CN115168028A - Task processing method and device - Google Patents

Abstract

The embodiment of the specification provides a task processing method and a task processing device, wherein the task processing method comprises the following steps: receiving a task allocation request aiming at a target virtual object, wherein the task allocation request carries a task to be allocated; determining at least two vCPUs corresponding to the target virtual object based on the task allocation request, and determining an initial vCPU from the at least two vCPUs, wherein the current resource occupation information of the initial vCPU is greater than that of other vCPUs; determining the initial vCPU as a target vCPU under the condition that the current resource occupation information of the initial vCPU is determined to meet the task processing condition of the task to be distributed; distributing the task to be distributed to the target vCPU for task execution; therefore, the problem of CPU scheduling loss caused by frequent subsidence of the idle vCPU to the host is avoided, and the physical CPU scheduling loss is further reduced.

Description

Task processing method and device

technical field

The embodiments of this specification relate to the field of computer technologies, and in particular, to a task processing method.

Background technique

With the development of virtualization technology, many Internet companies will use hardware virtualization capabilities for security isolation. In hardware virtualization scenarios, a source of CPU overhead that is difficult to eliminate is that each idle vCPU will frequently fall out of In the host, a large amount of CPU computing resources are consumed to execute the scheduling process on the host for the vCPU, resulting in high and difficult-to-optimize physical CPU scheduling losses. The host machine, the problem of physical CPU scheduling loss caused.

SUMMARY OF THE INVENTION

In view of this, the embodiments of this specification provide a task processing method. One or more embodiments of this specification also relate to a task processing apparatus, a computing device, a computer-readable storage medium, and a computer program, so as to solve the technical defects existing in the prior art.

According to a first aspect of the embodiments of the present specification, a task processing method is provided, including:

receiving a task assignment request for the target virtual object, wherein the task assignment request carries a task to be assigned;

Based on the task allocation request, at least two vCPUs corresponding to the target virtual object are determined, and an initial vCPU is determined from the at least two vCPUs, wherein the current resource occupancy information of the initial vCPU is greater than that of other vCPUs;

In the case of determining the current resource occupancy information of the initial vCPU and satisfying the task processing conditions of the to-be-allocated task, determining the initial vCPU as the target vCPU;

Allocate the to-be-allocated task to the target vCPU for task execution.

According to a second aspect of the embodiments of the present specification, a task processing apparatus is provided, including:

a receiving module, configured to receive a task assignment request for the target virtual object, wherein the task assignment request carries a task to be assigned;

a first determining module, configured to determine at least two vCPUs corresponding to the target virtual object based on the task allocation request, and determine an initial vCPU from the at least two vCPUs, wherein the current resources of the initial vCPU The occupancy information is larger than other vCPUs;

a second determining module, configured to determine the initial vCPU as the target vCPU under the condition that the current resource occupation information of the initial vCPU is determined and the task processing condition of the to-be-allocated task is satisfied;

The execution module is configured to allocate the to-be-allocated task to the target vCPU for task execution.

According to a third aspect of the embodiments of the present specification, a computing device is provided, including:

memory and processor;

The memory is used for storing computer-executable instructions, the processor is used for executing the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the task processing method are implemented.

According to a fourth aspect of the embodiments of the present specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the steps of the task processing method.

According to a fifth aspect of the embodiments of the present specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the task processing method.

The task processing method provided in this specification includes: receiving a task allocation request for a target virtual object, wherein the task allocation request carries a task to be allocated; and determining at least two tasks corresponding to the target virtual object based on the task allocation request. and determining an initial vCPU from the at least two vCPUs, wherein the current resource occupancy information of the initial vCPU is larger than that of other vCPUs; after determining the current resource occupancy information of the initial vCPU, it satisfies the requirement of the task to be allocated. In the case of task processing conditions, the initial vCPU is determined as the target vCPU; the to-be-allocated task is allocated to the target vCPU for task execution.

Specifically, the method can determine, from at least two vCPUs corresponding to the target virtual object, that the current resource occupancy information is greater than the initial vCPUs of other vCPUs in the at least two vCPUs; In the case of the task processing conditions of the task, the initial vCPU is determined as the target vCPU, and the to-be-allocated task is executed through the target vCPU, thereby avoiding the problem of CPU scheduling loss caused by the frequent trapping of idle vCPUs to the host. , which further reduces the physical CPU scheduling loss.

Description of drawings

1 is a flowchart of a task scheduling provided by an embodiment of the present specification;

2 is a schematic flowchart of a trapping operation provided by an embodiment of the present specification;

3 is a schematic diagram of a specific application scenario of a task processing method provided by an embodiment of this specification;

FIG. 4 is a process flow chart of a task processing method provided by an embodiment of the present specification;

5 is a flowchart of a task processing method provided by an embodiment of the present specification;

6 is a schematic structural diagram of a task processing apparatus provided by an embodiment of the present specification;

FIG. 7 is a structural block diagram of a computing device provided by an embodiment of the present specification.

Detailed ways

In the following description, numerous specific details are set forth in order to provide a thorough understanding of this specification. However, this specification can be implemented in many other ways different from those described herein, and those skilled in the art can make similar promotions without departing from the connotation of this specification. Therefore, this specification is not limited by the specific implementation disclosed below.

The terminology used in one or more embodiments of this specification is for the purpose of describing a particular embodiment only and is not intended to limit the one or more embodiments of this specification. As used in the specification or embodiments and the appended claims, the singular forms "a," "the," and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that the term "and/or" as used in this specification in one or more embodiments refers to and includes any and all possible combinations of one or more of the associated listed items.

It will be understood that although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited by these terms. These terms are only used to distinguish the same type of information from each other. For example, a first could be termed a second, and similarly, a second could be termed a first, without departing from the scope of one or more embodiments of this specification. Depending on the context, the word "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determining."

First, the terminology involved in one or more embodiments of the present specification is explained.

Hardware Virtualization: A technique that uses hardware instructions provided by a physical CPU to execute a virtual machine.

vCPU: The CPU that the virtual machine can perceive. Under hardware virtualization, the general vCPU implementation is a thread on the host, that is, the vCPU thread.

Guest: The virtual machine.

Host: The host where the virtual machine is running.

Trap; under hardware virtualization, the CPU exits from the state of executing the instructions/processes in the virtual machine (called non-root mode) to the state of executing the instructions/processes of the host machine (root mode).

Trap: In contrast to trap, it is from root mode into non-root mode.

HLT: that is, halt, when the CPU has nothing to do (no threads waiting to be scheduled), execute the HLT instruction of the CPU, indicating that the CPU is idle. Under hardware virtualization, the execution of HLT instructions by the vCPU in the guest will cause a trap to the host.

idle: The idle in scheduling means that there are no threads waiting to run on the CPU that can be scheduled, so it enters an idle state, that is, the idle state. In this case, Linux generally executes the HLT instruction.

Scheduling domain: Due to the different memory access rates caused by the Cache and NUMA shared by CPUs at different levels, the scheduler may be divided into different levels such as the same CPU core, the same CPU package, the same NUMA, and all CPUs. The CPUs in each level are called For a scheduling domain, for example, the scheduler generally tries to avoid scheduling threads to a different NUMA than the NUMA it was in before.

