Fault detection method and device

Abstract

The application provides a fault detection method, comprising: acquiring policy information of a resource group, wherein the resource group comprises a plurality of computing nodes; generating a detection policy according to the policy information, wherein the detection policy is used to instruct to generate a detection virtual machine in the plurality of computing nodes, and to make the detection virtual machine migrate between the plurality of computing nodes to detect whether any computing node in the plurality of computing nodes has a live migration capability, wherein the detection virtual machine is only used to detect whether the plurality of computing nodes have the live migration capability, and does not undertake other service processing work; receiving a detection result reported by the detection virtual machine; and determining whether any computing node in the plurality of computing nodes has the live migration capability according to the detection result. The above scheme can greatly improve the success rate of live migration of the virtual machine.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of cloud computing, and in particular to a fault detection method and device. BACKGROUND

[0002] Live Migration, also called dynamic migration or real-time migration, is a virtual machine save / resume process, which can save the running state of the entire virtual machine completely and quickly restore it to the original hardware platform or even a different hardware platform. After the restoration, the virtual machine still runs smoothly and users cannot perceive any difference.

[0003] However, the efficiency of the virtual machine live migration in the prior art is not high. SUMMARY

[0004] To solve the above problems, the present application provides a fault detection method and device, which can effectively improve the success rate of the live migration of the virtual machine.

[0005] In a first aspect, a fault detection method is provided, comprising:

[0006] obtaining policy information of a resource group, wherein the resource group comprises a plurality of computing nodes;

[0007] generating a detection policy according to the policy information, wherein the detection policy is used to instruct to generate a detection virtual machine in the plurality of computing nodes, and to make the detection virtual machine perform live migration between the plurality of computing nodes to detect whether any computing node in the plurality of computing nodes has live migration capability, the detection virtual machine is only used to detect whether the plurality of computing nodes have live migration capability and does not undertake other business processing work; receiving a detection result reported by the detection virtual machine; and determining whether any computing node in the plurality of computing nodes has live migration capability according to the detection result.

[0008] In some possible designs, the detection virtual machine meets one or more of the following conditions: the detection virtual machine occupies less computing resources than a normal virtual machine, the detection virtual machine occupies less storage resources than a normal virtual machine, and the detection virtual machine occupies less network resources than a normal virtual machine.

[0009] In some possible designs, the policy information comprises one or more of user demand, resource size of each computing node in the plurality of computing nodes, and resource occupation status of each computing node in the plurality of computing nodes.

[0010] In some possible design, the content of the detection strategy can include one or more of the number of detection virtual machines, the specification of the detection virtual machines, the detection period, and the detection route.

[0011] In a second aspect, a physical host is provided, including an obtaining module and a generating module,

[0012] The obtaining module is configured to obtain policy information of a resource group, where the resource group includes a plurality of computing nodes.

[0013] The generating module is configured to generate a detection strategy according to the policy information, where the detection strategy is used to instruct to generate a detection virtual machine in the plurality of computing nodes, and to make the detection virtual machine migrate between the plurality of computing nodes to detect whether any computing node in the plurality of computing nodes has a live migration capability, the detection virtual machine is only used to detect whether the plurality of computing nodes has a live migration capability, and does not undertake other business processing work; the receiving module is configured to receive a detection result reported by the detection virtual machine; and the determining module is configured to determine whether any computing node in the plurality of computing nodes has a live migration capability according to the detection result.

[0014] In some possible design, the detection virtual machine meets one or more of the following: the detection virtual machine occupies less computing resources than a normal virtual machine, the detection virtual machine occupies less storage resources than a normal virtual machine, and the detection virtual machine occupies less network resources than a normal virtual machine.

[0015] In some possible design, the policy information includes one or more of a user demand, a resource size of each computing node in the plurality of computing nodes, and a resource occupation status of each computing node in the plurality of computing nodes.

[0016] In some possible design, the content of the detection strategy can include one or more of the number of detection virtual machines, the specification of the detection virtual machines, the detection period, and the detection route.

