Method for isolating computing power resources of CPU (Central Processing Unit) at tenant side of cloud platform

Abstract

The invention discloses a cloud platform tenant side CPU computing power resource isolation method, and relates to the technical field of cloud computing. Comprising the steps of 1, creating a CPU computing power resource isolation system, 2, defining computing node CPU isolation related parameters through a CPU isolation configuration module and carrying out parameter configuration, 3, configuring isolcpus parameters in a target computing node grub through an operating system CPU initialization module, appointing a CPU core needing to be isolated according to the isolcpus parameters, and carrying out CPU isolation on the CPU core needing to be isolated according to the isolcpus parameters. And 4, reading CPU related parameters defined by the CPU isolation configuration module through a management side CPU initialization module, calculating a CPU core set reserved for the management side service by the node according to the number of CPU cores reserved for the management side service, and completing CPU core binding of the management side service of the target calculation node. 5, CPU related parameters defined by the CPU isolation configuration module are read through a tenant side CPU initialization module, when an automatic selection is true, cpushadset is calculated, according to the cpushadset and the cpushadset, initialization of a CPU set used by the kernel binding virtual machine and the kernel non-binding virtual machine is completed, and CPU isolation of the tenant side and the management side is achieved.

Description

TECHNICAL FIELD

[0001] The application discloses a method for CPU computing resource isolation on the tenant side of a cloud platform, and relates to the technical field of cloud computing. BACKGROUND

[0002] In a cloud computing environment, the mode that the management side service shares underlying physical resources such as CPU, memory, storage, etc. with tenant virtual machines greatly improves resource utilization and reduces the cost of cloud vendors. However, there may be a situation of contention in CPU resource allocation, which on the one hand affects the stable operation of the management side service, and on the other hand also affects the performance of the tenant virtual machine, seriously affecting the stability of the cloud platform and user experience.

[0003] Among them, the tenant side CPU resource contention problem is easy to lead to:

[0004] Sudden load: the load of some tenants may suddenly increase, occupying a large amount of CPU resources, affecting the normal operation of other tenants.

[0005] Unclear priority: lack of dynamic priority management mechanism, leading to key business tenants unable to obtain sufficient CPU resource guarantee.

[0006] Uneven resource allocation: static resource allocation strategy cannot adapt to the dynamic changes of tenant load, leading to low resource utilization or performance fluctuation.

[0007] And in the management side CPU resource contention problem, it is easy to lead to:

[0008] Management service performance decline: the management service cannot obtain sufficient CPU resources in time, affecting the stable and efficient operation of the cloud platform.

[0009] Tenant performance is affected: the high priority of the management service may occupy the CPU resources of the tenant application, leading to the performance decline of the tenant virtual machine.

[0010] Resource allocation conflict: lack of effective resource isolation mechanism, leading to resource allocation conflict between the management service and the tenant application.

[0011] The current OpenStack-based cloud platform supports the tenant-side CPU isolation function by modifying the CPU set cpu_shared_set used by the normal non-bonded core virtual machine, the CPU set cpu_dedicated_set used by the bonded core virtual machine, and the reserved CPU core number reserved_host_cpus in the configuration file. However, the configuration of the above CPU set is generally different for computing nodes of different models, and needs to rely on manual selection and allocation, and the CPU isolation related configuration is also affected by the operating system isolcpus configuration. This greatly increases the difficulty of configuration, and also requires the operation and maintenance personnel to have certain prior knowledge. SUMMARY

[0012] The present application provides a method for isolating CPU computing power resources on the tenant side of a cloud platform, which solves the problems of CPU resource contention and insufficient isolation on the tenant side and the management side of the computing node in a multi-tenant cloud platform. Only the number of CPU cores reserved for the management side needs to be configured, and the CPU isolation configuration of the computing node can be automatically initialized without the need for operation and maintenance personnel to modify the complex configuration of the OpenStack-based cloud platform for CPU isolation function, without the need for manual configuration of bonded cores for management side services and without the need for configuration of operating system CPU isolation configuration. This can initialize the CPU isolation related configuration at each level of the computing node, solve the pain points of CPU resource contention in a multi-tenant cloud platform, improve the resource utilization, service quality and operation and maintenance efficiency of the cloud platform, and provide reliable technical support for large-scale multi-tenant cloud environments.

