Method, device, system, medium and product for managing bare metal device

Abstract

The application discloses a management method, device, system, medium and product of a bare metal device. The method comprises: in a management mode of a running mode of the bare metal device, starting a management instance on an intelligent network card to control the working state of the management instance to switch from a shutdown state to a standby running state; in the running state, in response to installation instruction information, switching the working state from the running state to an operation state for executing an operation; and in the operation state, installing an operating system of the bare metal device based on the management instance to create a user instance of the bare metal device; and determining that the user instance is created, and then controlling the running mode to switch from the management mode to a user mode. In this way, the whole life cycle of the user instance is managed based on the management instance of the intelligent network card, and the management efficiency and flexibility of the bare metal server are improved.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a method, apparatus, system, medium and product for managing bare metal equipment. Background Technology

[0002] Bare metal servers refer to physical servers delivered as a complete machine. Compared to virtual machines that consume some of the host machine's resources, bare metal servers are provided to users as a whole physical server, and have the characteristics of no virtualization loss, high performance, and good reliability, making them suitable for high-performance computing and other scenarios.

[0003] In related technologies, the management of bare metal servers mainly relies on the following two core methods: management scheme based on PXE (Preboot Execute Environment) and management scheme based on cloud disk storage offloading + Cloud-init. (1) For the management scheme based on PXE, its advantage is that it can effectively manage the local resources of the bare metal server, but this method also has some disadvantages: PXE is highly dependent on the network environment and is easily affected by network conditions. Moreover, the entire process of PXE is relatively complex and usually takes a long time. Furthermore, different machine models have a certain impact on the stability of the PXE system's successful startup. (2) Bare metal management based on cloud disk storage offloading has the advantage of high deployment flexibility and strong adaptability of cloud disks, realizing the separation of bare metal installation and image deployment operations. However, it also faces the problems of high dependence on cloud storage services, time-consuming cloud disk creation, and insufficient flexibility in the management of local resources.

[0004] As can be seen from the above, existing bare metal management methods are inadequate in terms of stability, speed, scalability, and resource management, which limits their performance in efficient and flexible management. Summary of the Invention

[0005] In view of this, embodiments of this application provide a method, apparatus, system, medium, and product for managing bare metal, aiming to solve the problems of insufficient stability, speed, scalability, and resource management in the control of bare metal equipment.

[0006] The technical solution of this application embodiment is implemented as follows:

[0007] In a first aspect, embodiments of this application provide a method for managing a bare metal device, the bare metal device including a smart network interface card (NIC), the method comprising:

[0008] When the bare metal device is in management mode, the management instance on the smart network card is started to control the working state of the management instance to switch from the power-off state to the standby running state;

[0009] In the running state, in response to the installation instruction information, the working state is switched from the running state to the operation state of performing operations; and in the operation state, the operating system of the bare metal device is installed based on the management instance to create a user instance of the bare metal device;

[0010] Once the user instance is confirmed to have been created, the operating mode is switched from the management mode to the user mode.

[0011] In some embodiments, the method further includes:

[0012] In the user mode, obtain management request information for the user instance;

[0013] In response to the management request information, the operating mode is switched from the user mode to the management mode, and in the management mode, management operations are performed on the user instance based on the management instance.

[0014] In some embodiments, the method further includes:

[0015] The operating mode of the bare metal equipment is determined to be the empty machine mode;

[0016] In response to the first instruction, the idle machine mode is switched to the management mode.

[0017] In some embodiments, the method further includes:

[0018] In the running state, it is determined whether an installation order information has been received. If so, the installation instruction information is generated and sent.

[0019] In some embodiments, the method further includes:

[0020] In the running state, if the installation order information is received, the installation order type is obtained, which includes cloud disk system disk order type and local system disk order type;

[0021] If the installation order type is determined to be a local system disk order type, then the installation instruction information is generated and sent.

[0022] In some embodiments, the method further includes:

[0023] In the running state, if the installation order information is not received and the bare metal physical machine is not scheduled, then the heartbeat signal data of the management instance and the physical environment information of the smart network card are obtained;

[0024] Send the heartbeat signal data and the physical environment information.

[0025] In some embodiments, the management request information includes at least one of the following: a disk cleanup command, a local disk image creation command, and a system reinstallation command. The management operation performed on the user instance based on the management instance in the management mode includes:

[0026] In the management mode, the management instance is controlled to switch from the running state to the operating state;

[0027] In the operating state, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, at least one of the disk cleanup operation, local disk image operation, and operating system reinstallation operation is performed on the user instance based on the management instance to perform management operations on the user instance.

[0028] In some embodiments, the bare metal device includes a bare metal physical machine, and in the management mode, the management operation on the user instance based on the management instance includes:

[0029] In the management mode, the management instance is controlled to operate the bare metal physical machine in order to perform management operations on the user instance.

[0030] In some embodiments, controlling the switching of the operating mode from the management mode to the user mode includes:

[0031] Uninstall the management instance and switch the running state of the management instance to the shutdown state;

[0032] Mount the system disk and network module corresponding to the user instance, and start the user instance.

[0033] In some embodiments, the method further includes:

[0034] If the management operation is determined to be completed, the operating mode is switched from the management mode to the user mode.

[0035] In some embodiments, controlling the operation mode to switch from the user mode to the management mode includes: uninstalling the user instance and starting the management instance.

[0036] In some embodiments, the method further includes:

[0037] In the operating state, acquire maintenance instruction information for bare metal equipment;

[0038] In response to the maintenance instruction information, maintenance operations are performed on the bare metal equipment based on the management instance, and the maintenance operations include at least one of the following:

[0039] Hardware quality inspection operations, hardware repair operations, and configuration operations within the operating system.

[0040] Secondly, embodiments of this application also provide a management device for a bare metal device, the bare metal device including a smart network interface card (NIC), the management device comprising:

[0041] The first control module is used to start the management instance on the smart network card when the bare metal device is in management mode, so as to control the working state of the management instance to switch from the power-off state to the standby running state;

[0042] A switching module is used to switch the working state from the running state to the operation state of performing operations in response to installation instruction information in the running state; and in the operation state, to install the operating system of the bare metal device based on the management instance to create a user instance of the bare metal device;

[0043] The second control module is used to determine that the user instance has been created, and then control the operation mode to switch from the management mode to the user mode.

[0044] In some embodiments, the management device further includes:

[0045] The first acquisition module is used to acquire management request information for the user instance in the user mode.

[0046] The second control module is also configured to, in response to the management request information, control the operation mode to switch from the user mode to the management mode, and in the management mode, perform management operations on the user instance based on the management instance.

[0047] In some embodiments, the management device further includes:

[0048] The determination module is used to determine that the operating mode of the bare metal equipment is an empty machine mode;

[0049] The switching module is also used to switch the idle machine mode to the management mode in response to the first instruction information.

[0050] In some embodiments, the determining module is further configured to determine, in the running state, whether installation order information has been received, and if so, to generate and send the installation instruction information.

[0051] In some embodiments, the determining module is further configured to, in the running state, if the installation order information is received, obtain the installation order type, wherein the installation order type includes cloud disk system disk order type and local system disk order type;

[0052] If the installation order type is determined to be a local system disk order type, then the installation instruction information is generated and sent.

[0053] In some embodiments, the determining module is further configured to, in the running state, if the installation order information is not received and the bare metal physical machine is not scheduled, obtain the heartbeat signal data of the management instance and the physical environment information of the smart network card;

[0054] The management device further includes:

[0055] The transmitting module is used to transmit the heartbeat signal data and the physical environment information.

[0056] In some embodiments, the management request information includes at least one of the following: disk cleanup instructions, local disk image creation instructions, and system reinstallation instructions. The second control module is also configured to control the management instance to switch from the running state to the operating state in the management mode.

[0057] In the operating state, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, at least one of the disk cleanup operation, local disk image operation, and operating system reinstallation operation is performed on the user instance based on the management instance to perform management operations on the user instance.

[0058] In some embodiments, the bare metal device includes a bare metal physical machine, and the second control module is further configured to control the management instance to operate the bare metal physical machine in the management mode to perform management operations on the user instance.

[0059] In some embodiments, the second control module is further configured to unload the management instance and switch the running state of the management instance to a shutdown state;

[0060] Mount the system disk and network module corresponding to the user instance, and start the user instance.

[0061] In some embodiments, the second control module is further configured to control the operating mode to switch from the management mode to the user mode if it is determined that the management operation is completed.

[0062] In some embodiments, the second control module is further configured to unload the user instance and start the management instance.

[0063] In some embodiments, the management device further includes:

[0064] The second acquisition module is used to acquire maintenance instruction information for bare metal equipment in the running state;

[0065] The second control module is further configured to, in response to the maintenance instruction information, perform maintenance operations on the bare metal equipment based on the management instance, wherein the maintenance operations include at least one of the following:

[0066] Hardware quality inspection operations, hardware repair operations, and configuration operations within the operating system.

[0067] Thirdly, embodiments of this application also provide a management system for a bare metal device, the bare metal device including a smart network interface card (NIC), the management system including: a management device and a management instance component, the management device being connected to the management instance component, the management device being used to execute the method described in the first aspect, and the management instance component including:

[0068] The system disk component is used to load the disk image of the management instance, wherein the disk image of the management instance is a trimmed remote disk image of the smart network card;

[0069] The network components establish internal network connections with the internal network interface of the smart network card and external network connections with the external network module, respectively.

[0070] The execution component is further configured to invoke the system disk component to load the disk image in order to start the management instance.

[0071] In some embodiments, the bare metal device further includes a bare metal physical machine connected to the internal network interface, the bare metal physical machine including multiple local disks.

[0072] In some embodiments, the smart network interface card includes:

[0073] A first physical function module, which is connected to the internal network interface;

[0074] The second physical function module is connected to the system disk.

[0075] In some embodiments, the smart network interface card (NIC) is a first smart NIC with the control channel not enabled, and the network component includes:

[0076] The virtual switching network module is connected to both the first physical functional module and the external network module.