[0017] In a third aspect, a computing node is provided, including a memory and a processor, where the processor executes a program in the memory to run a computing service and a storage service, so as to execute the method in any one of the first aspect.

[0018] In a fourth aspect, a readable storage medium is provided, including instructions, when the instructions are run on a computing node, the computing node executes the method in any one of the first aspect.

[0019] In a fifth aspect, a computer program product is provided, which, when read and executed by a computer, causes the method according to any one of the first aspect to be performed.

[0020] In the above solution, before performing the heavy virtual machine live migration, a light virtual machine is used to detect whether the computing node has the live migration capability, and in the case that it is determined that the light virtual machine has the live migration capability, the heavy virtual machine is live migrated, thereby improving the success rate of the virtual machine live migration. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background art, the drawings needed to be used in the embodiments of the present application or the background art will be described below.

[0022] Figure 1 is a structural schematic diagram of a cloud system related to the present application;

[0023] Figure 2 is a schematic diagram of a specific cloud system related to the present application;

[0024] Figure 3 is a structural schematic diagram of a computing node provided by the present application;

[0025] Figure 4 is a schematic diagram of another specific cloud system related to the present application;

[0026] Figure 5 is a schematic diagram of another specific cloud system related to the present application;

[0027] Figure 6 is a structural schematic diagram of a physical host provided by the present application. DETAILED DESCRIPTION

[0028] Reference is made to Figure 1 , Figure 1This is a schematic diagram of the structure of a cloud system involved in this application. The owner of the cloud system deploys its own cloud computing infrastructure, namely, deploying computing resources (e.g., servers) 110, storage resources (e.g., memory) 120, and network resources (e.g., network interface cards) 130, etc. Then, the public cloud owner (e.g., a telecom operator) virtualizes the computing, storage, and network resources of the cloud computing infrastructure and provides corresponding computing, storage, and network services to cloud users (e.g., users). Among them, the operator can provide the following three services to users: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS).

[0029] IaaS provides users with access to cloud computing infrastructure, including processing, storage, networking, and other basic computing resources. Users can deploy and run any software, including operating systems and applications. Users do not manage or control any cloud computing infrastructure, but they can control the choice of operating system, storage space, application deployment, and may also have limited control over network components (e.g., firewalls, load balancers, etc.).

[0030] PaaS provides users with the service of deploying applications developed by users using vendor-provided programming languages ​​and tools (such as Java, Python, .NET, etc.) or acquired to cloud computing infrastructure. Users do not need to manage or control the underlying cloud computing infrastructure, including networks, servers, operating systems, storage, etc., but users can control the deployed applications and may also control the configuration of the hosting environment running the applications.

[0031] SaaS provides users with applications that run on cloud computing infrastructure. Users can access these applications on various devices through a client interface, such as a browser. Users do not need to manage or control any cloud computing infrastructure, including networks, servers, operating systems, storage, etc.

[0032] Users can utilize the services provided by the operator to run their own businesses, such as deep learning, artificial intelligence, and big data.

[0033] See Figure 2 , Figure 2 This is a schematic diagram of a specific cloud system involved in this application. The cloud system of this application includes: a physical host 210 and multiple resource groups 220.