[0013] The specific scheme provided by the present application is:

[0014] The present application provides a method for isolating CPU computing power resources on the tenant side of a cloud platform, which includes:

[0015] Step 1: Create a CPU computing power resource isolation system, which includes a CPU isolation configuration module, an operating system CPU initialization module, a management side CPU initialization module and a tenant side CPU initialization module,

[0016] Step 2: Define the CPU isolation related parameters and perform parameter configuration through the CPU isolation configuration module, the CPU isolation related parameters including whether to automatically select the CPU core auto_cpu_selection, the number of CPU cores reserved for the management side reserved_host_cpus, the CPU set cpu_shared_set used by the non-bonded core virtual machine, and the CPU set cpu_dedicated_set used by the bonded core virtual machine,

[0017] Step 3: Configure the isolcpus parameter in the target compute node's grub through the operating system CPU initialization module. Specify the CPU cores to be isolated based on the isolcpus parameter. The operating system CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, configure the CPU set corresponding to cpu_dedicated_set to grub and restart the compute node to complete the isolcpus initialization, ensuring that the CPU set in cpu_dedicated_set is not scheduled by the operating system kernel.

[0018] Step 4: Read the CPU-related parameters defined by the CPU isolation configuration module through the management-side CPU initialization module. When auto_cpu_selection is true, obtain reserved_host_cpus. Calculate the CPU core set reserved for the management-side service based on the number of CPU cores reserved for the management-side service, and complete the CPU core binding of the target computing node management-side service.

[0019] Step 5: The CPU initialization module on the tenant side reads the CPU-related parameters defined in the CPU isolation configuration module. When auto_cpu_selection is true, cpu_shared_set is calculated. Based on cpu_dedicated_set and cpu_shared_set, the CPU sets used by bound core VMs and non-bound core VMs are initialized to achieve CPU isolation between the tenant side and the management side.

[0020] Furthermore, when configuring parameters in step 2 of the method for isolating CPU computing power resources on the tenant side of a cloud platform, auto_cpu_selection is passed in true, which means that automatic selection of the CPU core to be isolated is enabled, and cpu_shared_set and cpu_dedicated_set can pass in comma-separated integers, integer ranges, and integers starting with ^ indicate that an integer in the integer range is to be excluded.

[0021] Furthermore, in step 4 of the method for isolating CPU computing power resources on the cloud platform tenant side, calculating the CPU core set of the node reserved for the management side service based on the number of CPU cores reserved for the management side service includes:

[0022] Obtain the CPU core set corresponding to each numa of the target computing node, remove the CPU set configured by nodeisolcpus, remove the CPU set configured by cpu_dedicated_set, take out a smallest CPU core from NUMA0 in turn, and add the reserved CPU set of the management side, so that the reserved CPU set is used to increase the binding core of the management side CPU when the management side service of the target computing node is started, and it is guaranteed that the management side service uses the reserved CPU core of the computing node.

[0023] Further, the step 5 of the cloud platform tenant side CPU computing resource isolation method comprises: obtaining the total CPU core set of the target computing node, removing the CPU set reserved for the management side calculated by the management side CPU initialization module, removing the CPU set configured by cpu_dedicated_set, removing the CPU set configured by nodeisolcpus, and finally rendering the final set to cpu_shared_set for use by the non-binding core virtual machine.

[0024] The application further provides a cloud platform tenant side CPU computing resource isolation system, which comprises a CPU isolation configuration module, an operating system CPU initialization module, a management side CPU initialization module and a tenant side CPU initialization module,

[0025] The CPU isolation configuration module defines and configures the CPU isolation related parameters of the computing node, and the CPU isolation related parameters comprise whether to automatically select the CPU core auto_cpu_selection, the number of CPU cores reserved for the management side reserved_host_cpus, the CPU set cpu_shared_set used by the non-binding core virtual machine, and the CPU set cpu_dedicated_set used by the binding core virtual machine.

[0026] The operating system CPU initialization module configures the isolcpus parameter in the grub of the target computing node, specifies the CPU core to be isolated according to the isolcpus parameter, reads the CPU related parameters defined by the CPU isolation configuration module, configures the CPU set corresponding to cpu_dedicated_set to the grub when auto_cpu_selection is true, and restarts the computing node to complete the isolcpus initialization, so that the CPU set in cpu_dedicated_set is not scheduled by the operating system kernel.

[0027] The management side CPU initialization module reads the CPU related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, the reserved_host_cpus are obtained, the CPU core set reserved for the management side service of the target computing node is calculated according to the number of CPU cores reserved for the management side service, and the CPU core binding of the management side service of the target computing node is completed.