[0077] In some embodiments, the management instance component further includes:

[0078] A first backup storage module is connected to the system disk component, and the first backup storage module includes a snapshot module or a block storage module.

[0079] In some embodiments, the smart network interface card (NIC) is a second smart NIC with an enabled control channel, the external network module is a virtualization tool module, and the network component is a third physical function module on the second smart NIC. The third physical function module is used to form the control channel between the first physical function module and the second physical function module. The control channel is used to establish an internal connection with the internal interface and an external network connection with the virtualization tool module.

[0080] In some embodiments, the management instance component includes:

[0081] The second backup storage module is connected to the system disk component, and the second backup storage module is a block storage module.

[0082] Fourthly, embodiments of this application provide a bare metal device, including: a processor and a memory for storing a computer program capable of running on the processor, wherein, when the processor runs the computer program, it executes the steps of the method described in the first aspect of embodiments of this application.

[0083] Fifthly, embodiments of this application provide a storage medium storing a computer program, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0084] In a sixth aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0085] The technical solution provided in this application embodiment is a bare metal management method. The bare metal device includes a smart network interface card (NIC). The method includes: when the bare metal device is in management mode, starting a management instance on the smart NIC to control the working state of the management instance to switch from a power-off state to a standby running state; in the running state, responding to installation instruction information, switching the working state from the running state to an operation state for performing operations; and in the operation state, installing an operating system for the bare metal device based on the management instance to create a user instance for the bare metal device; and after determining that the user instance has been created, controlling the running mode to switch from the management mode to the user mode.

[0086] Therefore, this application proposes a dynamic management method for management instances based on smart network interface cards (NICs), including the startup of management instances, switching of working states, task execution, and conversion of operating modes. This enables full lifecycle management of user instances by management instances based on smart NICs, improving the management efficiency and flexibility of bare metal servers. Specifically, this application dynamically starts management instances on smart NICs and flexibly switches the running, shutdown, and operation states of management instances according to business needs, ensuring rapid and efficient switching between management and user modes for bare metal devices. Simultaneously, this method supports parallel management of multiple instances on a single bare metal physical machine, significantly optimizing the bare metal device installation process, shortening delivery time, and improving resource utilization efficiency. Attached Figure Description

[0087] Figure 1 A flowchart illustrating the management method for bare metal equipment provided in this application embodiment;

[0088] Figure 2 This is a schematic diagram of the structure for single-machine multi-instance management provided in an embodiment of this application;

[0089] Figure 3 A schematic diagram illustrating the switching between user mode and management mode provided in an embodiment of this application;

[0090] Figure 4 This is a schematic diagram of the structure of the management instance component provided in the embodiments of this application;

[0091] Figure 5 A schematic diagram of the architecture of the EBC management instance component of the smart network interface card (NIC) for BF2 / BF3 provided as an application example of this application;

[0092] Figure 6 A schematic diagram of the architecture of the EBBC management instance component of the DPU provided as an application example of this application;

[0093] Figure 7 A flowchart illustrating the process of installing a local system disk based on a management instance, providing yet another application example of this application;

[0094] Figure 8 A schematic diagram of the structure of the management device for bare metal equipment provided in this application;

[0095] Figure 9 A schematic diagram of the bare metal device provided in this application. Detailed Implementation

[0096] The present application will now be described in further detail with reference to the accompanying drawings and embodiments.

[0097] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0098] In related technologies, the management of bare metal servers mainly relies on the following two core methods: PXE-based management solutions and cloud disk storage offloading + Cloud-init management solutions.

[0099] I. Management solution based on PXE (Preboot Execute Environment).

[0100] PXE is a protocol used to boot computers in a network environment. The basic workflow of this solution is as follows:

[0101] (1) Deploy PXE services: PXE-enabled network services need to be deployed in a dedicated environment.

[0102] (2) Startup process: When the bare metal server needs to perform operations such as installation, it needs to be restarted and enter PXE mode through the network card firmware support. Then the system obtains the IP address through the DHCP (Dynamic Host Configuration Protocol) protocol, and then downloads the bare metal server boot file and boot image from the PXE service environment.

[0103] (3) User operation: When a user needs to perform operations related to local resources on the bare metal server, the user first needs to complete the PXE process. Then, in the PXE environment, the user can continue to complete tasks such as operating system installation, password modification, local disk formatting, and image creation. After the user completes the operation, the server exits the PXE mode and switches back to the user's original operating system environment.

[0104] The advantage of PXE-based bare metal management is that it can achieve comprehensive and effective management of local resources on bare metal servers. However, this method also has some disadvantages: PXE is highly dependent on the network environment and is easily affected by network conditions. In addition, the entire PXE process is relatively complex and usually takes a long time. Furthermore, different machine models have a certain impact on the stability of the PXE system's successful startup.

[0105] II. Bare metal management based on cloud disk storage offloading.

[0106] This method primarily manages bare-metal services through Cinder (a cloud storage service component). The basic process is as follows:

[0107] (1) Cloud disk management: Users can perform management operations such as creating and deleting cloud disks through Cinder.

[0108] (2) Smart network card function: Combined with the smart network card, the bare metal physical machine can mount the cloud disk, that is, the bare metal server can directly use the cloud disk as its own storage resource.

[0109] (3) Installation separation: In the management scheme based on cloud disk storage unloading, there is no need to install bare metal disks locally, and the system disk can be replaced by cloud disk.

[0110] (4) Initialization operations: Bare metal installation, image deployment, and initialization are all completed on the Cinde side. Other user initialization requirements can be met through the capabilities provided by Cloud-int.

[0111] Bare metal management based on cloud disk storage offloading has the advantages of high deployment flexibility and adaptability of cloud disks, and it separates bare metal installation and image deployment operations. However, it also faces problems such as high dependence on cloud storage services, time-consuming cloud disk creation, and insufficient flexibility in local resource management.

[0112] Based on this, the bare metal control methods in related technologies have the following drawbacks:

[0113] I. PXE-based bare metal management:

[0114] (1) Stability depends on network environment: The stability of PXE is highly dependent on network quality. Since the boot files and PXE images need to be downloaded from the network, unstable network or limited network bandwidth may cause download failure, thus affecting the entire PXE boot process.

[0115] (2) The PXE operation process is relatively long: In the PXE environment, the physical machine needs to be restarted multiple times from entering PXE to completing the operation, which increases the PXE deployment time.

[0116] (3) Compatibility issues: The PXE boot file needs to be compatible with different types of smart cards and bare metal models. If it is not compatible, PXE will fail to start, which will affect the normal delivery of the server.

[0117] (4) Network security risks: PXE environments require dedicated independent networks. If they rely on the user’s network environment, network security risks will be introduced, especially in multi-tenant environments.

[0118] II. Bare Metal Management Based on Cloud Disk Storage Offloading:

[0119] (1) Limited management scalability: It is impossible to fully manage the local disk and other hardware resources of the bare metal server, and the scalability of supported functions is limited, especially in terms of supporting multiple hardware configurations. This makes it impossible for users to fully utilize the hardware performance when using bare metal service products, affecting the overall resource utilization rate.

[0120] (2) Stability depends on cloud storage services: Cloud disk management operations, such as creating, deleting and releasing cloud disk resources, involve a certain amount of time and are highly dependent on the stability and reliability of cloud storage services. If cloud storage services have problems, it will directly affect the management efficiency of bare metal servers and increase the complexity and risk of operation and maintenance.

[0121] As can be seen from the above, existing bare metal management methods are inadequate in terms of stability, speed, scalability, and resource management, which limits their performance in efficient and flexible management.

[0122] Based on this, embodiments of this application provide a bare metal device, such as... Figure 1 As shown, the bare metal device includes a smart network interface card (NIC) and a management instance component, with the management instance component located on the smart NIC.

[0123] Here, bare metal equipment refers to physical servers delivered as complete units. In practical applications, to meet the direct needs of customers, the operating system needs to be pre-installed and configured on the bare metal equipment before delivery.

[0124] Here, a smart network interface card (NIC) is a network interface card with hardware acceleration capabilities. It not only supports high-performance network communication but can also perform some computing tasks. The management instance component is a resident, lightweight instance that can be used to manage and control bare metal devices.

[0125] In this embodiment of the application, the management instance component is set on the smart network interface card (NIC), which serves as the operating platform for the management instance component and is responsible for providing network connectivity and computing resource support.

[0126] Therefore, this application proposes a dynamic management method for management instances based on smart network interface cards (NICs), including the startup of management instances, switching of working states, task execution, and conversion of operating modes. This enables full lifecycle management of user instances by management instances based on smart NICs, improving the management efficiency and flexibility of bare metal servers. Specifically, this application dynamically starts management instances on smart NICs and flexibly switches the running, shutdown, and operation states of management instances according to business needs, ensuring rapid and efficient switching between management and user modes for bare metal devices. Simultaneously, this method supports parallel management of multiple instances on a single bare metal physical machine, significantly optimizing the bare metal device installation process, shortening delivery time, and improving resource utilization efficiency.

[0127] Based on the above-mentioned bare metal equipment, this application provides a method for managing bare metal equipment.

[0128] It should be noted that, in this embodiment of the application, the instance metadata center (nova Master) and ironicmanager, as the core parts of the bare metal device management module, are jointly responsible for switching the management mode of the bare metal device and switching the running status of the managed instance, thereby realizing the full lifecycle management of user instances.

[0129] The Nova Master is the core in-memory data center that manages all instances (including user and management instances). It stores detailed information and status of all instances, the current status and operational information of bare-metal physical machines, and records of business requests over a specific period. In practice, when there are order requests, business operations, or maintenance needs, the Nova API on the control plane registers the request with the Nova Master. The Nova Master determines whether the request can be accepted based on the current status of the physical machine and the instance. If the request is accepted, the Nova Master will normally send the scheduling result to the Ironic Manager.

