maximum memory used by each group. That is, a total memory threshold is configured for each group, and the maximum memory used by each group does not exceed the total memory threshold. In particular implementations, a total memory threshold may be configured for each group through cgroup. Where cgroup is a mechanism to hierarchically organize processes, allocating system resources in a controlled manner along a hierarchy. cgroup is a short term for control group, a feature provided by the Linux kernel to limit and isolate the use of system resources by a set of processes. Specific management of different resources is accomplished by the division of work for each subsystem. When the memory used by the super fusion service set and/or the virtual machine set exceeds the corresponding total memory threshold, a trigger OOM (OutofMemory) is triggered.

In this embodiment, the super fusion service group is created by creating group1 and dividing mysql in the super fusion service into the group. The creation of a virtual machine group can be seen as follows.

#systemd-run--unit＝mysql--slice＝group1top-b

Further, the resources of the super fusion service group are limited by the following instructions, for example, group1 is set to cpu to use the upper limit 6 cores, and the memory is 32G:

#systemctlset-propertygroup.sliceCPUShares＝6000MemoryLimit＝32G

the CPU is set to use the upper limit 42 cores, and the memory is a 95G memory (a group is established by default as a group called machine. Slice, which is specially used for the virtual machine)

#systemctlset-propertymachine.sliceCPUShares＝42000MemoryLimit＝96G

Further, after the total memory threshold is configured for the super-converged service group and the virtual machine group, a corresponding memory threshold is configured for each service, that is, a corresponding upper memory limit is configured for each service to limit the maximum memory capacity used by each service. The order in which the total memory thresholds are configured for the super fusion service group and the virtual machine (service) group is not limited. In particular implementations, a corresponding memory threshold may be configured for each service through cgroup.

After the above-mentioned series of configurations are completed for the physical machine, the described resource isolation method can be implemented so as to make the physical machine reasonably allocate resources under the condition of limited CPU and memory resources, and ensure that the super-fusion service can be run with optimum performance and stability.

As shown in fig. 3, the resource isolation method includes the steps of:

firstly, judging whether the resources of the super fusion service group are sufficient, whether the resources of the virtual machine group are sufficient and whether the prepared resources are sufficient;

and finally, responding to insufficient resources of the super fusion service group, insufficient resources of the virtual machine group and insufficient preparation resources, configuring the resources of the virtual machine group to be maximum based on the scheduling service, killing the virtual machine which consumes the most memory based on a memory overflow mechanism, and further distributing the resources to the super fusion service based on the scheduling service.

Specifically, in order to ensure that the super fusion service can have sufficient resources to operate and avoid system crash, when the super fusion service is implemented, whether the resources of the super fusion service group are sufficient, whether the resources of the virtual machine group are sufficient, and whether the preparation resources are sufficient are firstly judged. When the resources of the three are insufficient, namely the resources of the super fusion service group are insufficient, the resources of the virtual machine group are insufficient and the prepared resources are insufficient, the virtual machine which consumes the most memory resources is killed at the moment, and the resources are distributed to the super fusion service for use through a scheduling service (schedule-service), so that the super fusion service can stably run. In the above-mentioned physical machine, when the resources of the super-fusion service group, the resources of the virtual machine group and the prepared resources are insufficient, the scheduling service of the operating system configures the resources of the virtual machine group to include 38 cores and 64G memory, and at this time, the cgroup schedules the OOM mechanism of the system to kill the virtual machine that consumes the most memory. That is, after the resources of the virtual machine set are configured to include 38 cores and 64G memory, the OOM mechanism of the system is triggered to kill the virtual machine that consumes the most memory.

In this embodiment, when judging whether the resources of the super fusion service group are sufficient, the following manner may be adopted:

and judging whether the resources of the super fusion service group are larger than a super fusion resource threshold value. And when the resources of the super fusion service group are larger than the super fusion resource threshold, determining that the resources of the super fusion service group are insufficient, otherwise, determining that the resources of the super fusion service group are sufficient. That is, when the resources used by the super fusion service exceed the pre-configured super fusion resource threshold, the resources of the super fusion service group are determined to be insufficient, otherwise, the resources of the super fusion service group are determined to be sufficient.

Similarly, when judging whether the resources of the virtual machine set are sufficient, the following method can be adopted to judge:

and judging whether the resources of the virtual machine set are larger than a virtual machine resource threshold value. And when the resources of the virtual machine set are larger than the virtual machine resource threshold, determining that the resources of the virtual machine set are insufficient, otherwise, determining that the resources of the virtual machine set are sufficient. That is, when the resources used by the virtual machine exceed the pre-configured virtual machine resource threshold, the resources of the virtual machine set are determined to be insufficient, otherwise, the resources of the virtual machine set are determined to be sufficient.