[0028] The tenant side CPU initialization module reads the CPU related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, the cpu_shared_set is calculated, and the initialization of the CPU set used by the bound core virtual machine and the non-bound core virtual machine is completed according to the cpu_dedicated_set and the cpu_shared_set, so as to realize the CPU isolation of the tenant side and the management side.

[0029] Further, when the CPU isolation configuration module of the cloud platform tenant side CPU computing resource isolation system performs parameter configuration, auto_cpu_selection is input as true, indicating that the automatic selection of the CPU core to be isolated is started, cpu_shared_set and cpu_dedicated_set can be input as integers, integer intervals, and integers starting with ^ to indicate the removal of certain integers in the integer interval.

[0030] Further, the management side CPU initialization module of the cloud platform tenant side CPU computing resource isolation system calculates the CPU core set reserved for the management side service of the target computing node according to the number of CPU cores reserved for the management side service, including:

[0031] The CPU core set corresponding to each numa of the target computing node is obtained, the CPU set configured by the node isolcpus is removed, the CPU set configured by the cpu_dedicated_set is removed, a smallest CPU core is sequentially taken out from NUMA0, and is added to the CPU set reserved for the management side. When the management side service of the target computing node is started, the reserved CPU set is used to increase the binding of the management side CPU, so that the management side service uses the CPU core reserved by the computing node.

[0032] Further, the tenant-side CPU initialization module of the system for isolating CPU computing resources of a cloud platform tenant side calculates cpu_shared_set, including: obtaining a complete CPU core set of a target computing node, removing a CPU set reserved for the management side calculated by a management-side CPU initialization module, removing a CPU set configured by cpu_dedicated_set, removing a CPU set configured by node_isolcpus, and finally rendering the final set to cpu_shared_set for use by a non-bonded core virtual machine.

[0033] The present application has the advantages of:

[0034] (1) A simple way is provided, in which, during initialization of a cloud platform computing node, an operation and maintenance personnel only needs to configure the number of CPU cores to be reserved for a management side service, and then the CPU isolation configuration of the computing node system, the management side process and the tenant side can be automatically completed, and the tenant side CPU isolation function is realized.

[0035] (2) The CPU isolation configuration module supports the functions of customizing isolated cores and automatically selecting CPU cores.

[0036] (3) After the operation and maintenance personnel runs the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate and configure the CPU set corresponding to the isolcpus of the operating system, restart the physical machine, and automatically complete the system-level CPU core isolation.

[0037] After the operation and maintenance personnel runs the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate the CPU set reserved for the management side process, and bind the process and the CPU core after the management side service is started, so as to automatically complete the CPU core isolation of the management side service.

[0038] (4) After the operation and maintenance personnel runs the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate the CPU set cpu_shared_set used by the ordinary non-bonded core virtual machine in the OpenStack configuration, the CPU set cpu_dedicated_set used by the bonded core virtual machine, and the number of reserved CPU cores reserved_host_cpus, automatically initialize the CPU isolation related configuration, and after the virtual machine is created, the corresponding configuration can be used to complete the CPU core isolation of the tenant side virtual machine. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a system application schematic diagram of the present application.

[0040] Figure 2 It is the parameter structure diagram of the CPU configuration module.

[0041] Figure 3 This is the core flow chart of the tenant-side CPU initialization module. DETAILED DESCRIPTION

[0042] The present invention will be further described below with reference to the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention.

[0043] Example 1

[0044] The present invention provides a method for isolating CPU computing power resources on a cloud platform tenant side, comprising:

[0045] Step 1: Create a CPU computing resource isolation system, which includes a CPU isolation configuration module, an operating system CPU initialization module, a management-side CPU initialization module, and a tenant-side CPU initialization module.