[0130] Here, the ironic-Manager is the central control node, primarily responsible for managing operations such as starting up the management instance, executing specific tasks, and shutting it down; as well as monitoring the status of bare metal devices and updating the Nova Master in real time with changes in the status of the management instance and the bare metal physical machine. In practical applications, the ironic Manager completes bare metal device management operations through interaction, such as starting the management instance, switching running states, and installing the operating system.

[0131] For example, the management instance is controlled by the central control node ironic-manager. The ironic-manager accepts requests from NovaMaster (which stores the state of the physical machine), and completes tasks such as starting the management instance, performing operations under the management instance, and shutting down and exiting the management instance. Based on the results of each stage, it updates the state of the management instance to NovaMaster.

[0132] When there is an order request, business operation, or operation and maintenance operation requirement, the request will first pass through the API layer of the control plane and ironic-Manager. Nova-api registers a request with NovaMaster. The metadata center determines whether the current bare metal physical machine can accept the corresponding request based on the current status of the bare metal physical machine and the status of the management instance. If it accepts, it will be normally sent to ironic-Manager, which is responsible for interacting with the metadata center to update the status of the management instance.

[0133] It should be noted that bare metal equipment has three main operating modes: empty machine mode, user mode, and management mode.

[0134] An idle machine refers to a bare metal device in an idle state (not taken over by a user or management instance). It belongs to neither user mode nor management mode. In other words, the machine is only in management mode when a management instance needs to be launched to perform operations. An empty physical machine belongs to neither user mode nor management mode.

[0135] User mode and management mode are the operating states of the bare metal physical machine, which are updated by ironic-Manager to NovaMaster, usually when the xx operation is completed.

[0136] User mode: Bare metal devices are hosted by user instances to run the user's business operating system (guestOS).

[0137] Management mode: Bare metal devices are managed by a managed instance to perform management tasks (such as installation, hardware testing, data cleanup, etc.).

[0138] When a bare metal device needs to perform a management task, Ironic Manager triggers management mode, starts a management instance, and switches the device from user mode to management mode. After the task is completed, the management instance exits and releases resources, and the device switches back to user mode.

[0139] Here, the ironic manager directly manages the instance and triggers the nova-executor (execution component) to execute tasks through commands (such as installation task commands). In management mode, the nova-executor runs in the management instance and is responsible for executing tasks issued by the ironic manager, such as operating system installation and hardware detection.

[0140] It should be noted that the management instance is the core functional module in management mode, and it has the following three working states, which can be dynamically switched according to the task. The management instance has multiple working states, including shutdown state, running state, and operation state. Shutdown state: The management instance is not started and does not consume resources. Running state (standby state): The management instance has started, but has not executed any specific tasks and is waiting for task instructions. Operation state (execution state): The management instance enters the task execution phase (such as installation, hardware detection, etc.). The operation state is the core operating state of the management instance, used to complete the management tasks of bare metal devices.

[0141] like Figure 1 As shown, this management method includes the following steps:

[0142] Step 110: With the bare metal device in management mode, start the management instance on the smart network card to control the management instance to switch from the power-off state to the standby running state.

[0143] Here, the bare metal device is in management mode. In this mode, the bare metal is managed by a management instance, and the management instance on the smart NIC is started. Initially, the management instance is in a powered-off state (i.e., not running). The startup process switches its state to a standby running state. The standby running state means the management instance has completed initialization but has not yet executed any specific tasks; it is in a ready state to receive commands at any time. Starting the management instance on the smart NIC controls the switch of its working state from the powered-off state to the standby running state, preparing it for subsequent tasks.

[0144] Here, the management instance is a virtualized environment or operating system deployed on a smart network interface card (NIC), and is a management tool for bare metal devices.

[0145] Step 120: In the running state, in response to the installation instruction information, switch the working state from the running state to the operation state to perform the operation; and in the operation state, install the operating system of the bare metal device based on the management instance to create a user instance of the bare metal device.

[0146] Here, in the running state of the management instance (i.e., the management instance has started but has not executed any tasks), it receives and responds to the installation instruction information, switching the working state of the management instance from running state to operation state (i.e., entering the specific task execution state). That is, to create a user instance, the management instance (which is normally in standby state) must first switch from standby state to powered-on running state, then to order operation state. In operation state, the management instance, based on the installation instruction information, executes the installation task, installs the operating system of the bare metal device, and thus creates the user instance of the bare metal device.

[0147] It should be noted that a user instance is an independent computing environment or operating system running on a bare metal device, provided to end users for running their business applications. Since the management instance in this application is located on a smart network interface card (NIC), compared to existing technologies, it can quickly respond to installation task commands, thereby efficiently completing the installation operation on the bare metal device to create user instances. Furthermore, efficient management of bare metal user instances can be achieved subsequently through the management instance on the smart NIC.

[0148] For example, since bare metal devices typically do not come pre-installed with an operating system, users need to first complete the installation task on the bare metal device, i.e., install the operating system, through the management instance after purchasing bare metal services. After the operating system is installed, the user instance can run normally on the device and support the user's business needs.

[0149] Step 130: Once the user instance creation is complete, switch the control running mode from management mode to user mode.

[0150] Here, once the user instance is created, the controller switches the bare metal device's operating mode from management mode to user mode. In user mode, the bare metal device is taken over and run by the user instance, and the user can directly use the device to complete the corresponding tasks. At this time, the management instance exits and stops managing the device.

[0151] Thus, this application proposes a novel management method for bare metal devices, enabling full lifecycle management of user instances by a management instance based on a smart network interface card (NIC), thereby improving the management efficiency and flexibility of bare metal servers. By dynamically launching management instances on demand and flexibly switching between running, shut-down, and operational states according to business scenarios, the method drives efficient transitions between management and user modes for bare metal devices, while also supporting parallel management of multiple instances on a single bare metal physical machine. This optimizes the bare metal device installation process, shortens delivery time, enhances resource isolation and management flexibility, and provides innovative technical support for efficient management and flexible delivery in complex business scenarios.

[0152] In some embodiments, the method further includes:

[0153] In user mode, obtain management request information for user instances;

[0154] In response to a management request, the control operation mode switches from user mode to management mode, and in management mode, management operations are performed on user instances based on management instances.

[0155] It is understandable that when the bare metal device is in user mode (user instance running state), if a management request for the user instance is received, the device will respond to the management request and trigger a mode switch to switch the bare metal device from user mode to management mode. In management mode, management operations are performed on the user instance based on the management instance.

[0156] In some embodiments, the method further includes:

[0157] The operating mode of the bare metal equipment is determined to be the idle machine mode;

[0158] In response to the first instruction message, switch from idle machine mode to management mode.

[0159] Here, in idle machine mode, this initial indication is typically issued by the control node (such as Nova Master or Ironic Manager), primarily used to switch the device from an "idle" state to a "manageable" state during the bare metal device's lifecycle. Common triggering scenarios include resource allocation, state changes, operation and maintenance management, and automated scheduling.

[0160] For example, when a user initiates a resource allocation request through a cloud platform or management system (such as purchasing or applying for bare metal equipment), the bare metal equipment is initially in an empty machine mode (unallocated and uninitialized). It needs to enter "management mode" for initialization and preparation work, such as installation or configuration. At this time, a first instruction message is generated and sent to instruct the bare metal equipment to switch from the empty machine mode to the management mode. In response to the first instruction message, the empty machine mode is switched to the management mode to facilitate the execution of subsequent management operations.

[0161] Below, in conjunction with Figure 2 This document describes the process of gradually transitioning bare metal equipment from an idle state (empty machine) to a managed mode for installation, and finally switching it to user mode to run user instances. Figure 2 This is a schematic diagram illustrating the operation of a single machine with multiple instances. HOST refers to the bare metal physical machine, and BF2 / BF3 / DPU refers to the smart network interface card.

[0162] 1. Empty machine mode

[0163] Empty machine mode is the default state for bare metal devices, used to wait for task scheduling.

[0164] HOST: The bare metal physical machine is in a no-task state, with no management or user instance loaded.

[0165] BF2 / BF3 / DPU: This indicates that the smart network interface card (NIC) component is in an idle state and has not been assigned any management tasks. When an installation task command is received, the system will switch the bare metal device to management mode.

[0166] At this point, the bare metal device is in its initial state, not scheduled or assigned any tasks, and the resources of the smart network card and the physical host are not occupied.

[0167] 2. Management Model

[0168] In managed mode, bare metal equipment is taken over by managed instances to complete installation tasks or other management operations.

[0169] mgr OS: Represents the operating system of the management instance, which is typically used to perform management tasks on bare metal devices, such as installation tasks (operating system installation), resource configuration (such as disk partitioning and network initialization), and maintenance tasks (such as hardware detection and diagnosis).

[0170] BF2 / BF3 / DPU: Smart NICs are responsible for managing instances and providing network and storage resource support.

[0171] In management mode, a series of management tasks are performed, including installation tasks (such as installing the operating system), hardware detection, and data cleanup.

[0172] In management mode, the system will switch to user mode once the task is completed (such as the installation task).

[0173] (3) User Mode

[0174] In user mode, bare metal physical machines are hosted by user instances, running a user operating system (Guest OS), and physical resources are exclusively owned by the user instances.

[0175] guest OS: The operating system for user instances, supporting end-user business applications and computing tasks.

[0176] BF2 / BF3 / DPU: Smart NICs host the network resources of user instances, providing communication and data transmission support.

[0177] Here, when a user instance needs maintenance operations (such as disk cleanup, system reinstallation, etc.), the control node triggers a switch; the user instance resources are unloaded, the management instance is restarted, and the device enters management mode.

[0178] Here, in user mode, management request information can be generated in response to operation task instructions for user instances. The management request information is used to instruct the execution component to perform management operations on user instances based on management instances.

[0179] In some embodiments, the method further includes:

[0180] In runtime, determine whether installation order information has been received. If so, generate and send installation instruction information.

[0181] Here, in the running state, specifically in the running state of the management instance (i.e., the management instance has been started but has not executed any tasks), it is determined whether an installation order information has been received. If an installation order has been received, then it is necessary to install the operating system on the bare metal device, thereby generating and sending installation instruction information.