NUMA (Non Uniform Memory Access): Non-uniform memory access.

CPU core: The core of the CPU processor.

CPU package: Refers to a physical entire CPU, and a CPU package may have multiple CPU cores.

With the development of virtualization technology, many virtual machines, cloud servers (ECS) or secure containers will use hardware virtualization capabilities for security isolation, and under hardware virtualization, a source of CPU overhead that is difficult to eliminate is: After each vCPU thread is trapped in the host, it will re-execute the scheduling process on the host. The scheduling process is: the scheduler judges that if there are no other tasks waiting to be scheduled on the vCPU, then it determines that the vCPU is idle. state, and the host CPU where the vCPU is located has no other threads waiting to run on the host, an idlebalance process will be executed. If there are other tasks waiting to be run on the corresponding host CPU, although the idle balance will not be executed, the runtime update of the vCPU will also be processed during the scheduling process, which is also time-consuming. The idle balance process will try to pull the processes waiting to run from other vCPUs, which will occupy more obvious vCPU running resources and further consume the computing resources of the physical CPU running the vCPU thread. Brings high and hard-to-optimize virtualized CPU costs (ie, scheduling costs).

For example, in the scenario where tcp ping-pong is executed in a secure container or virtual machine to send and receive packets, the above-mentioned problem of causing CPU consumption is more obvious, because the vCPU corresponding to the secure container or virtual machine, every few hundred microseconds or a few In milliseconds, the operation of sending and receiving packets is performed once, and a small amount of code is executed each time and it is trapped; and in the process of each trapping, in the case where the vCPU has already executed the scheduling process in the virtual machine, it is necessary to An additional execution of the scheduling process on the host causes the scheduling overhead of the vCPU to become twice that of the physical machine.

One of the reasons for the above problem is that the scheduler of a security container or virtual machine usually selects the most idle CPU in a certain scheduling domain when selecting a vCPU to execute a task, so as to achieve balanced scheduling. See Figure 1 for the relevant process. 1 is a flowchart of a task scheduling provided by an embodiment of this specification, which specifically includes the following steps:

Step 102: There is a new thread to run.

Step 104: Find the most idle vCPU in a certain scheduling domain.

Step 106: Schedule the thread to run and execute on the most idle vCPU.

Based on the above steps, it can be seen that using the scheduler as an example to ensure load balancing, new threads will be allocated to the most idle vCPU, so that each vCPU has tasks (that is, threads) that need to be executed; however, this also leads to each The vCPU will perform the trapping operation after a short period of time, which will result in more HLT trapping processing overhead and scheduling overhead.

Specifically, refer to FIG. 2 for a flow chart of a trapping operation performed by a vCPU. FIG. 2 is a schematic flowchart of a trapping operation provided by an embodiment of this specification, which specifically includes the following steps:

Step 202: idle state.

Specifically, the scheduler of the host machine determines from the multiple vCPUs corresponding to the virtual machine that the current state and the idle state are the vCPUs.

Step 204: Enter the virtual machine.

Specifically, the scheduler of the host machine schedules the idle vCPU to the corresponding virtual machine, so that the virtual machine can perceive the vCPU.

Step 206: Execute the instruction in the virtual machine.

Specifically, after the idle vCPU is scheduled to the corresponding virtual machine, when the virtual machine receives a new instruction, the vCPU can execute the instruction, for example, the vCPU can execute the instruction of sending and receiving packets.

Step 208: the virtual machine kernel scheduling process.

Specifically, after receiving the instruction for the virtual machine, the scheduler in the virtual machine kernel starts the scheduling process in the virtual machine kernel, and schedules the instruction to the vCPU corresponding to the virtual machine for execution.

Step 210: No callable, HLT traps.

Specifically, the scheduler in the virtual machine kernel can determine to perform the HLT trap operation on the vCPU when it is determined that no execution can be scheduled to the vCPU.

Step 212: Execute HLT trapping processing.

Specifically, the scheduler in the virtual machine kernel performs HLT trapping processing on the vCPU that has no task (eg, instruction, thread, etc.) to be executed, and traps the vCPU to the host machine.

Step 214: Execute the host scheduling process.

Specifically, after the vCPU is trapped in the host, the scheduler of the host can release the CPU computing resources corresponding to the vCPU; allocate the other thread based on the released CPU computing resources corresponding to the vCPU; wherein the other thread may be The physical CPU implements threads of other vCPUs, or threads that perform tasks of the host itself.

Step 216: If there is no scheduling, enter the idle state.

Specifically, the scheduler in the virtual machine kernel can adjust the current state of the vCPU to an idle state after performing the HLT trapping process on the vCPU that has no task execution.

It should be noted that when the virtual machine corresponds to multiple vCPUs, each vCPU will perform the steps from step 202 to step 216 above. For example, if the virtual machine corresponds to vCPU1 and vCPU2, then both vCPU1 and vCPU2 will perform the above steps. 202 - the steps of step 216.

Based on the above steps, it can be known that in the case that the vCPU frequently performs the trapping operation, more HLT trapping processing overhead and scheduling overhead will be caused.

For the above problems, this manual provides four solutions, among which:

The first solution to this trapping overhead is to make the guest not perform the trapping operation when executing the HLT, so that the scheduling process on the host does not need to be executed every time the guest changes from work to idle.

However, this solution is usually applicable to the situation where the physical CPU on the host can be exclusively executed by the vCPU. If multiple vCPUs run on one physical CPU, other vCPUs will also interrupt the vCPU of the HLT in the guest when they want to execute, resulting in If the HLT in the guest is interrupted, the trapping operation will still be performed. Therefore, the effect of this scheme cannot achieve the expected effect when the vCPU cannot exclusively use the physical CPU.

The second solution is to optimize the scheduling process on the guest or host and reduce the CPU overhead required by these scheduling processes. However, since the scheduling process of Linux (an operating system) is relatively complex, it has been proved by practice that this optimization operation is difficult to implement. Therefore, there is no better optimization for the Linux scheduling process in the industry at present.

The third solution is to make a dedicated scheduler for the vCPU. However, this solution does not solve the problem of the above-mentioned large trapping overhead. Because, even if the vCPU is scheduled through a dedicated scheduler, after the vCPU is trapped, it is inevitable to schedule and select other vCPUs to be executed, or idle scheduling (pulling vCPUs waiting to be executed on other CPUs) process, That is to say, this solution still has a large scheduling overhead.

The fourth solution is: closing the idle balance in the scheduler can also reduce the CPU scheduling overhead, but this solution does not accumulate tasks running on a few CPUs as much as possible, so the effect of reducing the CPU scheduling overhead is relatively low. weak.

Based on this, in the task processing method provided in this specification, after receiving a task allocation request for a target virtual object and carrying a task to be allocated, firstly, based on the task allocation request, at least two vCPUs corresponding to the target virtual object are determined, and from at least two vCPUs corresponding to the target virtual object are determined. An initial vCPU is determined from the two vCPUs, wherein the current resource occupation information of the initial vCPU is greater than other vCPUs other than the initial vCPU among the at least two vCPUs.