[0046] Step 2: Define and configure CPU isolation parameters for the compute node using the CPU isolation configuration module. These parameters include whether to automatically select CPU cores (auto_cpu_selection), the number of CPU cores reserved for the management side (reserved_host_cpus), the CPU set used by non-core VMs (cpu_shared_set), and the CPU set used by core-bound VMs (cpu_dedicated_set). The auto_cpu_selection parameter can be set to true or false. If true, automatic selection of CPU cores for isolation is enabled. In this case, you only need to specify the number of CPU cores reserved for the management side (reserved_host_cpus). The CPU set used by non-core VMs is automatically calculated based on this number and configured as cpu_shared_set. If false, automatic selection of CPU cores for isolation is disabled. In this case, you can specify cpu_shared_set to customize the CPU cores used by non-core VMs. You can also specify cpu_dedicated_set to configure the CPU set used by core-bound VMs. This configuration is not affected by auto_cpu_selection. cpu_shared_set and cpu_dedicated_set accept comma-separated integers, integer ranges, and integers starting with "^" to exclude an integer from the range. For example, if auto_cpu_selection is set to true and cpu_dedicated_set is configured, CPU core selection will be automatic.

[0047] Step 3: Configure the isolcpus parameter in the grub of the target computing node through the operating system CPU initialization module, and specify the CPU cores that need to be isolated according to the isolcpus parameter, so that they are not used by the kernel scheduler. The operating system CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module, and when auto_cpu_selection is true, the CPU set corresponding to cpu_dedicated_set is configured to grub and the computing node is restarted to complete the isolcpus initialization, ensuring that the CPU set in cpu_dedicated_set is not used by the operating system kernel scheduler.

[0048] Step 4: The management-side CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module, and when auto_cpu_selection is true, it obtains reserved_host_cpus and calculates the CPU core set reserved for the management-side service according to the number of CPU cores reserved for the management-side service, completing the CPU core binding of the management-side service of the target computing node. The CPU core set reserved for the management-side service of the computing node includes:

[0049] The CPU core set corresponding to each numa of the target computing node is obtained, the CPU set configured by node isolcpus is removed, the CPU set configured by cpu_dedicated_set is removed, and a smallest CPU core is sequentially taken out from NUMA0, and added to the CPU set reserved for the management side. When the management-side service of the target computing node is started, the reserved CPU set is used to increase the binding of the management-side CPU, ensuring that the management-side service uses the CPU core reserved by the computing node.

[0050] Step 5: The tenant-side CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module, and when auto_cpu_selection is true, it calculates cpu_shared_set and completes the initialization of the CPU set used by the bound-core virtual machine and the non-bound-core virtual machine according to cpu_dedicated_set and cpu_shared_set, realizing the isolation of the tenant-side and management-side CPU. The calculation of cpu_shared_set includes: obtaining the entire CPU core set of the target computing node, removing the CPU set reserved for the management side calculated by the management-side CPU initialization module, removing the CPU set configured by cpu_dedicated_set, and removing the CPU set configured by node isolcpus. The final set is rendered to cpu_shared_set for use by non-bound-core virtual machines.

[0051] At this time, the initialization of the CPU set used by the bound-core virtual machine and the non-bound-core virtual machine is completed.

[0052] The user can open the general ECS virtual machine or the high-performance bound-core virtual machine in a conventional manner, and after the virtual machine is successfully created, the CPU set initialized in the configuration file can be used to realize the tenant-side and management-side CPU isolation.

[0053] Embodiment 2

[0054] The application further provides a system for isolating CPU computing power resources on the tenant side of a cloud platform, comprising a CPU isolation configuration module, an operating system CPU initialization module, a management-side CPU initialization module, and a tenant-side CPU initialization module,

[0055] The CPU isolation configuration module defines and configures the CPU isolation related parameters of the computing node, and the CPU isolation related parameters comprise whether to automatically select the CPU core (auto_cpu_selection), the number of CPU cores reserved for the management side (reserved_host_cpus), the CPU set used by the non-bound-core virtual machine (cpu_shared_set), and the CPU set used by the bound-core virtual machine (cpu_dedicated_set),

[0056] The operating system CPU initialization module configures the isolcpus parameter in the grub of the target computing node, specifies the CPU core to be isolated according to the isolcpus parameter, reads the CPU related parameters defined by the CPU isolation configuration module, when auto_cpu_selection is true, configures the CPU set corresponding to cpu_dedicated_set to the grub and restarts the computing node, completes the initialization of isolcpus, and ensures that the CPU set in cpu_dedicated_set is not scheduled by the operating system kernel,

[0057] The management-side CPU initialization module reads the CPU related parameters defined by the CPU isolation configuration module, when auto_cpu_selection is true, obtains reserved_host_cpus, calculates the CPU core set reserved for the management-side service of the computing node according to the number of CPU cores reserved for the management-side service, and completes the CPU core binding of the management-side service of the target computing node,

[0058] The tenant-side CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module, and when auto_cpu_selection is true, calculates cpu_shared_set, and completes the initialization of the CPU set used by the bound-core virtual machine and the non-bound-core virtual machine according to cpu_dedicated_set and cpu_shared_set, to realize CPU isolation on the tenant side and the management side.