[0182] In some embodiments, the method further includes:

[0183] In runtime, if an installation order is received, the installation order type is obtained. The installation order type includes cloud disk system disk order type and local system disk order type.

[0184] If the installation order type is determined to be a local system disk order type, then an installation instruction message will be generated and sent.

[0185] Here, the installation order instruction information includes the installation order type (cloud disk system disk order type or local system disk order type).

[0186] Here, the cloud disk system disk order type refers to an order where the operating system is installed on a cloud disk, and the bare metal device accesses the system disk through remote cloud storage services. The local system disk order type refers to an order where the operating system is installed directly on the local disk of the bare metal physical machine.

[0187] It should be noted that cloud disk system disks are typically managed by cloud storage services (such as the Cinder module in OpenStack). Operations such as cloud disk creation, formatting, and mounting are automatically completed by the cloud service, decoupling them from the physical resources of the bare metal device. Since cloud disks are remote storage, operating system installation can be completed directly in the cloud, requiring no local intervention on the bare metal device.

[0188] Here, if the installation order type is determined to be a local system disk order type, then installation instruction information is generated and sent.

[0189] For example, for cloud disk system disk orders: the management instance does not perform any additional operations, but waits for the user cloud disk system disk on the Cinder side to be created. At this time, the management instance will unload its own system disk and service network card, and then the management instance will enter a shutdown silent state. Finally, it will mount the user instance resources and power on, and the machine will switch to user instance mode.

[0190] For local system disk orders (which require installation by a management instance): an installation instruction message is generated and sent. The management instance accesses the bare metal host side through its own network, completes the installation and related initialization operations on the local system disk, and after the operation is completed, the management instance resources are unloaded, the management instance enters a shutdown silent state, and the user instance system is started, and the machine switches to user instance mode.

[0191] Thus, the management instance is compatible with various bare metal models, including local system disk and cloud disk system disk models. Compared with the previous bare metal based on cloud disk storage offloading, the embodiments of this application provide more operation and management space through the management instance, enabling the bare metal server to support more user-defined operations and provide richer product capabilities.

[0192] In some embodiments, the method further includes:

[0193] In the running state, if no installation order information is received and the bare metal physical machine is not scheduled, the heartbeat signal data of the management instance and the physical environment information of the smart network card are obtained.

[0194] Send heartbeat signal data and physical environment information.

[0195] Here, heartbeat signal data is periodic status information sent by the management instance to the upper-level control node (such as Ironic Manager or NovaMaster) to indicate that the management instance is still operating normally.

[0196] Here, the physical environment information of the smart network card includes the hardware environment information of the smart network card, including but not limited to: hardware health status (such as the status of CPU, memory, and storage); network status (such as whether the network card is connected normally), etc.

[0197] For example, when the bare metal physical machine is idle and not scheduled, the management instance is in standby mode. In standby mode, the management instance periodically reports heartbeat signals to the metadata center and checks the current smart NIC hardware environment.

[0198] In some embodiments, the management request information includes at least one of the following: disk cleanup instructions, local disk image creation instructions, and system reinstallation instructions. In management mode, management operations are performed on user instances based on the management instance.

[0199] In management mode, control the switching of the management instance from running state to operating state;

[0200] In the operating state, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, at least one of the disk cleanup operation, local disk image operation, and operating system reinstallation operation is performed on the user instance based on the management instance to perform management operations on the user instance.

[0201] Understandably, disk cleanup commands are used to clean up stored data, such as data files on local disks, temporary files, or remnants of old operating system data. Local disk image creation commands are used to create an image file of the contents of a local disk on a bare-metal physical machine for later backup, migration, or cloning. System reinstallation commands are used to reinstall the operating system on a bare-metal physical machine, typically used for system failure repair, environment reset, or changing the operating system version.

[0202] Here, in management mode, the working state of the management instance switches from standby running state to operation state, indicating that the management instance has entered the task execution phase. At this time, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, the management instance performs at least one of the following operations on the user instance: disk cleanup operation, local disk image operation, and operating system reinstallation operation, in order to perform management operations on the user instance.

[0203] As you can understand, disk cleanup refers to clearing the storage space of a user instance and freeing up resources. Local disk imaging refers to creating a backup image of the user instance's local disk for recovery or copying. Operating system reinstallation refers to reinstalling the user instance's operating system to restore it to its initial state or troubleshoot system failures.

[0204] In some embodiments, the bare metal device includes a bare metal physical machine. In management mode, management operations are performed on user instances based on a management instance, including:

[0205] In management mode, the management instance controls the bare metal physical machine to perform management operations on user instances.

[0206] For example, a management instance can be controlled to access the host side of a bare-metal physical machine, thereby performing management operations on user instances. The host side typically refers to the physical machine's host environment, while the user instance is the operating system or computing environment running on the physical machine. By accessing the host side, the management instance indirectly operates on the user instance's resources (such as local disks), such as reinstalling the system or cleaning up the disk.

[0207] In some embodiments, controlling the switching of the operating mode from management mode to user mode includes:

[0208] Uninstall the management instance and switch its running state to shutdown.

[0209] Mount the system disk and network module corresponding to the user instance, and start the user instance.

[0210] Understandably, this releases resources used by the management instance to prepare for the mounting of resources for user instances. Specifically, unmounting the management instance's system disk (such as the management operating system image mgr OS) frees up storage resources used by the management instance. Unmounting the service network interface card used by the management instance disconnects it from the management network. Ensure that management instance resources no longer occupy the network interfaces of smart network interface cards or bare metal devices.

[0211] Understandably, switching the management instance's running state to the shutdown state shuts it down, releasing the computing and storage resources it occupies. This involves switching the management instance from its current running or operational state to the shutdown state. The shutdown state is the silent state of the management instance; in this state, the management instance does not occupy device resources but can be restarted if needed.

[0212] Here, the operating system and network resources of the user instance are mounted, preparing the user instance for operation. The system disk of the user instance (such as the cloud disk or local system disk required for the guest OS) is mounted, ensuring a correct connection between the operating system file system and the bare metal physical machine. Network components are mounted, allowing configuration of the network resources corresponding to the user instance (such as virtual switching networks, user network cards, etc.), ensuring that the user instance can communicate with the external network.

[0213] Here, the user instance is started, and the guest operating system is booted into running mode, so that the bare metal device is officially switched to user mode, ensuring that the user instance can take over the resources of the bare metal device and run user applications normally.

[0214] For example, such as Figure 3 As shown, Figure 3 This diagram illustrates the switching between management mode and user mode.

[0215] 1. User Mode (Left Side)

[0216] In user mode, bare metal devices are hosted by user instances to run the user's business operating system (guest OS) and applications.

[0217] The main modules include:

[0218] HOST: The instance operating system of the user running on the bare metal physical machine.

[0219] guest OS: The operating system of the user instance, such as Linux or Windows.

[0220] System disk: Mount the system disk corresponding to the user instance to store the operating system and necessary system files.

[0221] Data disk: Provides data storage functionality for user instances.

[0222] User network interface card (NIC): The network interface of a user instance, used to connect to external networks and enable network communication.

[0223] Admin (Management Network Interface): Provides an additional management network interface, but it is silent and does not participate in management operations.

[0224] BF2 / BF3 / DPU: Smart network interface cards (such as BF2, BF3, or DPU) are used to host network and data transmission resources for user instances.

[0225] When a user instance is running, the management instance is unloaded; the user's network card takes over network communication, and the management network (admin) is only used as a backup; the user instance has full control over the physical machine resources (CPU, memory, disk, network card).

[0226] 2. Management Model (Right Side)

[0227] In management mode, bare metal devices are hosted by management instances, which run on smart network interface cards (NICs) to perform device management tasks (such as operating system installation, resource configuration, and maintenance).

[0228] The main modules include:

[0229] HOST: The physical machine is in management mode and no user instances are running at the moment.

[0230] mgr OS (manager operating system) is responsible for device management tasks (such as installation tasks, hardware diagnostics, etc.).

[0231] System disk: The system disk component for managing instances, providing storage resources for managing the operating system.

[0232] Data disk: Used to manage instance data storage.

[0233] Service NIC: The network interface for the management instance, used to connect to the external management network and enable communication for management tasks.

[0234] Admin (Management Network Interface): Provides a dedicated network channel for management instances, ensuring isolation between the management network and the user network.

[0235] BF2 / BF3 / DPU: Smart NICs (BF2 / BF3 / DPU) manage the network and resources of the managed instance and perform management tasks.

[0236] The device is in management instance control mode; the service network interface card provides a management network interface and is physically isolated from the user network; the management instance runs on the smart network interface card and is responsible for performing operating system installation, maintenance and hardware resource management.

[0237] For example, controlling a bare metal device to switch from management mode to user mode, i.e., management mode → user mode, includes the following steps:

[0238] (1) After completing the management task (such as installation), the management instance resources are unloaded.

[0239] (2) Mount the system disk and network card of the user instance.

[0240] (3) Start the user instance and the device enters user mode.

[0241] In this way, the management instance enables dynamic switching between user mode and management mode, and efficiently completes maintenance tasks, ensuring flexible management of equipment resources and normal operation of user services.

[0242] In some embodiments, the method further includes:

[0243] If the management operation is confirmed to be complete, the control operation mode will switch from management mode to user mode.

[0244] Understandably, once the management operation is confirmed to be complete, the system will issue a command to switch the bare metal device's operating mode from the current management mode back to user mode. This switch signifies that the management instance's task has ended, and the bare metal device will be returned to the user instance for management.

[0245] In some embodiments, controlling the operation mode to switch from user mode to management mode includes: unloading a user instance and starting a management instance.

[0246] Here, when the bare metal physical machine is in user instance running mode, if a user needs to initiate operations involving local resources (such as local disk reinstallation, data cleanup, creating a local disk image, etc.), the steps are as follows:

[0247] (1) Unload the current user instance resources and mount the management instance resources. Start the management instance and the management instance status will enter the operation state.