Then, when the current resource occupancy information of the initial vCPU is determined and the task processing conditions of the to-be-allocated task are satisfied, the initial vCPU is determined as the target vCPU, and the to-be-allocated task is allocated to the target vCPU for task execution. The problem of CPU scheduling loss caused by frequent traps to the host machine further reduces the physical CPU scheduling loss.

Specifically, in this specification, a task processing method is provided, and this specification also relates to a task processing apparatus, a computing device, a computer-readable storage medium, and a computer program. In the following embodiments They are explained in detail one by one.

Fig. 3 shows a flowchart of a task processing method provided in an embodiment of the present specification in a specific application scenario, which specifically includes: a scheduler in the virtual machine kernel, when receiving a new thread A, determines There is a new thread A that needs to be allocated to the vCPU to run and execute. Based on this, the scheduler determines the scheduling domain B corresponding to the thread A from multiple scheduling domains; then the scheduler finds the busiest vCPU1 from the scheduling domain B, and the busiest vCPU1 can be understood as the scheduling domain The highest occupied vCPU in B. The scheduling domain B includes vCPU1, vCPU2, and vCPU3.

After that, if it is determined that the vCPU1 is too busy, find a second busy vCPU2 from the scheduling domain B; the second busy vCPU2 can be understood as the vCPU with the highest occupancy rate in the scheduling domain B except for vCPU1; that is, Yes, the vCPU with the second highest occupancy in this scheduling domain B.

In addition, it can be understood that the vCPU1 is too busy, and the occupancy rate of the vCPU1 reaches a preset occupancy rate threshold (for example, 70%) or more, then it is determined that the vCPU1 is too busy; that is, the vCPU1 is too busy. Unable to process new threads.

Finally, the scheduler in the virtual machine kernel schedules the new thread A to run and execute on the selected vCPU.

Based on the above content, in an embodiment provided in this specification, the scheduling content provided by the task processing method includes the following steps A1 to A4.

Step A1: There is a new thread to run.

Step A2: Find the busiest vCPU in a certain scheduling domain.

Step A3: If the vCPU is too busy, for example, the vCPU occupancy rate reaches above the threshold (70%), select a less busy vCPU.

Step A4: Schedule the thread to be executed on the selected vCPU.

Based on the above scheduling related content for FIG. 3 , the scheduling idea of the task processing method provided in this specification is to make the threads waiting for execution run on a few CPUs as much as possible to reduce HLT overhead. When there is a new task (thread) available to run, the scheduler of the virtual machine kernel selects the busiest CPU in a scheduling domain. If the CPU is too busy and exceeds a threshold (for example, the CPU usage exceeds 70%), a less busy CPU is reselected, and then the executable task is scheduled to be executed on the selected CPU. In this way, the threads waiting to be executed are concentrated to run on a few CPUs, so as to reduce the HLT overhead, avoid the problem of CPU scheduling loss caused by the frequent trapping of idle vCPUs to the host, and further reduce the physical CPU scheduling loss. .

Based on the above scheduling method, the scheduling strategy provided by the task processing method in this specification can make the threads waiting for execution run on a few CPUs as much as possible, and keep other vCPUs in an idle state, thereby reducing CPU scheduling loss. scheduling plan.

When the virtual machine corresponds to multiple vCPUs, the above scheduling policy will be executed on a few CPUs, while other vCPUs will execute the same steps as shown in Figure 2. For example, if the virtual machine corresponds to vCPU1 and vCPU2, the above scheduling policy will be executed on vCPU1, so that vCPU1 can select and execute more tasks in the guest, so vCPU1 is busy and less idle; therefore, resulting in less HLT for vCPU1 The scheduling process on the trap (idle) and the host; specifically, under the scheduling strategy provided by the task processing method provided in this specification, the HLT trapping and scheduling process for vCPU1 is shown in Figure 4, which is a diagram of this specification. A processing flow chart of a task processing method provided by an embodiment specifically includes the following steps.

For the specific process of vCPU performing the trapping operation, please refer to the following steps:

Step 402: idle state.

Specifically, the scheduler of the host machine determines from the multiple vCPUs corresponding to the virtual machine that the current state and the idle state are the vCPUs.

Step 404: Enter the virtual machine.

Specifically, the scheduler of the host machine schedules the idle vCPU to the corresponding virtual machine, so that the virtual machine can perceive the vCPU.

Step 406: Execute the thread in the virtual machine.

Specifically, after the idle vCPU is scheduled to the corresponding virtual machine, when the virtual machine receives a new thread, the virtual machine can execute the thread through the vCPU, for example, execute the thread for sending and receiving packets through the vCPU.

Step 408: the virtual machine kernel scheduling process.

Specifically, after receiving the thread for the virtual machine, the scheduler in the virtual machine kernel starts the scheduling process in the virtual machine kernel, and schedules the thread to the vCPU corresponding to the virtual machine for execution.

The scheduling process may be: when a new task (thread) can be run, the scheduler of the virtual machine kernel selects the most busy CPU in a scheduling domain. If the CPU is too busy and exceeds a threshold (for example, the CPU usage exceeds 70%), a less busy CPU is reselected, and then the executable task is scheduled to be executed on the selected CPU.

Step 410: The thread yields the CPU and calls other threads waiting for execution.

Specifically, through the above scheduling process, the threads waiting to be executed can be concentrated to run on a few CPUs as much as possible, so there are multiple threads waiting to be executed corresponding to the vCPU.

When the vCPU finishes executing a thread, the thread yields the vCPU; and the vCPU calls other threads waiting for execution.

Step 412: Execute other waiting threads.

Specifically, after the vCPU calls other threads waiting for execution, it will execute the other threads waiting for execution.

In addition, when the vCPU executes other waiting threads, the scheduler in the virtual machine kernel will continue to execute the scheduling process in the virtual machine kernel after receiving the thread for the virtual machine, and schedule the thread to the virtual machine kernel. Steps performed on the vCPU corresponding to the virtual machine.

Step 414: If there is no schedulable, enter the idle state.

Specifically, the scheduler in the virtual machine kernel can determine to perform the HLT trap operation on the vCPU when it is determined that no execution can be scheduled to the vCPU.

Step 416: Execute HTL trapping processing.

Specifically, the scheduler in the virtual machine kernel performs HLT trapping processing on the vCPU that has no task (eg, thread, thread, etc.) to be executed, and traps the vCPU to the host machine.

After the vCPU is trapped in the host, the scheduler of the host can release the CPU computing resources corresponding to the vCPU; allocate the other thread based on the released CPU computing resources corresponding to the vCPU; wherein, the other thread may be the physical Threads that implement other vCPUs on the CPU, or threads that execute the host's own tasks.

Moreover, referring to FIG. 4, it can be seen that the task processing method provided in this specification allows vCPU1 to execute the above steps 402 to 416, so that vCPU1 selects and executes more tasks in the guest; The same steps, so that when vCPU2 has fewer tasks to execute, it will not enter the guest and trap the guest regularly, but spend more time on the host, so the HLT of the idle vCPU2 is trapped and thus The resulting scheduling overhead on the host is also less. Further, the CPU overhead (part of the virtualization overhead) of each virtual machine will be very small.