[0034] Resource group 220 includes multiple compute nodes 221, and each compute node 221 can be virtualized into one or more virtual machines 222 using virtualization technology. For example... Figure 3 As shown, compute node 221 includes hardware resources 310, hypervisor 320, virtual machines (VMs) 330, operating system 340, and application software (APP) 350. Hardware resources 310 are the basic physical hardware in the compute node, which may include processors, memory, disks, and networks, etc. Virtual software 320 runs as an intermediate software layer between the hardware resources and the virtual machines. Examples include VMware vSphere, Hyper-V, Citrix XenServer, IBM PowerVM, Red Hat Enterprise Virtualization, Huawei FusionSphere, KVM, Xen, VirtualBSD, etc., allowing multiple virtual machines to share a single set of basic physical hardware. Virtual machine 330 refers to a complete computer system simulated by software, possessing full hardware system functionality and running in a completely isolated environment. An operating system (OS) 340 is a computer program that manages and controls computer hardware and software resources. It serves as the interface between the user and the computer, and also as the interface between computer hardware and other software. Examples include SUN Solaris, IBM-AIX, HP-UX, FreeBSD, Red Hat Linux, CentOS, Debian, Ubuntu, Windows Server 2003, Windows Server 2008, and Windows Server 2008R2. Application software 350 is a collection of various programming languages ​​available to users and applications written in those languages, such as office software, ICQ software, and browser software. Virtualization software 320 can employ a hypervisor, which coordinates access to all the underlying physical hardware and virtual machines on a compute node. When the compute node starts and executes the hypervisor, it allocates appropriate hardware resources, such as processors, memory, disks, and networks, to each virtual machine and loads the operating system for all virtual machines.

[0035] Physical host 210 includes a cloud management platform 211. The cloud management platform 211 can be AWS (Amazon Web Services), OpenStack, CloudStack, Hadoop, Apache Mesos, Docker-based Kubernetes, Swarm, etc., without specific limitations. The cloud management platform 211 can include a computing service 224. The computing service 224 can be a computing organization controller within the cloud management platform 211, a component used for deploying virtual machines, etc. Optionally, the cloud management platform 211 may also include storage services and network services, etc. (not shown in the figure). The storage service can be a component within the cloud management platform used to provide distributed, persistent virtual object storage, etc. The network service can be a component within the cloud management platform 211 used to provide network topology management for network nodes. Taking OpenStack as an example, the computing service 224 can be Nova, the storage service can be Swift, and the network service can be Neutron. In addition, physical host 210 also includes an operating system 225 and hardware 226. The hardware 226 can include a physical network interface card 227.

[0036] Virtual machines on compute nodes 221 within the same resource group 220 can be hot-migrated to each other with the help of compute service 224. However, virtual machines on compute nodes 221 in different resource groups 220 cannot be hot-migrated to each other.

[0037] Continue with Figure 2 Taking the example shown, assume there are resource group 1 and resource group 2. Resource group 1 includes compute node 1, compute node 2 and compute node 3. Compute node 1 is equipped with virtual machine 1, compute node 2 is equipped with virtual machine 2 and compute node 3 is equipped with virtual machine 3. Resource group 2 includes compute node 3, compute node 4 and compute node 5. Compute node 3 is equipped with virtual machine 3, compute node 4 is equipped with virtual machine 4 and compute node 5 is equipped with virtual machine 5.

[0038] In resource group 1, virtual machines 1, 2, and 3 can freely migrate between compute nodes 1, 2, and 3. For example, virtual machine 1 can migrate from compute node 1 to compute node 2, virtual machine 2 can migrate from compute node 2 to compute node 3, and virtual machine 3 can migrate from compute node 3 to compute node 1. It is understood that, besides the above example, virtual machines 1 to 3 can also use other migration methods, which will not be elaborated here.

[0039] In resource group 2, virtual machines 4, 5, and 6 can freely migrate between compute nodes 4, 5, and 6. For example, virtual machine 4 can migrate from compute node 4 to compute node 5, virtual machine 5 can migrate from compute node 5 to compute node 6, and virtual machine 6 can migrate from compute node 6 to compute node 4. It is understood that, besides the above example, virtual machines 4 and 6 can also use other migration methods, which will not be elaborated here.

[0040] However, virtual machines 1, 2, and 3 cannot be migrated to compute nodes 4, 5, and 6, and virtual machines 4, 5, and 6 cannot be migrated to compute nodes 1, 2, and 3.

[0041] As cloud systems continue to expand in scale and various resource types increase, the complexity of environment configuration increases dramatically, and the hardware failure rate also continues to rise. As a result, the success rate of hot migration of virtual machines is not high, which affects the stability of upper-layer services and greatly impacts the user experience.