[0248] (2) Access the HOST side of the bare metal physical machine through the management instance network, complete the corresponding business requirements HOST side operation, exit the management instance and enter the shutdown state, remount the user instance resources, power on the user instance, and restore the user instance running mode.

[0249] For example, such as Figure 3 As shown, User Mode → Management Mode.

[0250] User instances need to perform management tasks (such as data disk cleanup, local disk image creation, etc.). The device is either idle or scheduled for a new installation task.

[0251] Switchover operation: Unload the user instance operating system (guest OS). Load the management instance operating system (mgr OS). Configure the service network interface card (Service NIC) and the disk resources corresponding to the management instance. Thus, the bare metal device is taken over by the management instance and enters management mode.

[0252] In some embodiments, the method further includes:

[0253] In runtime, acquire maintenance instructions for bare metal equipment;

[0254] In response to maintenance instructions, maintenance operations are performed on bare metal equipment based on a management instance. These operations include at least one of the following:

[0255] Hardware quality inspection operations, hardware repair operations, and internal configuration operations of the operating system.

[0256] Here, in addition to creating and managing user instances based on management instances (manipulating bare metal local resources), it can also be used for hardware quality inspection and repair of bare metal servers authorized by OPs (Operations Personnel). Other business needs include user OS system-related settings such as HT (Hyper-Threading), NUMA (Non-Uniform Memory Access), and NPS (Node Performance Setting).

[0257] This application embodiment also provides a management system for a bare metal device, the bare metal device including a smart network interface card (NIC), and the management system including: a management device and a management instance component, the management device being connected to the management instance component, and the management device being used to execute the above-described management method, such as... Figure 4 As shown, the management instance components include: system disk component, network component, and execution component.

[0258] Here, the core task of the system disk component is to load the disk image of the management instance, which is a remote disk image of the trimmed smart network interface card.

[0259] Understandably, the system disk of the management instance is a remote local disk based on a smart network card. Its image has been specially adapted and trimmed to minimize the space occupied after optimization. It provides predefined operation and management functions to ensure the high performance and stability of the management instance and hardware resources.

[0260] Here, the system disk component is used to load the trimmed smart network interface card (NIC) remote disk image. Once the image is loaded, it is mounted to the virtual device on the system disk, providing a runtime environment for the management instance. This image contains a trimmed operating system and comes pre-installed with management tools. With the image loading and initialization process complete, the management instance starts and enters a running state.

[0261] Here, the network component establishes an internal network connection with the smart network card's internal network interface and an external network connection with the external network module. That is, the network component is responsible for establishing a connection with the smart network card's internal network interface, while simultaneously enabling communication with the external network through the external network module.

[0262] It is understandable that internal network connectivity refers to the interaction between the management instance and the internal resources of bare metal devices (such as bare metal physical machines). External network connectivity refers to the interaction between the management instance and external control platforms (such as Ironic, Nova, etc.) to receive commands and report status.

[0263] Here, the network component in this embodiment of the application uses an independent service account network card (admin IP), which can achieve physical isolation from the user instance network. This design effectively improves the security and reliability of the network and avoids conflicts or interference between the management network and the user business network.

[0264] Here, the execution component is the central hub for implementing bare metal device management tasks, and it is connected to the system disk component and the external network module, respectively.

[0265] Here, the execution component (nova-executor) connects to and interacts with the system disk component. It is used to call the system disk component to load the disk image of the management instance to start the management instance. This disk image is a trimmed remote disk image that provides the necessary management tools and installation programs.

[0266] The execution component connects to an external network module, thereby interacting with the outside world through the external network module (e.g., a control plane, virtualization platform, or cloud management platform for bare metal devices). In this embodiment, after starting the management instance, the execution component of the management instance can create user instances based on the management instance and perform operation management on the user instances.

[0267] Here, the execution component is also used, in runtime, in response to installation instruction information, to install the operating system (system OS) of the bare metal device based on the management instance to create a user instance.

[0268] For example, the system disk component stores the images and tools required for the management instance to run, and it forms the basis for the execution component to perform its tasks. During the installation process, the execution component needs to load a trimmed management instance image from the system disk component; and invoke the installation program within the image to install the operating system on the bare metal device.

[0269] Thus, this application proposes a novel management instance component architecture for bare metal devices, realizing full lifecycle management of user instances by management instances based on smart network interface cards (NICs), thereby improving the management efficiency and flexibility of bare metal servers. (1) By having the management instance component reside on the smart NIC, independent of the local hardware resources and external PXE environment of the bare metal device, the dependency on network boot service (PXE) is eliminated, breaking through the dependence of traditional technologies on PXE environment and cloud disk storage services. (2) The management instance component includes a system disk component, a network component, and an execution component, which can quickly load the trimmed and optimized remote disk image and directly start the management instance. The management instance responds to instructions through the execution component, completing the installation of the operating system, the creation and management of user instances locally on the bare metal device, realizing an efficient installation process for bare metal devices and greatly shortening the server delivery time. (3) Through the connection between the network component and the internal network interface of the smart NIC and the external network module, the management instance can realize dynamic control and efficient communication of the bare metal device, which not only improves the management efficiency of the bare metal device, but also enhances the network security and stability of the device.

[0270] In some embodiments, such as Figure 4 As shown, the bare metal device also includes: a bare metal physical machine, which is connected to an internal network interface, and the bare metal physical machine includes multiple local disks.

[0271] Understandably, the bare metal host is the core hardware of a bare metal server, connected to the internal network interface of a smart network card, and used to provide physical computing resources (such as CPU, memory, and disk).

[0272] Understandably, the bare metal physical machine is the object of management instance operations, and the smart network interface card (NIC) communicates with it through its internal network interface. Therefore, the execution component is also used to respond to management request information, operate the bare metal physical machine based on the management instance, and perform management operations on the user instance. This ensures that the management instance on the smart NIC can interact with the bare metal physical machine through the internal network interface to complete the operation and management tasks of the bare metal device.

[0273] It is understandable that a bare metal physical machine includes multiple local disks, that is, multiple physical storage devices that are directly connected to the physical machine's motherboard or storage controller, supporting data storage and operating system operation.

[0274] In some embodiments, the smart network interface card includes:

[0275] The first physical function module is connected to the internal network interface.

[0276] The second physical function module is connected to the system disk component.

[0277] It is understandable that the Physical Function (PF) module in a smart network interface card (NIC) is a hardware-level functional interface that directly maps to the host system to provide specific network or storage functions. Each PF module typically contains its own hardware resources (such as DMA queues) and independent configuration space.

[0278] Here, the internal network interface is the eth0 interface, which is the main network interface of the bare metal physical machine (host). The internal network interface is mainly used for communication between the management instance and the bare metal physical machine. (1) Hardware status detection: The management instance accesses the hardware information of the bare metal physical machine through the eth0 interface. (2) Operating system installation: In the installation task, the management instance interacts with the storage device of the physical machine through the eth0 interface to transfer installation data. (3) Operation and maintenance tasks: For example, the data flow of tasks such as disk cleanup and local disk image creation is completed through the internal network interface.

[0279] Here, the physical function module includes at least two PF modules. The first physical function module (PF1) is connected to the internal network interface (eth0 interface), providing a communication channel between the management instance and the bare metal physical machine. With the hardware support of PF1, the smart network card can efficiently process internal network data streams, achieving low-latency and high-bandwidth network communication.

[0280] Here, the second physical function module (PF2), connected to the system disk component, supports access to a trimmed remote disk image on the smart network interface card. For example, when starting the management instance, the trimmed image is loaded from the system disk component via PF2. During the operation of the management instance, management tools or configuration files in the system disk component are accessed via PF2.

[0281] In some embodiments, the smart network interface card (NIC) is a first smart NIC with the control channel not enabled, and the network components include:

[0282] The virtual switching network module is connected to both the first physical function module and the external network module.

[0283] Here, the smart NIC is the first smart NIC with its control channel disabled. The first smart NIC is a BF2 / BF3 smart NIC. Therefore, network communication needs to be implemented indirectly through virtualization. The network components include a Virtual Switch Network Module (BVS), which is connected to both the first physical function module and the external network module to manage traffic on both the internal and external networks. This Virtual Switch Network Module is connected to the first physical function module to distribute data traffic from internal network interfaces (such as eth0 of a bare-metal physical machine) to the external network module.

[0284] For example, for BF2 / BF3 cards, since the SOC-side control channel is not open, the management instance needs to connect to the outside world via the BVS virtual network. Therefore, this network component includes a virtual switching network module, which is connected to the first physical function module and the external network module respectively.

[0285] For example, when the execution component installs the operating system (System OS) on the bare metal device based on the management instance, it communicates with the bare metal physical machine through the first physical function module PF1 and BVS to complete the installation, partitioning and configuration of the operating system.

[0286] After the management instance is installed, the System OS runs on the bare metal physical machine, providing an operating environment for user instances to be created.

[0287] In some embodiments, the management instance component further includes:

[0288] The first backup storage module is connected to the system disk component. The first backup storage module includes a snapshot module or a block storage module.

[0289] Understandably, the first smart network interface card (BF2 / BF3 card) itself does not support a control channel and relies on virtualization modules (such as BVS) for network communication. This limits the data management capabilities of the management instance, thus requiring the first backup storage module to provide additional storage support.

[0290] Here, if the smart network card is the first smart network card, the management instance component also includes a first backup storage module. The first backup storage module includes a snapshot module or a block storage module and is connected to the system disk component to store snapshots and running data of the management instance.

[0291] In some embodiments, the smart network interface card (NIC) is a second smart NIC with an enabled control channel, the external network module is a virtualization tool (libvirt) module, and the network component is a third physical function module on the second smart NIC. The third physical function module is used to form a control channel between the first physical function module and the second physical function module. The control channel is used to establish an internal connection with the internal interface and an external network connection with the virtualization tool module.

[0292] As is understandable, a control channel is a dedicated communication channel for management tasks, used to connect internal interfaces (the management network of bare metal devices) to external networks (control platforms, virtualization tools, etc.). Through the control channel, management instances can directly access external networks without relying on complex virtual switching network modules (such as the BVS module in the first smart NIC).