Based on the above steps, it can be seen that the main difference between the task processing method provided in this specification and the general scheduling strategy is:

The main purpose of the general scheduling strategy is to balance the load of each CPU as much as possible, and the benefit is that the overall performance of the application is higher, but in the case of virtualization, each vCPU will often switch between work and idle, resulting in frequent HLT is trapped, which in turn leads to thousands of scheduling processes per second per vCPU on the host and CPU overhead.

Based on this, the task processing method provided in this specification changes the traditional scheduler's strategy of balancing the tasks of each CPU/vCPU as much as possible, and reduces the HLT under virtualization by stacking tasks on individual or few CPUs within a certain threshold. trapping, and the resulting scheduling overhead (CPU overhead) effect on the host is suitable for high-frequency trapping and short-lived scenarios, such as network ping-pong sending and receiving packets, or function computing (function compute) and other scenarios.

At the same time, for the fourth solution above, in the task processing method provided in this specification, since the HLT trapping of the vCPU is not turned off, there is no restriction that the vCPU needs to monopolize the CPU to be effective. Since the idle balance processing flow of the scheduler is not modified, the difficulty of optimizing the idle balance overhead is not involved. Since a scheduler for vCPUs is not specifically implemented, the difficulty of implementing a new scheduling policy for vCPUs is not involved. Since the above solution cannot accumulate running tasks on a few CPUs, the effect of reducing the CPU scheduling overhead is not as good as that achieved by the task processing method provided in this specification.

Based on this, the task processing method provided in this specification achieves the effect of reducing the CPU scheduling overhead under virtualization by making small changes to the existing scheduling strategy in selecting the CPU strategy, especially in some high-frequency wake-up scenarios. , such as the scenario of TCP ping-pong sending and receiving packets, and there is no obvious impact on performance in these scenarios.

Fig. 5 shows a flowchart of a task processing method provided according to an embodiment of the present specification, which specifically includes the following steps.

Step 502: Receive a task assignment request for the target virtual object, wherein the task assignment request carries the task to be assigned.

In practical applications, the task processing method provided in this specification can be applied to a scheduling module in a target virtual object, and the scheduling module can be understood as a module deployed in the target virtual object for allocating a vCPU to a task to be allocated; for example, The scheduling module may be a scheduler.

Wherein, the virtual unit can be understood as a virtual machine unit generated by hardware virtualization technology and used to perform tasks to be assigned. For example, the virtual unit can be understood as a virtual machine, a security container, a cloud server, etc., which is not described in this specification. Specific limitations; in the embodiments provided in this specification, the task processing method is described by taking the target virtual object as a virtual machine as an example.

The task allocation request may be understood as a request for instructing the scheduler to allocate a vCPU for the task to be allocated to execute the task.

The to-be-allocated task may be understood as a task that requires the vCPU to run and execute. For example, the to-be-allocated task includes but is not limited to an instruction to be executed, a thread to be executed, and the like. The instruction to be executed and the thread to be executed may be set according to actual application scenarios, which are not specifically limited in this specification. For example, the thread to be executed includes, but is not limited to, a thread that implements the ping-pong packet sending and receiving operation. The to-be-executed instruction may be an instruction that needs to be executed in the process of implementing the ping-pong packet sending and receiving operation, and the like.

Specifically, the scheduler in the target virtual object can receive a task allocation request for the target virtual object, where the task allocation request for the target virtual object can be understood as a task allocation request that the target virtual object needs to execute.

The task processing method is described below by taking the application of the task processing method provided in this specification in the scenario of allocating a packet sending and receiving thread to a vCPU as an example, wherein the target virtual object is a target virtual machine, and the task to be allocated is a packet sending and receiving thread , the task allocation request is a request for a packet sending and receiving thread to allocate an operation corresponding to the vCPU, and the packet sending and receiving thread may be a thread that implements a ping-pong sending and receiving operation of packets.

Based on this, the scheduler of the target virtual machine can receive a task allocation request for allocating the packet sending and receiving thread to the vCPU for execution, and the task allocation request carries the packet sending and receiving thread.

Step 504: Based on the task allocation request, determine at least two vCPUs corresponding to the target virtual object, and determine an initial vCPU from the at least two vCPUs, wherein the current resource occupancy information of the initial vCPU is larger than other vCPUs .

Wherein, the current resource occupancy information may be understood as information representing the current occupancy of computing resources of the vCPU, for example, the current resource occupancy information may be the occupancy rate of the vCPU. The initial vCPU may be understood as the vCPU with the highest resource usage among the at least two vCPUs. The other vCPU may be understood as other vCPUs other than the initial vCPU among the at least two vCPUs.

Specifically, after receiving the task allocation request, the scheduler can determine at least two vCPUs corresponding to the target virtual object based on the task allocation request, and then the scheduler determines current resource occupation information from the at least two vCPUs An initial vCPU larger than the other vCPUs.

Following the above example, after receiving a request for allocating a vCPU for running and executing a packet sending and receiving thread, the scheduler can determine at least two vCPUs corresponding to the virtual machine based on the request, for example, vCPU1, vCPU2, and vCPU3.

After that, the scheduler determines the occupancy rate of each vCPU, and determines the vCPU with the highest occupancy rate from vCPU1, vCPU2, and vCPU3; for example, the occupancy rate of vCPU1 is 65%, the occupancy rate of vCPU2 is 50%, and the occupancy rate of vCPU3 is 50%. 0%, the scheduler selects vCPU1 with the highest occupancy rate from vCPU1, vCPU2, and vCPU3.

In the embodiments provided in this specification, the scheduler can determine the initial vCPU from at least two vCPUs in a sorting manner, so as to ensure that the to-be-allocated task can be subsequently processed based on the vCPU with the highest current resource occupancy information, thereby avoiding idleness The problem of CPU scheduling loss caused by the frequent trapping of vCPUs to the host machine further reduces the physical CPU scheduling loss. Specifically, the determining the initial vCPU from the at least two vCPUs includes:

Determine the current resource occupancy information of each vCPU;

Sort each vCPU in descending order based on the current resource occupancy information, and obtain a descending sorting result of each vCPU;

The vCPU at the first position in the descending sorting result is determined as the initial vCPU.

Following the above example, the scheduler determines the occupancy rate of each vCPU, for example, the occupancy rate of vCPU1 is 75%, the occupancy rate of vCPU2 is 50%, and the occupancy rate of vCPU3 is 0%; Sort in descending order to obtain the descending sorting result corresponding to vCPU, such as vCPU1→vCPU2→vCPU3.

After that, the scheduler selects the vCPU1 in the first position from the descending sorting result as the vCPU with the highest occupancy rate.

In an embodiment provided in this specification, in order to save scheduling overhead, the task processing method provided in this specification will select the busiest vCPU from all vCPUs, and subsequently execute tasks based on the selected vCPU. At the same time, due to the different memory access rates caused by the Cache (cache) and NUMA shared by vCPUs at different levels, the scheduler may be divided into different levels such as the same CPU core, the same CPU package, the same NUMA, and all CPUs. The CPUs in each level is called a scheduling domain. Based on this, the unified task processing method provided in this specification can also take into account the running performance of the task. Through the scheduling domain selection strategy, the busiest CPU is selected from the CPUs of the same NUMA; try to avoid scheduling threads to the same NUMA as possible. Instead of going up to other NUMA, the thread is scheduled to the NUMA where it was before, so as to ensure the execution efficiency of the thread. The scheduling domain selection strategy can be any strategy that can select the busiest CPU from a group of CPUs. Which CPUs are selected specifically, that is, the scheduling domain selection strategy does not affect this solution, which is not specifically limited in this specification. . Among them, a scheduling domain selection strategy provided by this embodiment is as follows.