Further, the resource isolation method further comprises:

and in response to the insufficient resources of the super-fusion service group, the insufficient resources of the virtual machine group and the insufficient preparation resources, triggering a memory overflow mechanism at the moment, and further killing the process with the maximum memory consumption in the super-fusion service group based on the memory overflow mechanism so as to ensure that the super-fusion service can stably run.

Further, the resource isolation method further comprises:

and when the resources of the super fusion service group are sufficient, further judging whether the resources of the super fusion service group are not used uniformly, and when the resources of the super fusion service group are not used uniformly, carrying out service migration among NUMA nodes. In the physical machine, when the resource usage in the super-converged service group is uneven, the scheduling service (Schedule-service) can dynamically perform service migration between NUMA nodes through the cgroup (control group), for example, migrate the service in the NUMA0 node to the NUMA1, so as to ensure the stable operation of the super-converged server. A Cgroup may allocate resources such as CPU time, system memory, network bandwidth, or a combination of these resources for a user-defined group of tasks running in the system.

As shown in fig. 4, in this embodiment, service migration between NUMA nodes may be performed by:

first, the load of the CPU and the memory for the super fusion service in each NUMA node is obtained. In practice, the load of the CPU and the memory for the super fusion service in each NUMA node can be periodically obtained by the scheduling service. The load of the CPU and the memory for the super fusion service in each NUMA node is obtained through each 1min of the scheduling service. The time can be configured according to the actual requirements.

Finally, judging whether the load difference value of the CPU or the memory among different NUMA nodes is larger than a preset value; the preset value here is configured when the scheduling service is installed. When the load difference is greater than a preset value, if the difference between the load 0 and the load 1 is greater than 50% of the sum of NUMA node resources, further judging whether a service with the load and the difference being close to each other and smaller than the difference exists. When there is a service whose load and difference are close and less than the difference, the service is migrated, i.e., to another NUMA node. If there is a service with a load and a difference close to and less than the difference in the NUMA0 node, the service is migrated to the NUMA1 node to ensure that the service can run stably.

When no service with the load close to the difference value and smaller than the difference value exists, the service with the lowest load in the NUMA node with the highest load is migrated, for example, the service with the lowest load in the NUMA0 node is migrated to the NUMA1 node, so that the service can be ensured to run stably.

In this embodiment, the migration of the service may be performed by the following instructions:

cgclassify-gsubsystems:path_to_cgrouppidlist

where subsystems are comma separated lists of subsystems, or start processes in a hierarchy associated with all available subsystems.

Path_to_cgroup is a path to cgroup in its hierarchy;

pidlist is a list of Process Identifiers (PIDs) separated by spaces.

If the process with PID 1701 is moved to group1 in the cgroup, the following command is used:

cgclassify-gcpu,memory:group11701。

as shown in fig. 5, a resource isolation device according to a preferred embodiment of the present invention includes a judging module and a resource allocation module. The judging module is used for judging whether the resources of the super fusion service group are sufficient, whether the resources of the virtual machine group are sufficient and whether the prepared resources are sufficient or not; the resource allocation module is used for responding to insufficient resources of the super fusion service group, insufficient resources of the virtual machine group and insufficient preparation resources, configuring the resources of the virtual machine group to be maximum based on the scheduling service, killing the virtual machine which consumes the most memory based on the memory overflow mechanism, and further allocating the resources to the super fusion service based on the scheduling service.

Specifically, in order to ensure that the super fusion service can have sufficient resources to operate and avoid system crash, when the super fusion service is implemented, a judging module is firstly used for judging whether the resources of the super fusion service group are sufficient, judging whether the resources of the virtual machine group are sufficient and judging whether the preparation resources are sufficient. When the resources of the three are insufficient, namely the resources of the super fusion service group are insufficient, the resources of the virtual machine group are insufficient and the prepared resources are insufficient, the virtual machine which consumes the most memory resources is killed at the moment, and the resources are distributed to the super fusion service for use through a scheduling service (schedule-service), so that the super fusion service can stably run. In the above-mentioned physical machine, when the resources of the super-fusion service group, the resources of the virtual machine group and the prepared resources are insufficient, the scheduling service of the operating system configures the resources of the virtual machine group to include 38 cores and 64G memory, and at this time, the cgroup schedules the OOM mechanism of the system to kill the virtual machine that consumes the most memory. That is, after the resources of the virtual machine set are configured to include 38 cores and 64G memory, the OOM mechanism of the system is triggered to kill the virtual machine that consumes the most memory.