[0293] Here, the second smart network card can be a DPU card. For a DPU card, the network component is the third physical function module on the DPU card. The network connection is made directly through the third physical function module on the card, namely the admin PF module (Administrative Physical Function).

[0294] Here, the third physical function module is used to establish an internal connection with the internal interface and an external connection with the external interface.

[0295] For example, the management instance directly accesses the bare metal physical machine resources through the admin PF module to complete the installation and initialization of the systemOS.

[0296] In some embodiments, the management instance component includes:

[0297] The second backup storage module is connected to the system disk component and is a block storage module.

[0298] Here, if the smart network card is the second network card and the second backup storage module is a block storage (blk) module connected to the system disk component, the block storage module, as a high-performance storage method, matches the high-bandwidth communication capability of the second smart network card, providing faster storage read and write speeds for the management instance.

[0299] The following is a detailed explanation of this application with reference to an application example.

[0300] like Figure 5 and Figure 6 As shown, Figure 5 and Figure 6The diagrams show the architecture of two bare-metal cloud server management instance components. Figure 5 A schematic diagram of the architecture of the EBC (Enhanced Baseboard Controller) management instance component for BF2 / BF3 smart network interface cards; Figure 6 This is a schematic diagram of the architecture of the EBBC (Enhanced Baseboard Bus Controller) management instance component for the DPU, where the smart network interface card (NIC) is used.

[0301] The architecture of these two management instance components will be explained in detail below.

[0302] I. The smart network card is the EBC management instance component of BF2 / BF3 (i.e., the aforementioned first smart network card).

[0303] like Figure 5 As shown, the main components of this architecture include: a management instance component, a smart network interface card (BF2 / BF3), a bare metal physical machine, and an external network module. The management instance component is located on the smart network interface card.

[0304] (1) Manage instance components.

[0305] 1. System disk component (admin img): Used to load and store the trimmed remote disk image.

[0306] 2. Network Components (BVS, Virtual Switch Network Module): Establishes internal network connections with the smart network card's internal network interface (eth0) and external network connections with the external network module. For example, in... Figure 5 In this configuration, the virtual switching network module is connected to both the first physical function module and the external network module.

[0307] 3. Execution component (nova-executor): Used to call the system disk component to load the disk image, start the management instance, and perform management tasks such as installing the operating system (system OS).

[0308] 4. Snapshot / Block Storage Module (snap / blk): Used for backup and storage management tasks, and connected to the system disk component.

[0309] (2) Smart network card.

[0310] 5. PF Module (Physical Function): This module includes a first physical function (PF1) module and a second physical function (PF2) module. These two physical function modules are responsible for communication with internal and external network modules. The management instance component implements internal and external communication of the bare metal device through the PF module and the BVS module.

[0311] (3) Bare metal physical mechanism.

[0312] 6. Disk: The disk is the local disk of the bare metal physical machine, used to store the operating system (systemos) and provide the data storage space required for user instances to run.

[0313] (4) External network module.

[0314] The external network modules include: ironic (bare metal resource management module), csoc (central scheduling and control module), and port (network interface module). Here, the port interface is connected to the virtual switching module, and ironic is connected to the aforementioned execution components.

[0315] In practical applications, the CSOC receives management commands from the upper-layer platform (such as Nova) and communicates with the CSOC (Central Dispatch and Control Module) through the Port interface to realize virtual switching network (BVS). The commands are then sent to the execution component, which executes the management commands and connects to the internal network interface (eth0) of the bare metal physical machine through the physical function module (PF) of the smart network card to complete the management operation of the bare metal device.

[0316] II. A schematic diagram of the structure of the EBBC (Enhanced Baseboard Bus Controller) management instance component of the smart network card as the DPU (i.e., the aforementioned second smart network card).

[0317] like Figure 6 As shown, the main components of this architecture include: a management instance component, a smart network interface card (DPU), a bare metal physical machine, and an external network module, wherein the management instance component is located on the smart network interface card.

[0318] In practical applications, for smart network interface cards (NICs) of the DPU type, such as Figure 6As shown, the execution component communicates directly with external network modules (such as Libvirt) through the AdminPF (Admin Physical Function) module (i.e., the aforementioned third physical function module) to receive management commands, thereby supporting efficient data transmission. This eliminates the need to rely on the virtual switching network module (such as BVS) of the BF2 / BF3 architecture. Furthermore, it communicates with the internal network interface (eth0) of the bare metal physical machine through the physical function module (PF) to achieve internal communication, ensuring that the execution component executes the management command and completes the management operations of the bare metal device.

[0319] The main modules involved in this architecture are described below.

[0320] (1) Manage instance components.

[0321] 1. System disk component (admin img): Used to load and store the trimmed remote disk image.

[0322] 2. The network component is the third physical function module (admin PF). The third physical function module is used to form a control channel between the first physical function module and the second physical function module. The control channel is used to establish internal connections with internal interfaces and external network connections with virtualization tool modules.

[0323] 3. Execution component (nova-executor): Used to call the system disk component to load the disk image, start the management instance, and perform management tasks such as installing the operating system (system OS).

[0324] 4. Block storage module (blk): Used for backup and storage management tasks, and connected to the system disk component.

[0325] (2) Smart network card.

[0326] 5. PF Module (Physical Function): This module includes a first physical function module (PF1) and a second physical function module (PF2). These two physical function modules are responsible for communication with internal and external network modules.

[0327] (3) Bare metal physical mechanism.

[0328] 6. Disk: The disk is the local disk of the bare metal physical machine, used to store the operating system (systemos) and provide the data storage space required for user instances to run.

[0329] (4) External network module.

[0330] The external network module is a virtualization tool module responsible for task scheduling and managing communication between instances and external platforms, and establishing external network connections with the third physical function module.

[0331] Thus, based on the two management instance architecture designs described above, the following can be achieved:

[0332] (1) System disk design: The system disk of the management instance is a remote local disk based on the smart network card. Its image has been specially adapted and trimmed to minimize the space occupied after optimization. It provides predefined operation and management functions to ensure the high performance and stability of the management instance and hardware resources.

[0333] (2) Network Design: For BF2 / BF3 cards, since the SOC-side management channel is not enabled, the management instance connects to the outside world via the BVS virtual network. For the self-developed Taihang DPU card, the network connection is directly established through the admin PF on the card. To ensure the network security of the management instance, the network card of the management instance uses an independent service account network card, which is isolated from the user instance network.

[0334] (3) Dynamic activation: By controlling the behavior of the management instance through state control, the management instance is dynamically activated when it is necessary to use the capabilities of the management instance (such as operating bare metal local resources). By utilizing the network of the management instance, the operation and management of server-side hardware resources can be directly initiated, ensuring the manageability and flexibility of bare metal resources.

[0335] The embodiments of this application will now be described in detail with reference to another application example.

[0336] Existing bare metal server management solutions suffer from shortcomings in stability, speed, scalability, and resource management, limiting their performance in efficient and flexible management. These deficiencies provide a good entry point for this solution. By combining smart network interface cards (NICs) and dynamic management instances, this solution aims to improve the management efficiency and flexibility of bare metal servers.

[0337] This application example proposes a method for managing bare metal operations based on management instances on smart network interface cards (NICs). It designs an efficient management mode lifecycle, system disk, and network mode management scheme, which differs from the traditional PXE-based approach. This avoids the cumbersome upfront processes required when operating local resources of bare metal physical machines, thereby improving overall management efficiency.

[0338] The EBC (Elastic Bare Metal Cloud Server) management instance in this application example is a resident instance on a bare metal physical machine, possessing multiple states (running, shut down, and operational), which can flexibly switch according to business scenarios to efficiently complete various business operations. The management instance is compatible with various bare metal machine models, including local system disk and cloud disk system disk models. Compared to previous bare metal servers based on cloud disk storage offloading, the management instance provides more operational and management space, enabling bare metal servers to support more user-defined operations and offering richer product capabilities.

[0339] Its main technical means include:

[0340] (1) Management instance architecture design, such as Figure 5 and Figure 6 As shown:

[0341] System disk design: The system disk of the management instance is a remote local disk based on the smart network card. Its image has been specially adapted and trimmed to minimize the space occupied after optimization. It provides predefined operation and management functions to ensure the high performance and stability of the management instance and hardware resources.

[0342] Network Design: For BF2 / BF3 cards, since the SOC-side management channel is not enabled, management instances connect to the outside world via the BVS virtual network. For the self-developed Taihang DPU card, network connections are made directly through the admin PF on the card. To ensure the network security of management instances, the network interface cards (NICs) of management instances use independent service account NICs, achieving isolation from user instance networks.

[0343] Dynamic startup: By controlling the behavior of the management instance through state control, the management instance is dynamically started when it is necessary to use the capabilities of the management instance (such as operating bare metal local resources). By utilizing the network of the management instance, operations and management of server-side hardware resources can be directly initiated, ensuring the manageability and flexibility of bare metal resources.

[0344] (2) Single machine multi-instance management, such as Figure 2 As shown,

[0345] Instance Metadata Center: The instance metadata center stores the status of management instances that reside on bare metal physical machines and dynamically modifies the status of management instances based on upstream business requests, thereby controlling the behavior of management instances.

[0346] State transitions and resource management:

[0347] When the bare metal physical machine is idle and not scheduled, the management instance is in standby mode. In this state, the management instance periodically reports heartbeat signals to the metadata center and periodically checks the hardware environment of the current physical machine through internal operation and maintenance scripts, reporting the machine's health status to the operation and maintenance center.

[0348] When a bare metal physical machine is scheduled and receives an upstream order request, the management instance switches to the "in operation" state and processes the different order types separately.

[0349] For cloud disk system disk orders: The management instance does not perform any additional operations, but waits for the user cloud disk system disk on the Cinder side to be created. At this time, the management instance will unmount its own system disk and service network card, and then the management instance will enter a shutdown silent state. Finally, it will mount the user instance resources and power on, and the machine will switch to user instance mode.