[0042] To address the aforementioned issues, this application presents a schematic diagram of another specific cloud system. For example... Figure 4 As shown, the cloud system of this application includes: a physical host 410 and multiple resource groups 420.

[0043] Resource group 420 includes multiple compute nodes 421, and each compute node 421 can be virtualized into one or more virtual machines 422 using virtualization technology. Specifically, Figure 4 The computing node 421 can be found in [reference needed]. Figure 3 The computation nodes shown and their related descriptions will not be repeated here.

[0044] Physical host 410 includes a cloud management platform 411. The cloud management platform 411 can be AWS (Amazon Web Services), OpenStack, CloudStack, Hadoop, Apache Mesos, Docker-based Kubernetes, Swarm, etc., without specific limitations. The cloud management platform 411 can include virtual machine hot migration detection 422, operation and maintenance service 423, and computing service 424. The virtual machine hot migration detection 422 is used to determine whether the computing nodes 421 in resource group 420 have the ability to hot migrate virtual machines and automatically isolate faulty nodes (i.e., computing nodes that do not have hot migration capabilities), and report the faulty nodes to the operation and maintenance service 423. The operation and maintenance service 423 is used to receive the faulty nodes reported by the virtual machine hot migration detection 422, thereby realizing the control of business operation and maintenance needs, processes, and results, as well as the control of the overall system operating status, thereby realizing the recording and tracking of application changes, the tracking of changes based on different regions, the prediction and prevention of environmental and application risks, etc. The operation and maintenance service 423 can be Service CM, Alibaba Cloud Tianji, AWS Apollo, etc. Computing service 424 may be a computing organization controller in cloud management platform 411, a component used for deploying virtual machines, etc. Optionally, cloud management platform 411 may also include storage services and network services, etc. (not shown in the figure). Storage services may be components in the cloud management platform used to provide distributed, persistent virtual object storage, etc. Network services may be components in cloud management platform 411 used to provide network topology management for network nodes. Taking OpenStack as an example, computing service 424 may be Nova, storage services may be Swift, and network services may be Neutron. Furthermore, physical host 410 also includes operating system 425 and hardware 426. Hardware 426 may include physical network interface card 427.

[0045] In one specific implementation, such as Figure 5As shown, the virtual machine hot migration detection 422 may include a resource scheduling module 4221, a migration execution module 4222, and a fault handling module 4223. The resource scheduling module 4221 receives user requests input through the interface, obtains the resource size and resource occupancy status of each compute node in each resource group through the compute service 424, and determines a detection strategy based on the user requests, the resource size, and the resource occupancy status of each compute node in each resource group. The migration execution module 4222 calls the compute service 424 to generate a probe virtual machine within each compute node according to the detection strategy, performs hot migration of the probe virtual machine, and records the results of the hot migration. The results of the hot migration include whether the probe virtual machine successfully migrated from one compute node to another, and the duration required for the hot migration, etc. The fault handling module 4223 analyzes the faulty nodes based on the hot migration results, reports the faulty nodes to the operations and maintenance department, and automatically isolates the faulty nodes through the compute service 424.

[0046] In one specific implementation, the compute node 421 in resource group 420 has the ability to hot-migrate virtual machines under the following two conditions: First, the environment configuration of the compute node to which the virtual machine is to be migrated is the same as the environment configuration of the compute node from which the virtual machine is to be migrated; second, neither the compute node to which the virtual machine is to be migrated nor the compute node from which the virtual machine is to be migrated has experienced a failure. The environment configuration of the compute node includes: system resource isolation configuration, system kernel parameter configuration, compute node migration domain configuration, physical network configuration, etc. Failures occurring on the compute node include: motherboard failure, CPU failure, memory failure, hard disk failure, network card failure, etc.

[0047] In one specific implementation, the virtual machine hot migration detector 422 can generate multiple probe virtual machines within the resource group according to the detection strategy, and enable the multiple probe virtual machines to migrate between computing nodes within the resource group.