[0059] The information interaction and execution process between the modules in the system are based on the same concept as the method embodiments of the application, and the specific content can be referred to the description in the method embodiments of the application, which will not be repeated here.

[0060] Similarly, the advantages of the device of the application are:

[0061] (1) When the cloud platform computing node is initialized, the operation and maintenance personnel only need to configure the number of CPU cores reserved for the management side service, and the CPU isolation configuration of the computing node system, the management side process and the tenant side can be automatically completed, to realize the tenant-side CPU isolation function.

[0062] (2) The CPU isolation configuration module supports the functions of customizing the isolated cores and automatically selecting the CPU cores.

[0063] (3) After the operation and maintenance personnel run the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate and configure the CPU set corresponding to the operating system isolcpus, restart the physical machine, and automatically complete the system-level CPU core isolation.

[0064] After the operation and maintenance personnel run the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate the CPU set reserved for the management side process, and bind the process and the CPU core after the management side service is started, to automatically complete the CPU core isolation of the management side service.

[0065] (4) After the operation and maintenance personnel run the computing node CPU initialization program, the program will read the number of CPU cores reserved for the management side service according to the CPU isolation configuration module, calculate the CPU set cpu_shared_set used by the ordinary non-bound-core virtual machine in the OpenStack configuration, the CPU set cpu_dedicated_set used by the bound-core virtual machine, and the number of reserved CPU cores reserved_host_cpus, automatically initialize the CPU isolation-related configuration, and after the virtual machine is created, the corresponding configuration can be used to complete the CPU core isolation of the tenant-side virtual machine.

[0066] It should be noted that not all steps and modules in the above processes and system structures are necessary, and some steps or modules can be omitted according to actual needs. The execution order of each step is not fixed and can be adjusted as needed. The system structure described in each of the above embodiments can be a physical structure or a logical structure, that is, some modules can be implemented by the same physical entity, or some modules can be implemented by multiple physical entities, or can be implemented by some components in multiple independent devices together.

[0067] The above-described embodiments are merely preferred embodiments of the present application and are not intended to limit the protection scope of the present application. Any equivalent replacement or transformation made by those skilled in the art based on the present application shall fall within the protection scope of the present application. The protection scope of the present application is defined by the claims.

Claims

1. A method for isolating CPU computing resources on the tenant side of a cloud platform, characterized by: include: Step 1: Create a CPU computing resource isolation system, which includes a CPU isolation configuration module, an operating system CPU initialization module, a management-side CPU initialization module, and a tenant-side CPU initialization module. Step 2: Define and configure the CPU isolation parameters of the compute node through the CPU isolation configuration module. The CPU isolation parameters include whether to automatically select the CPU core auto_cpu_selection, the number of CPU cores reserved for the management side reserved_host_cpus, the CPU set cpu_shared_set used by non-core bound virtual machines, and the CPU set cpu_dedicated_set used by core bound virtual machines. Step 3: Configure the isolcpus parameter in the target compute node's grub through the operating system CPU initialization module. Specify the CPU cores to be isolated based on the isolcpus parameter. The operating system CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, configure the CPU set corresponding to cpu_dedicated_set to grub and restart the compute node to complete the isolcpus initialization, ensuring that the CPU set in cpu_dedicated_set is not scheduled by the operating system kernel. Step 4: Read the CPU-related parameters defined by the CPU isolation configuration module through the management-side CPU initialization module. When auto_cpu_selection is true, obtain reserved_host_cpus. Calculate the CPU core set reserved for the management-side service based on the number of CPU cores reserved for the management-side service, and complete the CPU core binding of the target computing node management-side service. Step 5: The CPU initialization module on the tenant side reads the CPU-related parameters defined in the CPU isolation configuration module. When auto_cpu_selection is true, cpu_shared_set is calculated. Based on cpu_dedicated_set and cpu_shared_set, the CPU sets used by bound core VMs and non-bound core VMs are initialized to achieve CPU isolation between the tenant side and the management side.