The determining an initial vCPU from the at least two vCPUs includes:

determining a vCPU set corresponding to each vCPU, wherein the vCPU set includes vCPUs of the same type;

Determine the target vCPU set corresponding to the task type of the to-be-allocated task based on the historical task execution information;

An initial vCPU is determined from the target vCPU set, wherein the current resource occupation information of the initial vCPU is greater than other vCPUs in the target vCPU set.

The vCPU set can be understood as a scheduling domain, and the same type of vCPUs can be understood as vCPUs sharing the same Cache and NUMA.

The historical task execution information may be understood as data generated by at least two vCPUs in the process of executing the historical task. The historical task execution information includes the task type corresponding to the historically executed task of each vCPU.

The task type can be understood as information representing the type of a task to be allocated. For example, in the case that the task to be allocated is a thread for sending and receiving packets, the task type can be a type for sending and receiving packets.

Specifically, the scheduler determines the vCPU set corresponding to each vCPU, and determines the historical task execution information generated by each vCPU during the historical task execution process; based on the historical task execution information, determines the target corresponding to the task type of the task to be allocated. vCPU set; and determining an initial vCPU from the target vCPU set, wherein the current resource occupation information of the initial vCPU is greater than other vCPUs in the target vCPU set. It should be noted that, for the explanation of the determination of the initial vCPU in the embodiment of this specification, reference may be made to the corresponding or corresponding content in the foregoing embodiment, and details are not described here in this embodiment.

Following the above example, the scheduler determines the scheduling domain to which each vCPU belongs, and determines the historical execution records generated in the process of executing threads according to the history of each vCPU; and the scheduling domain A to which the vCPU belongs; and the vCPU with the highest occupancy rate is selected from the scheduling domain A, that is, the busiest vCPU is selected from the scheduling domain A.

Step 506: In the case that the current resource occupancy information of the initial vCPU is determined and the task processing condition of the to-be-allocated task is satisfied, determine the initial vCPU as the target vCPU.

The target vCPU refers to the vCPU allocated to the task to be allocated and capable of performing task processing on the task to be allocated.

The task processing condition of the to-be-allocated task may be set according to the actual application scenario, which is not specifically limited in this specification. For example, the task processing condition of the to-be-allocated task may be that the initial vCPU needs to be an unloaded vCPU, and the initial vCPU The occupancy rate needs to be less than a preset threshold (eg 70%), etc.

Specifically, in an embodiment provided in this specification, the initial vCPU is determined as the target vCPU under the condition that the current resource occupation information of the initial vCPU is determined and the task processing condition of the to-be-allocated task is satisfied, include:

When the current resource occupation information of the initial vCPU is determined and the preset load condition is satisfied, the initial vCPU is determined as the target vCPU.

Wherein, the preset load condition can be understood as the occupancy rate of the initial vCPU is less than the preset occupancy rate threshold.

Following the above example, after determining the vCPU with the highest occupancy rate from the multiple vCPUs of the virtual machine, in order to ensure the smooth execution of the packet sending and receiving thread, the scheduler will determine whether the vCPU with the highest occupancy rate is under load. Specifically, When determining the vCPU with the highest occupancy rate, the scheduler determines whether the occupancy rate is less than the preset occupancy rate threshold (70%). If so, it is determined that the vCPU is not loaded and the packet sending and receiving thread can be successfully executed. Based on this, The vCPU with the highest occupancy rate is used as the vCPU that executes the packet sending and receiving thread. The preset occupancy rate threshold may be set according to an actual application scenario, which is not specifically limited in this specification.

Further, in the implementation provided in this specification, if the corresponding occupancy rate of the vCPU with the highest occupancy rate is greater than or equal to the preset occupancy rate threshold, it is determined that the vCPU is in a load state and cannot successfully execute the packet sending and receiving thread, then restart the process. Determine a new vCPU, and continue to judge whether the occupancy rate of the new vCPU is less than the preset occupancy rate threshold, until a vCPU with a occupancy rate less than the preset occupancy rate threshold and a relatively high occupancy rate is found, so as to ensure the pending tasks While executing smoothly, it can also avoid the problem of CPU scheduling loss caused by the frequent trapping of idle vCPUs to the host, and further reduce the physical CPU scheduling loss. Specifically, the determination from the at least two vCPUs After the initial vCPU, also includes:

In the case where the current resource occupancy information of the initial vCPU is determined and does not meet the preset load condition, a candidate vCPU is determined from the other vCPUs, wherein the current resource occupancy information of the candidate vCPU satisfies the preset load condition ;

A target vCPU is determined from the candidate vCPUs, wherein the current resource occupation information of the target vCPU is greater than other candidate vCPUs in the candidate vCPUs except the target vCPU.

The candidate vCPU may be understood as a vCPU whose occupancy rate is less than the preset occupancy rate threshold among other vCPUs.

Following the above example, when the scheduler determines the vCPU1 with the highest current occupancy rate, and determines that the occupancy rate (75%) of this vCPU1 is greater than or equal to the preset occupancy rate threshold (70%), in the load state, the scheduler can start from Among vCPU2 and vCPU3, the vCPU whose occupancy rate is less than the preset occupancy rate threshold is determined. Since the occupancy rate of vCPU2 is 50% and the occupancy rate of vCPU3 is 0%, the scheduler can select vCPU2 and vCPU3, and determine from vCPU2 and vCPU3. The vCPU with the highest utilization rate is vCPU2. After that, the scheduler uses the vCPU2 as the vCPU that executes the packet sending and receiving thread.

In the embodiments provided in this specification, in order to ensure the smooth execution of the to-be-allocated task, after the scheduler determines from at least two vCPUs that the current resource occupancy information is greater than the initial vCPUs of other vCPUs, it also needs to determine the remaining calculation of the initial vCPUs Whether the resource can run the to-be-allocated task smoothly, so as to ensure the smooth execution of the to-be-allocated task, specifically, when determining the current resource occupancy information of the initial vCPU and satisfying the task processing conditions of the to-be-allocated task , determine the initial vCPU as the target vCPU, including:

determining the resource occupation information of the to-be-allocated task;

Determine remaining resource information corresponding to the initial vCPU based on the current resource occupation information of the initial vCPU;

When the remaining resource information is greater than or equal to the resource occupation information, the initial vCPU is determined as the target vCPU.

The resource occupation information of the to-be-allocated task may be understood as computing resources required by the vCPU to run and execute the to-be-allocated task. For example, in the case of a thread for sending and receiving packets of a task to be allocated, the resource occupancy information may be understood as the occupancy rate required by the vCPU to execute the thread for sending and receiving packets.

The remaining resource information corresponding to the initial vCPU can be understood as the information representing the computing resources currently remaining in the initial vCPU and capable of executing the task to be allocated. For example, the remaining resource information can be the remaining resource occupancy rate; the occupancy rate corresponding to the initial vCPU In the case of 55%, the remaining resource occupancy rate of the initial vCPU can be 45%.