[0048] The detection strategy can include the number, specifications, detection cycle, detection route, etc., of the virtual machines being detected. The number of virtual machines within a resource group can be one or more. The number of virtual machines being detected in different resource groups can be the same or different. The specifications of the virtual machine refer to the amount of computing, storage, and network resources it consumes. Higher specifications mean more computing, storage, and network resources are consumed, and vice versa. Virtual machines within the same resource group can have the same or different specifications, and the specifications of virtual machines in different resource groups can also be the same or different. The detection cycle for virtual machines can be in the range of seconds, minutes, hours, etc., without specific limitations. The detection cycle for virtual machines within the same resource group can be the same or different, and the detection cycle for virtual machines in different resource groups can also be the same or different. The probe path can be the route taken to detect the hot migration of virtual machines within a resource group's compute nodes. For example, the probe path for virtual machine 1 within the resource group could be: detect virtual machine 1 migrating from compute node 1 to compute node 2, then from compute node 2 to compute node 3, and finally from compute node 3 back to compute node 1. The compute nodes traversed by the probe path can be repeated or non-repeating. For example, a non-repeating probe path could be: detect virtual machine 1 migrating from compute node 1 to compute node 2, from compute node 2 to compute node 3, and from compute node 3 back to compute node 1. A repeated probe path could be: detect virtual machine 1 migrating from compute node 1 to compute node 2, from compute node 2 to compute node 3, from compute node 3 back to compute node 2, and from compute node 2 back to compute node 1. It should be understood that the probe strategy can also include other elements, such as probe time periods (e.g., not probing during peak periods), etc., which will not be listed here.

[0049] Factors influencing the detection strategy include: user demand, the resource size of computing nodes within the resource group, and the resource utilization of computing nodes within the resource group. The resources of computing nodes within the resource group include processor processing speed, memory capacity, and bandwidth, among others. When the user requirement is high reliability for hot migration, the number of probe virtual machines generated within the resource group can be set relatively high; the specifications of the probe virtual machines generated within the resource group can be set relatively high; the probe cycle of the probe virtual machines generated within the resource group can be set relatively short. When the user requirement is low resource consumption, the number of probe virtual machines generated within the resource group can be set relatively low; the specifications of the probe virtual machines generated within the resource group can be set relatively low; the probe cycle of the probe virtual machines generated within the resource group can be set relatively long; when the resource size of the compute nodes within the resource group is relatively large, the number of probe virtual machines generated within the resource group can be set relatively high; when the resource size of the compute nodes within the resource group is relatively large, the specifications of the probe virtual machines generated within the resource group can be set relatively high; when the resource size of the compute nodes within the resource group is relatively large, the specifications of the probe virtual machines generated within the resource group can be set relatively high; when the resource size of the compute nodes within the resource group is relatively large, the specifications of the probe virtual machines generated within the resource group can be set relatively high. When the source size is large, the detection cycle of the probe virtual machines generated within the resource group can be set to a shorter duration. When the resource utilization of the compute nodes within the resource group is low, the number of probe virtual machines generated within the resource group can be set to a larger duration. When the resource utilization of the compute nodes within the resource group is low, the specifications of the probe virtual machines generated within the resource group can be set to a lower duration. When the resource utilization of the compute nodes within the resource group is low, the detection cycle of the probe virtual machines generated within the resource group can be set to a longer duration. When the user's requirement is to focus on monitoring the reliability of a specific compute node, the detection route of the probe virtual machines generated within the resource group can be set to traverse that compute node multiple times. When the resource size of the compute nodes within the resource group is large, the detection route of the probe virtual machines generated within the resource group can be set to traverse that compute node multiple times. When the resource utilization of the compute nodes within the resource group is low, the detection route of the probe virtual machines generated within the resource group can be set to traverse that compute node multiple times. It should be understood that other factors can also influence the detection strategy, such as the failure rate of compute nodes, etc., which will not be listed here.