Further, in the implementation, when the judging module judges whether the resources of the super fusion service group are insufficient through the following steps, the judging module can judge by the following modes:

Similarly, when judging whether the resources of the virtual machine set are insufficient, the following method can be adopted to judge:

The embodiment of the invention also discloses an electronic device, which can realize the above-mentioned method for isolating resources under super fusion, and can reasonably allocate resources under the condition of limited CPU and memory resources, thereby ensuring that the super fusion service runs with optimal performance and stability. The electronic device may include, but is not limited to: personal computers, server computers, workstations, desktop computers, laptop computers, notebook computers, mobile electronic devices, smart phones, tablet computers, cellular phones, personal Digital Assistants (PDAs), handsets, messaging devices, wearable electronic devices, consumer electronic devices, and the like. The electronic equipment can realize the above-mentioned method for isolating resources under the super fusion, can reasonably allocate resources under the condition of limited CPU and memory resources, and ensures that the super fusion service runs with optimal performance and stability. In particular, the electronic device comprises at least one memory, at least one processor, and a computer program, the at least one memory being coupled to the at least one processor, wherein the computer program is stored in the memory and is executable in the processor, such as the computer program being a super-converged resource isolation program or the like. In practice, the processor may implement the steps of the above method when executing the computer program, such as determining whether the resources of the super fusion service group are sufficient and whether the resources of the virtual machine group are sufficient.

The computer program herein may be divided into one or more units, which are stored in and executed by the memory to accomplish the present invention. Wherein one or more of the units may be a series of computer program instruction segments capable of performing the specified functions, the computer program instruction segments being adapted to describe the execution of a computer program in said electronic device.

It should be noted that the electronic device herein includes, but is not limited to, a memory, a processor, and a computer program as described above, but may also include other devices, such as an input device (e.g., a keyboard, etc.) for inputting instructions, a display screen for displaying negotiation results, a communication interface, etc., which communicate with each other via a bus.

An embodiment of the present invention also discloses a computer readable storage medium, where a computer program is stored, where the computer program can implement the above-mentioned resource isolation method under super fusion when executed by a processor. Wherein the computer program comprises computer program code, which may be in the form of source code, executable files, or in some intermediate form, etc., the computer readable medium may comprise any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, randomAccessMemory), etc.

It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The foregoing descriptions of specific exemplary embodiments of the present invention are presented for purposes of illustration and description. It is not intended to limit the invention to the precise form disclosed, and obviously many modifications and variations are possible in light of the above teaching. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application to thereby enable one skilled in the art to make and utilize the invention in various exemplary embodiments and with various modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims and their equivalents.

Claims

1. The method for isolating the resources under the super fusion is characterized by comprising the following steps of:

2. The method for isolating resources under super fusion according to claim 1, further comprising:

3. The method for isolating resources under super fusion according to claim 1, further comprising:

4. The method for isolating resources under super fusion according to claim 3, wherein said migration of super fusion services between NUMA nodes comprises the steps of:

5. The method for resource isolation under super fusion of claim 4, wherein said migration of super fusion services between NUMA nodes further comprises:

in response to there being no service with a load approaching and less than the difference, migrating the least loaded service from the highest loaded NUMA nodes to another NUMA node.

6. The method of resource isolation under hyper-fusion of claim 1, wherein the hyper-fusion service comprises a software defined networking service and a software defined storage service configured to use resources in the same NUMA node.

7. The method for resource isolation under super fusion according to claim 1, wherein said determining whether resources of the super fusion service group are sufficient comprises:

8. The method for resource isolation under super fusion according to claim 1, wherein said determining whether resources of the virtual machine set are sufficient comprises:

9. The method for isolating resources under super fusion according to claim 1, wherein the preparation resources are composed of resources reserved by all NUMA nodes, each NUMA node is configured with resources for super fusion services and resources for virtual machines, the resources for super fusion services in all NUMA nodes are configured as the super fusion service group, and the resources for virtual machines in all NUMA nodes are configured as the virtual machine group.

10. The utility model provides a resource isolation device under super fusion, its characterized in that, resource isolation device under super fusion includes:

11. An electronic device, the electronic device comprising:

at least one processor; and

at least one memory coupled to the at least one processor and storing a computer program for execution by the at least one processor, the computer program, when executed by the at least one processor, causing the electronic device to perform the method of any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that a computer program is stored thereon, which computer program, when being executed by a machine, implements the method of any of claims 1 to 9.