Specifically, after determining the initial vCPU from the at least two vCPUs, the scheduler can determine the resource occupancy information required to process the task to be allocated, and determine the remaining resources corresponding to the initial vCPU based on the current resource occupancy information of the initial vCPU information.

When it is determined that the remaining resource information is greater than or equal to the resource occupation information, it is determined that the initial vCPU can successfully execute the to-be-allocated task, and therefore, the initial vCPU is determined as the target vCPU.

Following the above example, after determining the vCPU with the highest occupancy rate, the scheduler can determine that the resource occupancy rate required to execute the packet sending and receiving thread is 30%, and determine the remaining computing resource remaining rate of the vCPU based on the current occupancy rate of the vCPU. is 50%, and when it is determined that the remaining computing resource rate of 35% is greater than or equal to the resource occupancy rate of 15%, the scheduler uses the vCPU with the highest current occupancy rate as the vCPU for processing the packet sending and receiving thread.

Further, in the embodiment provided in this specification, in the case that the residual rate of computing resources of the vCPU is less than the resource occupancy rate of the task to be allocated, in order to ensure the smooth execution of the task to be allocated, a vCPU will be re-selected, and the execution will continue. Comparing the residual rate of computing resources with the resource occupancy rate, until a vCPU whose residual rate of computing resources is greater than or equal to the resource occupancy rate is found, specifically, determining the residual rate corresponding to the initial vCPU based on the current resource occupancy information of the initial vCPU After the resource information, it also includes:

In the case that the remaining resource information is smaller than the resource occupation information, the step of determining an initial vCPU from the at least two vCPUs is continued until the remaining resource information is greater than or equal to the resource occupation information.

Wherein, for the explanation of continuing to determine the initial vCPU from at least two vCPUs in this embodiment, reference may be made to the corresponding or corresponding content in the foregoing embodiment, which will not be repeated in this specification.

Following the above example, the remaining computing resource rate of the vCPU1 is 25%, and the resource occupancy rate corresponding to the packet sending and receiving thread is 30%. The required resource occupancy ratio is compared to determine that the remaining computing resource ratio of 25% is less than the resource occupancy ratio of 30%. After that, the scheduler determines the vCPU with the highest occupancy ratio from vCPU2 and vCPU3, that is, vCPU2. After that, the scheduler determines that the remaining rate of computing resources of the vCPU2 is 50%, and compares the remaining rate of 50% of computing resources with the resource occupancy rate of 30% required for executing the packet sending and receiving thread, and determines that the remaining rate of computing resources of 50% is greater than the resource occupancy rate. The rate is 30%, so that the vCPU2 is determined as the vCPU that executes the packet sending and receiving thread.

Step 508: Allocate the to-be-allocated task to the target vCPU for task execution.

Following the above example, schedule this executable packet sending and receiving thread to the selected vCPU2 for execution.

In the embodiments provided in this specification, allocating the to-be-allocated task to the target vCPU for task execution includes:

Allocate the to-be-allocated task to a to-be-executed task queue corresponding to the target vCPU, and perform task execution in the to-be-executed task queue.

The to-be-executed task queue may be understood as a queue for storing to-be-allocated tasks to be executed by the target vCPU.

Following the above example, the vCPU in the virtual machine has a corresponding task queue, and the task queue stores the tasks to be executed that the vCPU needs to execute. Based on this, after determining the vCPU that executes the packet sending and receiving thread, the scheduler can send the packet sending and receiving thread to a task queue corresponding to the vCPU, and the vCPU executes the packet sending and receiving thread in the task queue.

In the embodiment provided in this specification, when the scheduler in the virtual machine determines that there is no other task waiting to be scheduled and running on the vCPU, it determines that the vCPU is in an idle state, and performs a trapping operation on the vCPU, thereby avoiding The vCPU without task execution wastes the computing resources of the physical CPU. Specifically, after allocating the to-be-allocated task to the target vCPU for task execution, the method further includes:

When it is determined that the queue of tasks to be executed corresponding to the target vCPU is empty, the current running state of the target vCPU is adjusted to an idle state, and the physical computing resources corresponding to the target vCPU are released.

The physical computing resources include, but are not limited to, physical vCPU computing resources corresponding to the target vCPU, physical storage resources corresponding to the target vCPU, and the like.

Specifically, when the scheduler determines that there is no to-be-allocated task waiting to be executed in the to-be-executed task queue, it adjusts the current running state of the target vCPU to an idle state; in practical applications, each task scheduling will be scheduled Therefore, when a task is scheduled by the scheduler, the scheduler can determine that the target vCPU currently has tasks waiting to be executed by making a judgment.

Afterwards, the scheduler releases the physical computing resources corresponding to the target vCPU, wherein the releasing of the physical computing resources corresponding to the target vCPU can be understood as performing a trapping operation on the target vCPU; for the explanation of performing a trapping operation on the vCPU, please refer to the above Corresponding or corresponding content in the embodiment, or any method for implementing the trapping operation on the vCPU may be referred to, and this embodiment will not describe it in detail in this embodiment.

Further, in an embodiment provided in this specification, the scheduling idea of the task processing method provided in this specification is to make the threads waiting for execution run on a few CPUs as much as possible, so that in addition to the above few CPUs, The other vCPUs perform fewer tasks, so that one of the other vCPUs is in an idle state, thereby further achieving the purpose of scheduling overhead. However, when the scheduler currently has many task allocation requests for the target virtual object, in order to ensure smooth processing of the task allocation requests and avoid the load on the target virtual object, the scheduler will put the idle state The vCPU is adjusted to the running state, and the task to be allocated carried in the task allocation request is executed by the vCPU, so as to avoid the problem of load on the target virtual object. Specifically, after receiving the task allocation request for the target virtual object, also include:

In the case of determining that the attribute information of the task allocation request satisfies the vCPU state adjustment condition, determine at least two vCPUs corresponding to the target virtual object;

Determining the vCPU whose current running state is in the idle state among the at least two vCPUs as the target vCPU;

The current running state of the target vCPU is adjusted to the running state, and the to-be-allocated task is allocated to the target vCPU for task execution.

The attribute information of the task allocation request may be understood as the quantity information of the task allocation request; correspondingly, if the attribute information satisfies the vCPU state adjustment condition, it may be understood that the quantity information of the task allocation request is greater than the preset quantity threshold; The number threshold can be set according to the actual application scenario, for example, 1000, 10000 and so on.

Wherein, adjusting the current running state of the target vCPU to the running state can be understood as performing a trapping operation on the target vCPU; and for the explanation of the trapping operation performed by the vCPU, refer to the corresponding or corresponding content in the above embodiment, or refer to any A way to implement the trapping operation on the vCPU, which is not described in detail in this specification.

Specifically, the scheduler can determine the attribute information of the task allocation request, and when it is determined that the attribute information of the task allocation request satisfies the vCPU state adjustment condition, the scheduler can determine at least two vCPUs corresponding to the target virtual object; and Among at least two vCPUs, a vCPU whose current running state is idle is determined as the target vCPU; then, the current running state of the target vCPU is adjusted to the running state, and the task to be allocated is allocated to the target vCPU for task execution.