[0050] In one specific implementation, the virtual machine hot migration probe 422 determines whether the compute node 421 in the resource group 420 has the ability to hot migrate virtual machines by generating multiple probe virtual machines within the resource group and enabling these probe virtual machines to migrate between compute nodes within the resource group. This determines whether each compute node in the resource group has the ability to hot migrate virtual machines and automatically isolates faulty nodes (i.e., compute nodes that do not have the ability to hot migrate). The probe virtual machine is only used to detect whether a compute node has the ability to hot migrate and does not undertake other business processing tasks. Generally speaking, probe virtual machines have the following characteristics compared to normal virtual machines: (1) probe virtual machines do not need to undertake daily business processing tasks, while normal virtual machines do; (2) probe virtual machines occupy fewer resources, while normal virtual machines occupy more resources. Specifically, probe virtual machines occupy fewer compute resources than normal virtual machines, probe virtual machines occupy fewer storage resources than normal virtual machines, probe virtual machines occupy fewer network resources than normal virtual machines, etc.

[0051] For example, suppose virtual machine hot migration probe 422 can generate probe virtual machine 1 in compute node 1, probe virtual machine 2 in compute node 2, and probe virtual machine 3 in compute node 3 within resource group 1. Specifically, probe virtual machine 1 migrates from compute node 1 to compute node 2, probe virtual machine 2 migrates from compute node 2 to compute node 3, and probe virtual machine 3 migrates from compute node 3 to compute node 1.

[0052] If virtual machine 1 can be hot-migrated from compute node 1 to compute node 2 in the resource group, virtual machine 2 can be hot-migrated from compute node 2 to compute node 3 in the resource group, and virtual machine 3 can be hot-migrated from compute node 3 to compute node 1 in the resource group, then it can be determined that compute node 1, compute node 2, and compute node 3 all have hot-migrating capabilities, and there are no faulty nodes in the resource group.

[0053] If virtual machine 1 can be hot-migrated from compute node 1 to compute node 2 within the resource group, virtual machine 2 cannot be hot-migrated from compute node 2 to compute node 3 within the resource group, and virtual machine 3 cannot be hot-migrated from compute node 3 to compute node 1 within the resource group, then it can be determined that compute node 1 and compute node 2 have hot-migrating capabilities, while compute node 3 does not have hot-migrating capabilities, and compute node 3 will be automatically isolated.

[0054] If virtual machine 1 cannot be hot-migrated from compute node 1 to compute node 2 in the resource group, virtual machine 2 can be hot-migrated from compute node 2 to compute node 3 in the resource group, and virtual machine 3 cannot be hot-migrated from compute node 3 to compute node 1 in the resource group, then it can be determined that compute node 2 and compute node 3 have hot-migrating capabilities, compute node 1 does not have hot-migrating capabilities, and compute node 1 will be automatically isolated.

[0055] If virtual machine 1 cannot be hot-migrated from compute node 1 to compute node 2 in the resource group, virtual machine 2 cannot be hot-migrated from compute node 2 to compute node 3 in the resource group, and virtual machine 3 can be hot-migrated from compute node 3 to compute node 1 in the resource group, then it can be determined that compute node 1 and compute node 3 have hot-migrating capabilities, compute node 2 does not have hot-migrating capabilities, and compute node 2 will be automatically isolated.

[0056] If virtual machine 1 cannot be hot-migrated from compute node 1 to compute node 2 in the resource group, virtual machine 2 cannot be hot-migrated from compute node 2 to compute node 3 in the resource group, and virtual machine 3 cannot be hot-migrated from compute node 3 to compute node 1 in the resource group, then it can be determined that at least two compute nodes, compute node 2 and compute node 3, do not have hot-migrating capabilities, and compute node 1 to compute node 3 will be automatically isolated.