[0350] For local system disk orders: The management instance accesses the bare metal HOST side through its own network to complete the installation and related initialization operations on the local system disk. After the operation is completed, the management instance resources are unloaded, the management instance enters a shutdown silent state, and the user instance system is started, and the machine switches to user instance mode.

[0351] When a bare-metal physical machine is in user instance running mode, if a user needs to initiate operations involving local resources (such as local disk reinstallation, data cleanup, creating a local disk image, etc.), the steps are as follows:

[0352] Unmount the current user instance resources and mount the management instance resources. Start the management instance, and the management instance will enter the "operating" state. Access the bare metal physical machine's HOST side through the management instance network to complete the corresponding business operation requirements.

[0353] After the operation on the HOST side is completed, the management instance will exit and enter the shutdown state. The user instance resources will be remounted, the user instance will be powered on, and the user instance running mode will be restored.

[0354] (3) Manage instance mode switching, such as Figure 3 As shown:

[0355] Switching from user mode to management mode: Before a user instance needs to perform an operation, the control plane actively triggers a mode switch, including unloading the user's device on the smart card, mounting the management instance service network card and the management instance's remote local disk, and starting the management instance. At this time, the machine is managed by the management instance and is in management mode.

[0356] Switching from management mode to user mode: After the user instance operation is completed, the control plane actively triggers the mode switch, including unloading the management instance's device on the smart card, mounting the user instance's network card and system disk, and starting the user instance. At this time, the machine is managed by the user instance and is in user mode.

[0357] The following section explains the main operations for managing instances.

[0358] I. Local system disk installation and delivery under management instance.

[0359] (1) The management instance status changes to "Operating" and the management instance executes the process of creating bare metal.

[0360] (2) Log in to the server host side via the management instance IP.

[0361] (3) Download the image from Glance, partition and format the local disk on the host side, complete the image installation of the local system disk, set the system disk boot entry, HT / NUMA, etc.

[0362] (4) Clean up the network card and system disk of the management instance, shut down the management instance, and exit the management instance operating system.

[0363] (5) Mount the network card and cloud disk of the user instance (if the user has purchased a cloud disk in the order).

[0364] like Figure 7 As shown, Figure 7 This illustrates the process of installing a local system disk onto a bare metal device, where the operating system is directly installed onto the local disk of the bare metal device. The main modules involved include:

[0365] First, upper-layer service and control components.

[0366] 1. Nova-API: Provides the entry point for installation requests. It passes user order information to the scheduling module (Scheduler).

[0367] 2. ILO-API / IPMI: Provides hardware control functions for bare metal devices, including power management and status checks.

[0368] 3. CSOC: As the management and scheduling center, it is responsible for interacting with various components, including system switching and task scheduling.

[0369] 4. SNIC: Controls the network configuration and resource allocation of smart network interface cards (such as BF2, BF3, DPU).

[0370] 5. Ironic: Control node, used to manage the lifecycle of bare metal equipment, including installation, status management, resource allocation, etc.

[0371] Second, the bare metal physical host structure.

[0372] exist Figure 7 In the diagram, the physical machine is divided into two main areas:

[0373] 1. Management Mode (Service OS)

[0374] Service-img (Service Image): A management instance image loaded on the smart network interface card (BF2 / BF3 / DPU).

[0375] Service OS: An operating system that runs in management mode and is mainly used to perform management tasks, such as system installation, hardware detection, and image creation.

[0376] Snic-agent: Responsible for network management and resource scheduling of smart network interface cards.

[0377] Nova-agent: Responsible for interacting with Nova to control installation tasks and update status.

[0378] OVS / BVS: OVS (Open vSwitch): Provides network switching functionality. BVS: Virtual switching component used for network virtualization.

[0379] 2. User Mode (Guest OS)

[0380] Guest OS: The user instance operating system. After installation, the user instance enters the running mode.

[0381] Disk (data disk and system disk): Local disk resources, including the system disk (stores the operating system) and the data disk (stores user data).

[0382] The following describes the steps of the installation process.

[0383] Step 701: Check Hardware Status (Check ILO)

[0384] Here, the hardware information is obtained by calling the hardware control interface (such as ILO-API or IPMI) via CSOC. The hardware status of the bare metal device is checked to ensure it is in normal operating condition and ready for installation.

[0385] Step 702: Create User Resource

[0386] Here, Ironic Manager interacts with the network control service (Neutron Server) to create and allocate resources.

[0387] Step 703: Switch to Admin Mode

[0388] Switch the bare metal device from idle state to management mode and mount the management instance image (Admin IMG) to prepare for performing management tasks.

[0389] Step 704: Install the operating system (Install OS)

[0390] Install the user operating system on the local disk of the bare metal device. The management instance (Service OS) loads the installation script.

[0391] Step 705: Unplug Service Network

[0392] After completing the operating system installation, uninstall the service network card used by the management instance to release network resources and prepare to switch to user instance mode.

[0393] Step 706: Detach Service Volume

[0394] Release the service volume occupied by the management instance to free up disk resources and ensure that resources are available for user instances.

[0395] Step 707: Allocate User Resources

[0396] Allocate the necessary network and disk resources to user instances to ensure that user instances have the necessary operating environment, including: mounting user networks and allocating user disks (system disk and data disk).

[0397] Step 708: Execute the installation scripts (SSH Service OS to Execute Install Scripts)

[0398] Access the Service OS environment of the management instance via SSH, execute the installation script, and ensure that the operating system image is installed to the local disk.

[0399] Step 709: Download Guest OS.

[0400] Download the user's operating system image through the image storage service (Glance API) and store it on the local disk of the bare metal device.

[0401] Step 710: Reboot the device.

[0402] Reboot the bare metal device to load and start the user operating system (Guest OS).

[0403] Step 711: Callback of Task Status

[0404] The execution status of the installation task is sent back to the control node, the task result is updated, and the upstream service is notified that the installation is complete.

[0405] II. Releasing user instances under the management instance:

[0406] (1) Uninstall the user instance's main network card and system disk (if it is a cloud disk system disk);

[0407] (2) Mount the main network card and system disk of the management instance;

[0408] (3) Reboot into the management instance operating system, and the machine will switch to management instance mode;

[0409] (4) Execute the dd command within the management instance to clean up the local disk data on the server;

[0410] (5) Continue to release other resources of the user instance.

[0411] III. Creating an image of a user's local disk under a management instance:

[0412] (1) Uninstall the user instance's main network card and system disk (if it is a cloud disk system disk);

[0413] (2) Mount the main network card and system disk of the management instance;

[0414] (3) Reboot into the management instance operating system, and the machine will switch to management instance mode;

[0415] (4) The management instance status is switched to "operating" and the management instance executes the image creation process;

[0416] (5) Log in to the HOST side through the management instance network, compress the corresponding partition of the user instance, complete the format conversion and package it;

[0417] Upload the file to Glance as a mirror storage and modify the mirror status;

[0418] (6) Uninstall the primary network card and system disk of the management instance, and shut down the management instance;

[0419] (7) Mount back the user instance's primary network interface card and system disk (if it is a cloud disk system disk);

[0420] (8) Reboot into the user instance operating system, and the machine switches to user instance mode;

[0421] IV. Reinstalling the system on a user's local disk under a management instance:

[0422] (1) Uninstall the primary network interface card of the user instance;

[0423] (2) Mount the main network card and system disk of the management instance;

[0424] (3) Reboot into the management instance operating system, and the machine will switch to management instance mode;

[0425] (4) The management instance performs a system reinstallation process, connects to the HOST side via the management instance network, reformats the local disk partition, pulls the image from Glance and installs it to the local disk, sets the boot entry to uninstall the management instance's main network card and system disk, and shuts down the management instance.

[0426] (5) Mount back the primary network interface card of the user instance;

[0427] (6) Reboot into the user instance operating system and switch the machine to user instance mode.

[0428] Thus, this application example does not rely on a PXE environment or cloud storage services. Instead, it uses a management instance designed on the smart card to host the operation and behavior of bare metal, which is a new type of bare metal management strategy.

[0429] In order to implement the method of the embodiments of this application, the embodiments of this application also provide a management device for bare metal equipment. The management device for bare metal equipment corresponds to the above-mentioned bare metal management method. The steps in the above-mentioned bare metal management method embodiments are also fully applicable to the management device for bare metal equipment embodiments.

[0430] In some embodiments, the aforementioned management device may be a management apparatus for the bare metal device.

[0431] like Figure 8 As shown, the management device 800 for the bare metal device includes: a first control module 801, a switching module 802, and a second control module 803. The first control module 801 is used to start a management instance on the smart network card when the bare metal device is in management mode, thereby controlling the management instance's working state to switch from a power-off state to a standby running state. The switching module 802 is used to, in the running state, in response to installation instruction information, switch the working state from the running state to the operation state of performing operations; and in the operation state, install the operating system of the bare metal device based on the management instance to create a user instance of the bare metal device. The second control module 803 is used to, upon determining that the user instance creation is complete, control the operating mode to switch from management mode to user mode.

[0432] In some embodiments, the management device 800 further includes a first acquisition module 804, which is used to acquire management request information for a user instance in user mode; the second control module 803 is also used to control the operation mode to switch from user mode to management mode in response to the management request information, and to perform management operations on the user instance based on the management instance in management mode.

[0433] In some embodiments, the management device 800 further includes a determination module 805 for determining that the operating mode of the bare metal device is an empty machine mode; and a switching module 802 for switching the empty machine mode to the management mode in response to a first indication message.

[0434] In some embodiments, the determining module 805 is further configured to determine, in the running state, whether installation order information has been received, and if so, to generate and send installation instruction information.

[0435] In some embodiments, the determining module 805 is further configured to, in the running state, if installation order information is received, obtain the installation order type, which includes cloud disk system disk order type and local system disk order type; if the installation order type is determined to be local system disk order type, then generate and send installation instruction information.