Following the above example, the scheduler can determine the number of currently received task allocation requests. In practical applications, each task scheduling will go through the scheduler. Therefore, when a task is scheduled through the scheduler, the scheduler can determine Receives the number of task assignment requests received. Based on this, when the scheduler determines that the number of requests is greater than the preset number threshold, in order to avoid the problem of load on the virtual machine, the scheduler determines multiple vCPUs corresponding to the virtual machine, and determines from the multiple vCPUs that they are idle The vCPU in the state; the vCPU is determined as the vCPU that processes the packet sending and receiving thread carried in the request, and the trapping operation is performed on the vCPU, so that the vCPU can run the packet sending and receiving thread.

The task processing method provided in this specification can determine, from at least two vCPUs corresponding to the target virtual object, that the current resource occupancy information is greater than the initial vCPUs of other vCPUs in the at least two vCPUs; when determining the current resource occupancy information of the initial vCPU, When the task processing conditions of the task to be allocated are met, the initial vCPU is determined as the target vCPU, and the target vCPU is used to execute the to-be-allocated task, thereby avoiding CPU scheduling caused by frequent trapping of idle vCPUs to the host. The problem of loss further reduces the physical CPU scheduling loss.

Further, in the task processing method provided in this specification, the scheduler on the Host can also select the scheduling policy provided by the task processing method, and the scheduler on the Host can also achieve the effect of reducing the CPU scheduling overhead based on the scheduling policy, but Since the scheduler on the Host does not involve the two-layer scheduling of the guest and the host and the traps brought by virtualization, the effect of reducing the CPU scheduling overhead is not as obvious as that under virtualization.

Corresponding to the foregoing method embodiments, the present specification also provides an embodiment of a task processing apparatus, and FIG. 6 shows a schematic structural diagram of a task processing apparatus provided by an embodiment of the present specification. As shown in Figure 6, the device includes:

A receiving module 602, configured to receive a task assignment request for the target virtual object, wherein the task assignment request carries a task to be assigned;

The first determining module 604 is configured to, based on the task allocation request, determine at least two vCPUs corresponding to the target virtual object, and determine an initial vCPU from the at least two vCPUs, wherein the current value of the initial vCPU is The resource occupancy information is larger than other vCPUs;

The second determination module 606 is configured to determine the initial vCPU as the target vCPU under the condition that the current resource occupation information of the initial vCPU is determined and the task processing condition of the to-be-allocated task is satisfied;

The execution module 608 is configured to allocate the to-be-allocated task to the target vCPU for task execution.

Optionally, the second determining module 606 is further configured to:

When the current resource occupation information of the initial vCPU is determined and the preset load condition is satisfied, the initial vCPU is determined as the target vCPU.

Optionally, the task processing apparatus further includes a third determination module configured to:

In the case where the current resource occupancy information of the initial vCPU is determined and does not meet the preset load condition, a candidate vCPU is determined from the other vCPUs, wherein the current resource occupancy information of the candidate vCPU satisfies the preset load condition ;

A target vCPU is determined from the candidate vCPUs, wherein the current resource occupation information of the target vCPU is greater than other candidate vCPUs in the candidate vCPUs except the target vCPU.

Optionally, the first determining module 604 is further configured to:

Determine a vCPU set corresponding to each vCPU, wherein the vCPU set includes vCPUs of the same type; based on historical task execution information, determine a target vCPU set corresponding to the task type of the to-be-allocated task;

An initial vCPU is determined from the target vCPU set, wherein the current resource occupation information of the initial vCPU is greater than other vCPUs in the target vCPU set.

Optionally, the first determining module 604 is further configured to:

Determine the current resource occupancy information of each vCPU;

Sort each vCPU in descending order based on the current resource occupancy information, and obtain a descending sorting result of each vCPU;

The vCPU at the first position in the descending sorting result is determined as the initial vCPU.

Optionally, the second determining module 606 is further configured to:

determining the resource occupation information of the to-be-allocated task;

Determine remaining resource information corresponding to the initial vCPU based on the current resource occupation information of the initial vCPU;

When the remaining resource information is greater than or equal to the resource occupation information, the initial vCPU is determined as the target vCPU.

Optionally, the task processing apparatus further includes a fourth determination module configured to:

In the case that the remaining resource information is smaller than the resource occupation information, the step of determining an initial vCPU from the at least two vCPUs is continued until the remaining resource information is greater than or equal to the resource occupation information.

Optionally, the execution module 608 is further configured to:

Allocate the to-be-allocated task to a to-be-executed task queue corresponding to the target vCPU, and perform task execution in the to-be-executed task queue.

Optionally, the task processing apparatus further includes a first state adjustment module configured to:

When it is determined that the queue of tasks to be executed corresponding to the target vCPU is empty, the current running state of the target vCPU is adjusted to an idle state, and the physical computing resources corresponding to the target vCPU are released.

Optionally, the task processing apparatus further includes a second state adjustment module configured to:

In the case of determining that the attribute information of the task allocation request satisfies the vCPU state adjustment condition, determine at least two vCPUs corresponding to the target virtual object;

Determining the vCPU whose current running state is in the idle state among the at least two vCPUs as the target vCPU;

The current running state of the target vCPU is adjusted to the running state, and the to-be-allocated task is allocated to the target vCPU for task execution.

The task processing apparatus provided in this specification can determine, from at least two vCPUs corresponding to the target virtual object, that the current resource occupancy information is greater than the initial vCPUs of other vCPUs in the at least two vCPUs; when determining the current resource occupancy information of the initial vCPU, When the task processing conditions of the task to be allocated are met, the initial vCPU is determined as the target vCPU, and the target vCPU is used to execute the to-be-allocated task, thereby avoiding CPU scheduling caused by frequent trapping of idle vCPUs to the host. The problem of loss further reduces the physical CPU scheduling loss.

The above is a schematic solution of a task processing apparatus according to this embodiment. It should be noted that the technical solution of the task processing device and the technical solution of the above-mentioned task processing method belong to the same concept. For details that are not described in detail in the technical solution of the task processing device, please refer to the description of the technical solution of the above-mentioned task processing method. .

FIG. 7 shows a structural block diagram of a computing device 700 provided according to an embodiment of the present specification. Components of the computing device 700 include, but are not limited to, memory 710 and processor 720 . The processor 720 is connected with the memory 710 through the bus 730, and the database 750 is used for storing data.

Computing device 700 also includes access device 740 that enables computing device 700 to communicate via one or more networks 760 . Examples of such networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. Access device 740 may include one or more of any type of network interface (eg, network interface card (NIC)), wired or wireless, such as IEEE 802.11 wireless local area network (WLAN) wireless interface, World Interoperability for Microwave Access ( Wi-MAX) interface, Ethernet interface, Universal Serial Bus (USB) interface, cellular network interface, Bluetooth interface, Near Field Communication (NFC) interface, and the like.

In one embodiment of the present specification, the above-described components of computing device 700 and other components not shown in FIG. 7 may also be connected to each other, such as through a bus. It should be understood that the structural block diagram of the computing device shown in FIG. 7 is only for the purpose of example, rather than limiting the scope of the present specification. Those skilled in the art can add or replace other components as required.