[0057] It should be understood that the above example only uses virtual machine hot migration detection 422 to determine whether a compute node has hot migration capability by detecting whether the virtual machine's hot migration is successful. However, in practical applications, virtual machine hot migration detection 422 can also combine the detection of the failure reasons when the virtual machine fails to determine whether the compute node has hot migration capability. Failure reasons can include: insufficient resources on the target host, incorrect resource isolation configuration on the target host, incorrect OS kernel configuration on the target host, unreachable network on the target host, and the absence of the target host compute agent process, etc. Based on the different failure reasons and error codes retrieved from the migration results, it can be linked with the historical case database to determine whether the failure is due to compute node failure or human configuration error, ultimately determining whether the compute node has hot migration capability.

[0058] See Figure 6 , Figure 6 This is a schematic diagram of the structure of a physical host provided in this application. The physical host may typically include a first processor 610, a first memory 620, a smart network card 630, and a bus 640.

[0059] The first processor 610 can be one or more general-purpose processors. These general-purpose processors can be any type of device capable of processing electronic instructions, including a central processing unit (CPU), microprocessor, microcontroller, main processor, controller, and application-specific integrated circuit (ASIC), etc. The first processor 610 executes various types of digital storage instructions, such as software or firmware programs stored in the first memory 620. In a specific embodiment, the first processor 610 can be an x86 processor, etc. The first processor 610 sends commands to the first memory 620 through a physical interface to complete storage-related tasks. For example, the commands that the first processor 610 can provide include read commands, write commands, copy commands, and erase commands, etc. The commands can specify operations related to specific pages and blocks of the first memory 620. An operating system can be installed within the first processor 610, and a cloud management platform can be installed on the operating system.

[0060] The cloud management platform can be AWS (Amazon Web Services), OpenStack, CloudStack, Hadoop, Apache Mesos, Docker-based Kubernetes, Swarm, etc., without specific limitations. The cloud management platform 211 can include computing services, storage services, and network services. Computing services can be Nova, storage services can be Swift, and network services can be Neutron, etc. Of course, the cloud management platform 211 can also include other services, such as authentication services, image services, monitoring services, cluster services, database services, front-end services, etc. Authentication services can be Keystone, image services can be Glance, monitoring services can be Ceilometer, cluster services can be Heat, database services can be Trove, and front-end services can be Horizon, etc., without further examples.

[0061] The first memory 620 may include random access memory (RAM), flash memory, or RAM, read-only memory (ROM), hard disk drive (HDD), or solid-state drive (SSD).

[0062] The smart network interface card (NIC) 630, also known as a network interface controller, network interface card, or local area network (LAN) adapter, has a unique MAC address, which is burned into a read-only memory chip by the manufacturer during production. The smart NIC 630 includes a second processor 631, a second memory 632, and a transceiver 633. The second processor 631 is similar to the first processor 620, but its performance requirements can be lower. In one specific embodiment, the second processor 631 can be an ARM processor, etc. The second memory 632 can also be flash memory, HDD, or SDD, and its storage capacity can be smaller than that of the first memory 620. The transceiver 633 can receive and send messages, and upload the received messages to the second processor 631 for processing. The smart NIC 630 may also include multiple ports, which can be any one or more of the following interface types: thick cable interface, thin cable interface, and twisted pair interface.

[0063] The first processor 610 reads the program code from the first memory 620 and executes the following method:

[0064] Obtain policy information for a resource group, wherein the resource group includes multiple computing nodes;

[0065] A detection strategy is generated based on the strategy information. The detection strategy is used to instruct the generation of a detection virtual machine in the plurality of computing nodes, and to enable the detection virtual machine to perform hot migration between the plurality of computing nodes to detect whether any computing node in the plurality of computing nodes has hot migration capability. The detection virtual machine is only used to detect whether the plurality of computing nodes have hot migration capability and does not undertake other business processing work.

[0066] Receive the detection results reported by the detection virtual machine;

[0067] Based on the detection results, determine whether any one of the plurality of computing nodes has hot migration capability.