2. A method for isolating CPU computing power resources on a cloud platform tenant side according to claim 1, characterized in that When configuring parameters in step 2, set true to auto_cpu_selection to enable automatic selection of isolated CPU cores. You can set comma-separated integers or integer ranges to cpu_shared_set and cpu_dedicated_set. Integers starting with ^ indicate that an integer in the range is to be excluded.

3. A method for isolating CPU computing power resources on a cloud platform tenant side according to claim 1, characterized in that In step 4, the CPU core set reserved for the management-side service is calculated based on the number of CPU cores reserved for the management-side service, including: Get the CPU core set corresponding to each NUMA of the target compute node, remove the CPU set configured by the node isolcpus, remove the CPU set configured by cpu_dedicated_set, and start from NUMA0. Take out the smallest CPU core in turn and add it to the CPU set reserved by the management side. When the management side service of the target compute node starts, use the reserved CPU set to increase the bound core to the management side CPU to ensure that all management side services use the CPU cores reserved by the compute node.

4. A method for isolating CPU computing power resources on a cloud platform tenant side according to claim 1, characterized in that Calculating cpu_shared_set in step 5 includes: obtaining the set of all CPU cores of the target computing node, removing the CPU set reserved for the management side calculated by the CPU initialization module on the management side, removing the CPU set configured by cpu_dedicated_set, removing the CPU set configured by node isolcpus, and finally rendering the set to cpu_shared_set for use by non-core-bound virtual machines.

5. A system for isolating CPU computing resources on the tenant side of a cloud platform, characterized by It includes CPU isolation configuration module, operating system CPU initialization module, management side CPU initialization module and tenant side CPU initialization module. The CPU isolation configuration module defines and configures the CPU isolation related parameters of the computing node. The CPU isolation related parameters include whether to automatically select the CPU core auto_cpu_selection, the number of CPU cores reserved for the management side reserved_host_cpus, the CPU set cpu_shared_set used by non-core bound virtual machines, and the CPU set cpu_dedicated_set used by core bound virtual machines. The operating system CPU initialization module configures the isolcpus parameter in the target computing node grub, and specifies the CPU core to be isolated according to the isolcpus parameter. The operating system CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, the CPU set corresponding to cpu_dedicated_set is configured in grub and the computing node is restarted to complete the isolcpus initialization, ensuring that the CPU set in cpu_dedicated_set is not scheduled by the operating system kernel. The management-side CPU initialization module reads the CPU-related parameters defined by the CPU isolation configuration module. When auto_cpu_selection is true, it obtains reserved_host_cpus and calculates the CPU core set reserved for the management-side service based on the number of CPU cores reserved for the management-side service. This completes the CPU core binding of the target computing node management-side service. The tenant-side CPU initialization module reads the CPU-related parameters defined in the CPU isolation configuration module. When auto_cpu_selection is true, it calculates cpu_shared_set and initializes the CPU sets used by core-bound VMs and non-core-bound VMs based on cpu_dedicated_set and cpu_shared_set, thus achieving CPU isolation between the tenant and management sides.

6. A system for isolating CPU computing power resources on a cloud platform tenant side according to claim 5, characterized in that When configuring parameters for the CPU isolation configuration module, passing true to auto_cpu_selection enables automatic selection of the CPU core to be isolated. Comma-separated integers or integer ranges can be passed to cpu_shared_set and cpu_dedicated_set. Integers starting with ^ indicate excluding an integer within the integer range.

7. A system for isolating CPU computing power resources on a cloud platform tenant side according to claim 5, characterized in that The management-side CPU initialization module calculates the set of CPU cores reserved for management-side services on the node based on the number of CPU cores reserved for management-side services, including: Get the CPU core set corresponding to each NUMA of the target compute node, remove the CPU set configured by the node isolcpus, remove the CPU set configured by cpu_dedicated_set, and start from NUMA0. Take out the smallest CPU core in turn and add it to the CPU set reserved by the management side. When the management side service of the target compute node starts, use the reserved CPU set to increase the bound core to the management side CPU to ensure that all management side services use the CPU cores reserved by the compute node.

8. A method for isolating CPU computing power resources on a cloud platform tenant side according to claim 5, characterized in that The tenant-side CPU initialization module calculates cpu_shared_set, including obtaining the set of all CPU cores in the target compute node, removing the set of CPUs reserved for the management side calculated by the management-side CPU initialization module, removing the set of CPUs configured in cpu_dedicated_set, and removing the set of CPUs configured in the node isolcpus. The final set is rendered into cpu_shared_set for use by non-core-bound VMs.

Images