Computing device 700 may be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (eg, tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (eg, smart phones) ), wearable computing devices (eg, smart watches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or PCs. Computing device 700 may also be a mobile or stationary server.

The processor 720 is configured to execute the following computer-executable instructions, and when the computer-executable instructions are executed by the processor 720, implement the steps of the above task processing method.

The above is a schematic solution of a computing device according to this embodiment. It should be noted that the technical solution of the computing device and the technical solution of the above-mentioned task processing method belong to the same concept. For details not described in detail in the technical solution of the computing device, refer to the description of the technical solution of the above-mentioned task processing method.

An embodiment of the present specification further provides a computer-readable storage medium, which stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the steps of the above task processing method.

The above is a schematic solution of a computer-readable storage medium of this embodiment. It should be noted that the technical solution of the storage medium and the technical solution of the above-mentioned task processing method belong to the same concept, and the details not described in detail in the technical solution of the storage medium can be referred to the description of the technical solution of the above-mentioned task processing method.

An embodiment of the present specification further provides a computer program, wherein, when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned task processing method.

The above is a schematic solution of a computer program of this embodiment. It should be noted that the technical solution of the computer program and the technical solution of the above-mentioned task processing method belong to the same concept. For details that are not described in detail in the technical solution of the computer program, refer to the description of the technical solution of the above-mentioned task processing method.

The foregoing describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that in the embodiments and still achieve desirable results. Additionally, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

The computer instructions include computer program code, which may be in source code form, object code form, an executable file, some intermediate form, or the like. The computer-readable medium may include: 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, Read-Only Memory) , Random Access Memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable media may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable media Electric carrier signals and telecommunication signals are not included.

It should be noted that, for the convenience of description, the foregoing method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the embodiments of this specification are not limited by the described action sequences. Limitation, because certain steps may be performed in other orders or simultaneously according to embodiments of the present specification. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily all necessary for the embodiments of the specification.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

The preferred embodiments of the present specification disclosed above are provided only to aid in the elaboration of the present specification. Alternative embodiments are not intended to exhaust all details, nor do they limit the invention to only the described embodiments. Obviously, many modifications and changes can be made in accordance with the contents of the embodiments of the present specification. These embodiments are selected and described in this specification to better explain the principles and practical applications of the embodiments of this specification, so that those skilled in the art can well understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims (13)

1. A method of task processing, comprising:

receiving a task allocation request aiming at a target virtual object, wherein the task allocation request carries a task to be allocated;

determining at least two vCPUs corresponding to the target virtual object based on the task allocation request, and determining an initial vCPU from the at least two vCPUs, wherein the current resource occupation information of the initial vCPU is greater than that of other vCPUs;

under the condition that the current resource occupation information of the initial vCPU is determined to meet the task processing condition of the task to be distributed, determining the initial vCPU as a target vCPU;

and distributing the task to be distributed to the target vCPU for task execution.

2. The task processing method according to claim 1, wherein the determining that the initial vCPU is the target vCPU when determining that the current resource occupation information of the initial vCPU satisfies the task processing condition of the task to be allocated includes:

and under the condition that the current resource occupation information of the initial vCPU is determined to meet the preset load condition, determining the initial vCPU as a target vCPU.

3. The task processing method according to claim 1, further comprising, after determining an initial vCPU from the at least two vcpus:

determining candidate vCPUs from other vCPUs under the condition that the current resource occupation information of the initial vCPU is determined and does not meet the preset load condition, wherein the current resource occupation information of the candidate vCPUs meets the preset load condition;

and determining a target vCPU from the candidate vCPUs, wherein the current resource occupation information of the target vCPU is larger than other candidate vCPUs except the target vCPU in the candidate vCPUs.

4. The task processing method according to claim 1, wherein the determining an initial vCPU from the at least two vcpus comprises:

determining a vCPU set corresponding to each vCPU, wherein the vCPU sets comprise the vCPUs of the same type;

determining a target vCPU set corresponding to the task type of the task to be distributed based on historical task execution information;

and determining an initial vCPU from the target vCPU set, wherein the current resource occupation information of the initial vCPU is larger than other vCPUs in the target vCPU set.

5. The task processing method according to claim 1, wherein the determining an initial vCPU from the at least two vcpus comprises:

determining current resource occupation information of each vCPU;

sequencing each vCPU in a descending order based on the current resource occupation information to obtain a sequencing result of each vCPU in the descending order;

and determining the vCPU at the first position in the descending sorting result as an initial vCPU.

6. The task processing method according to claim 1, wherein the determining that the initial vCPU is a target vCPU when determining that the current resource occupation information of the initial vCPU satisfies the task processing condition of the task to be allocated includes:

determining resource occupation information of the tasks to be distributed;

determining residual resource information corresponding to the initial vCPU based on the current resource occupation information of the initial vCPU;

and determining the initial vCPU as a target vCPU under the condition that the residual resource information is greater than or equal to the resource occupation information.

7. The task processing method according to claim 6, after determining the remaining resource information corresponding to the initial vCPU based on the current resource occupancy information of the initial vCPU, further comprising:

and under the condition that the residual resource information is smaller than the resource occupation information, continuously executing the step of determining the initial vCPU from the at least two vCPUs until the residual resource information is larger than or equal to the resource occupation information.

8. The task processing method according to claim 1, wherein the allocating the task to be allocated to the target vCPU for task execution includes:

and distributing the tasks to be distributed to the task queues to be executed corresponding to the target vCPU, and executing the tasks in the task queues to be executed.

9. The task processing method according to claim 7, further comprising, after the task is executed by allocating the task to be allocated to the target vCPU:

and under the condition that the task queue to be executed corresponding to the target vCPU is determined to be empty, adjusting the current running state of the target vCPU to be an idle state, and releasing the physical computing resource corresponding to the target vCPU.

10. The task processing method according to claim 1, further comprising, after receiving the task allocation request for the target virtual object:

under the condition that the attribute information of the task allocation request is determined to meet the vCPU state adjustment condition, determining at least two vCPUs corresponding to the target virtual object;

determining the vCPU of which the current operation state is an idle state in the at least two vCPUs as a target vCPU;

and adjusting the current running state of the target vCPU to be a running state, and distributing the task to be distributed to the target vCPU for task execution.

11. A task processing device comprising:

the system comprises a receiving module, a processing module and a processing module, wherein the receiving module is configured to receive a task allocation request aiming at a target virtual object, and the task allocation request carries a task to be allocated;

a first determining module configured to determine, based on the task allocation request, at least two vcpus corresponding to the target virtual object, and determine an initial vCPU from the at least two vcpus, where current resource occupation information of the initial vCPU is greater than that of other vcpus;

the second determining module is configured to determine that the initial vCPU is a target vCPU under the condition that the current resource occupation information of the initial vCPU is determined to meet the task processing condition of the task to be distributed;

and the execution module is configured to allocate the task to be allocated to the target vCPU for task execution.

12. A computing device, comprising:

a memory and a processor;

the memory is configured to store computer-executable instructions, and the processor is configured to execute the computer-executable instructions, which when executed by the processor implement the steps of the task processing method of any of claims 1 to 10.

13. A computer-readable storage medium storing computer-executable instructions which, when executed by a processor, implement the steps of the task processing method of any one of claims 1 to 10.

Images