[0068] Optionally, the probe virtual machine meets one or more of the following criteria: the probe virtual machine occupies less computing resources than the normal virtual machine, the probe virtual machine occupies less storage resources than the normal virtual machine, and the probe virtual machine occupies less network resources than the normal virtual machine.

[0069] Optionally, the strategy information includes one or more of the following: user requirements, the resource size of each computing node among the plurality of computing nodes, and the resource occupancy status of each computing node among the plurality of computing nodes.

[0070] Optionally, the detection strategy may include one or more of the following: the number of virtual machines to be detected, their specifications, the detection cycle, and the detection route.

[0071] For simplicity, the working process of the physical host is not described in detail here. Please refer to [link to relevant documentation]. Figures 1 to 4 And related statements, which will not be described in detail here.

[0072] In the above scheme, before performing hot migration of heavyweight virtual machines, a lightweight probe virtual machine is first used to detect whether the compute node has hot migration capability. If it is determined that the probe virtual machine has hot migration capability, then the heavyweight virtual machine is hot migrated, thereby improving the success rate of hot migration of virtual machines.

[0073] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, storage disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

Claims

1. A fault detection method, characterized in that, include: Obtain policy information for a resource group, wherein the resource group includes multiple computing nodes; A detection policy is generated based on the policy information, wherein the detection policy is used to instruct the generation of a detection virtual machine in the plurality of computing nodes, and to instruct the detection virtual machine to perform hot migration between the plurality of computing nodes to detect whether any computing node among the plurality of computing nodes has hot migration capability, wherein the detection virtual machine meets one or more of the following criteria: the detection virtual machine occupies less computing resources than the normal virtual machine, the detection virtual machine occupies less storage resources than the normal virtual machine, and the detection virtual machine occupies less network resources than the normal virtual machine. Receive the detection results reported by the detection virtual machine; Based on the detection results, determine whether any one of the plurality of computing nodes has hot migration capability.

2. The method according to claim 1, characterized in that, The strategy information includes one or more of the following: user requirements, the resource size of each computing node among the plurality of computing nodes, and the resource occupancy status of each computing node among the plurality of computing nodes.

3. The method according to claim 1 or 2, characterized in that, The detection strategy includes one or more of the following: the number of virtual machines to be detected, their specifications, the detection cycle, and the detection route.

4. A physical host, characterized in that, include: The module includes an acquisition module, a generation module, a receiving module, and a determination module. The acquisition module is used to acquire the strategy information of the resource group, wherein the resource group includes multiple computing nodes; The generation module is configured to generate a detection strategy based on the strategy information. The detection strategy is configured to instruct the generation of a detection virtual machine in the plurality of computing nodes and to instruct the detection virtual machine to perform hot migration between the plurality of computing nodes to detect whether any computing node among the plurality of computing nodes has hot migration capability. The detection virtual machine meets one or more of the following criteria: the detection virtual machine occupies less computing resources than the normal virtual machine, the detection virtual machine occupies less storage resources than the normal virtual machine, and the detection virtual machine occupies less network resources than the normal virtual machine. The receiving module is used to receive the detection results reported by the detection virtual machine; The determining module is used to determine, based on the detection results, whether any one of the plurality of computing nodes has hot migration capability.

5. The host computer according to claim 4, characterized in that, The strategy information includes one or more of the following: user requirements, the resource size of each computing node among the plurality of computing nodes, and the resource occupancy status of each computing node among the plurality of computing nodes.

6. The host computer according to claim 4 or 5, characterized in that, The detection strategy includes one or more of the following: the number of virtual machines to be detected, their specifications, the detection cycle, and the detection route.

7. A computing node, characterized in that, It includes a memory and a processor, the processor executing programs in the memory to run computing and storage services, thereby performing the method as claimed in any one of claims 1 to 3.

8. A readable storage medium, characterized in that, Includes instructions that, when executed on a computing node, cause the computing node to perform the method as described in any one of claims 1 to 3.

Images