[0436] In some embodiments, the determining module 805 is further configured to, in the running state, if no installation order information is received and the bare metal physical machine is not scheduled, obtain the heartbeat signal data of the management instance and the physical environment information of the smart network card; the management device further includes a sending module 806, configured to send the heartbeat signal data and the physical environment information.

[0437] In some embodiments, the second control module 803 is used to control the management instance to switch from running state to operating state in management mode; in operating state, in response to at least one of disk cleanup command, local disk image creation command and system reinstallation command, the second control module 803 performs at least one of disk cleanup operation, local disk image operation and operating system reinstallation operation on the user instance based on the management instance, so as to perform management operation on the user instance.

[0438] In some embodiments, the bare metal device includes a bare metal physical machine, and the second control module 803 is further configured to control the management instance to operate the bare metal physical machine in management mode to perform management operations on the user instance.

[0439] In some embodiments, the second control module 803 is further configured to unload the management instance and switch the running state of the management instance to the shutdown state; mount the system disk corresponding to the user instance and the network module corresponding to the user instance, and start the user instance.

[0440] In some embodiments, the second control module 803 is further configured to switch the operating mode from management mode to user mode if it is determined that the management operation is completed.

[0441] In some embodiments, the second control module 803 is further configured to unload the user instance and start the management instance.

[0442] In some embodiments, the management device further includes a second acquisition module 807, which is further configured to acquire maintenance instruction information for the bare metal equipment in a running state; the second control module 803 is further configured to perform maintenance operations on the bare metal equipment based on a management instance in response to the maintenance instruction information, the maintenance operations including at least one of the following:

[0443] Hardware quality inspection operations, hardware repair operations, and internal configuration operations of the operating system.

[0444] In practical applications, the first control module 801, the switching module 802, the second control module 803, the acquisition module 804, the determination module 805, the sending module 806, and the second acquisition module 807 can be implemented by a processor in the management device of the bare metal equipment. Of course, the processor needs to run a computer program in memory to implement its functions.

[0445] It should be noted that the management device for bare metal equipment provided in the above embodiments is only illustrated by the division of the above-described program modules when controlling bare metal. In actual applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. In addition, the management device for bare metal equipment and the bare metal management method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0446] Based on the hardware implementation of the above program modules, and in order to implement the method of the embodiments of this application, the embodiments of this application also provide a bare metal device. Figure 9 This is merely an exemplary structure of the bare metal device, not the entire structure; implementation is possible as needed. Figure 9 The diagram shows part or all of the structure. For example... Figure 9 As shown, the bare metal device 900 provided in this embodiment includes at least one processor 901, a memory 902, a user interface 903, and at least one network interface 904. The various components in the bare metal device 900 are coupled together via a bus system 905. It can be understood that the bus system 905 is used to implement communication between these components. In addition to a data bus, the bus system 905 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 9The general labeled all buses as Bus System 905.

[0447] The user interface 903 may include a monitor, keyboard, mouse, trackball, click wheel, buttons, touchpad, or touch screen.

[0448] The memory 902 in this embodiment is used to store various types of data to support the operation of the bare metal device. Examples of such data include any computer program used to operate on the bare metal device.

[0449] The bare metal control method for bare metal devices disclosed in this application can be applied to or implemented by processor 901. Processor 901 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the bare metal control method for bare metal devices can be completed by integrated logic circuits in the hardware of processor 901 or by instructions in software form. The processor 901 can be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 901 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules can be located in a storage medium, specifically memory 902. Processor 901 reads information from memory 902 and, in conjunction with its hardware, completes the steps of the bare metal control method for bare metal devices provided in the embodiments of this application.

[0450] In an exemplary embodiment, the bare metal device may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned method.

[0451] It is understood that memory 902 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or... Volatile memory can be random access memory (RAM), used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM). The memories described in the embodiments of this application are intended to include, but are not limited to, these and any other suitable types of memory.

[0452] In an exemplary embodiment, this application also provides a computer storage medium, specifically a computer-readable storage medium storing a computer program thereon. This computer program can be executed by a processor to complete the steps of the method described in this application embodiment. The computer-readable storage medium can be a memory such as ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0453] In an exemplary embodiment, this application also provides a computer program product, including a computer program that can be executed by a processor 901 of a bare metal device to complete the steps of the method of this application embodiment.

[0454] It should be noted that terms such as "first" and "second" are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0455] Furthermore, the technical solutions described in the embodiments of this application can be combined arbitrarily without conflict.

[0456] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0457] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for managing bare metal equipment, characterized in that, The bare metal device includes a smart network interface card, and the method includes: When the bare metal device is in management mode, the management instance on the smart network card is started to control the working state of the management instance to switch from the power-off state to the standby running state; In the running state, in response to the installation instruction information, the working state is switched from the running state to the operation state of performing operations; and in the operation state, the operating system of the bare metal device is installed based on the management instance to create a user instance of the bare metal device; Once the user instance is confirmed to have been created, the operating mode is switched from the management mode to the user mode.

2. The method according to claim 1, characterized in that, The method further includes: In the user mode, obtain management request information for the user instance; In response to the management request information, the operating mode is switched from the user mode to the management mode, and in the management mode, management operations are performed on the user instance based on the management instance.

3. The method according to claim 1, characterized in that, The method further includes: The operating mode of the bare metal equipment is determined to be the empty machine mode; In response to the first instruction, the idle machine mode is switched to the management mode.

4. The method according to claim 1, characterized in that, The method further includes: In the running state, it is determined whether an installation order information has been received. If so, the installation instruction information is generated and sent.

5. The method according to claim 4, characterized in that, The method further includes: In the running state, if the installation order information is received, the installation order type is obtained, which includes cloud disk system disk order type and local system disk order type; If the installation order type is determined to be a local system disk order type, then the installation instruction information is generated and sent.

6. The method according to claim 4, characterized in that, The method further includes: In the running state, if the installation order information is not received and the bare metal physical machine is not scheduled, then the heartbeat signal data of the management instance and the physical environment information of the smart network card are obtained; Send the heartbeat signal data and the physical environment information.

7. The method according to claim 2, characterized in that, The management request information includes at least one of the following: disk cleanup command, local disk image creation command, and system reinstallation command. In the management mode, the management operation performed on the user instance based on the management instance includes: In the management mode, the management instance is controlled to switch from the running state to the operating state; In the operating state, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, at least one of the disk cleanup operation, local disk image operation, and operating system reinstallation operation is performed on the user instance based on the management instance to perform management operations on the user instance.

8. The method according to claim 2, characterized in that, The bare metal device includes a bare metal physical machine. In the management mode, the management operation performed on the user instance based on the management instance includes: In the management mode, the management instance is controlled to operate the bare metal physical machine in order to perform management operations on the user instance.

9. The method according to claim 1, characterized in that, The control of switching the operating mode from the management mode to the user mode includes: Uninstall the management instance and switch the running state of the management instance to the shutdown state; Mount the system disk and network module corresponding to the user instance, and start the user instance.

10. The method according to claim 2, characterized in that, The method further includes: If the management operation is determined to be completed, the operating mode is switched from the management mode to the user mode.

11. The management method according to claim 2, characterized in that, The control of switching the operating mode from the user mode to the management mode includes: uninstalling the user instance and starting the management instance.

12. The method according to claim 1, characterized in that, The method further includes: In the operating state, acquire maintenance instruction information for bare metal equipment; In response to the maintenance instruction information, maintenance operations are performed on the bare metal equipment based on the management instance, and the maintenance operations include at least one of the following: Hardware quality inspection operations, hardware repair operations, and configuration operations within the operating system.

13. A management device for bare metal equipment, characterized in that, The bare metal device includes a smart network interface card (NIC), and the management device includes: The first control module is used to start the management instance on the smart network card when the bare metal device is in management mode, so as to control the working state of the management instance to switch from the power-off state to the standby running state; A switching module is used to switch the working state from the running state to the operation state of performing operations in response to installation instruction information in the running state; and in the operation state, to install the operating system of the bare metal device based on the management instance to create a user instance of the bare metal device; The second control module is used to determine that the user instance has been created, and then control the operation mode to switch from the management mode to the user mode.

14. The management device according to claim 13, characterized in that, The management device further includes: The first acquisition module is used to acquire management request information for the user instance in the user mode. The second control module is also configured to, in response to the management request information, control the operation mode to switch from the user mode to the management mode, and in the management mode, perform management operations on the user instance based on the management instance.

15. The management device according to claim 13, characterized in that, The management device further includes: The determination module is used to determine that the operating mode of the bare metal equipment is an empty machine mode; The switching module is also used to switch the idle machine mode to the management mode in response to the first instruction information.

16. The management device according to claim 13, characterized in that, The determining module is also used to determine, in the running state, whether installation order information has been received, and if so, to generate and send the installation instruction information.

17. The management device according to claim 16, characterized in that, The determining module is further configured to, in the running state, if the installation order information is received, obtain the installation order type, wherein the installation order type includes cloud disk system disk order type and local system disk order type; If the installation order type is determined to be a local system disk order type, then the installation instruction information is generated and sent.

18. The management device according to claim 16, characterized in that, The determining module is further configured to, in the running state, if the installation order information is not received and the bare metal physical machine is not scheduled, obtain the heartbeat signal data of the management instance and the physical environment information of the smart network card; The management device further includes: The transmitting module is used to transmit the heartbeat signal data and the physical environment information.

19. The management device according to claim 14, characterized in that, The management request information includes at least one of the following: disk cleanup command, local disk image creation command, and system reinstallation command. The second control module is also used to control the management instance to switch from the running state to the operating state in the management mode. In the operating state, in response to at least one of the disk cleanup command, local disk image creation command, and system reinstallation command, at least one of the disk cleanup operation, local disk image operation, and operating system reinstallation operation is performed on the user instance based on the management instance to perform management operations on the user instance.

20. The management device according to claim 14, characterized in that, The bare metal device includes a bare metal physical machine, and the second control module is further configured to control the management instance to operate the bare metal physical machine in the management mode in order to perform management operations on the user instance.

21. The management device according to claim 13, characterized in that, The second control module is also used to unload the management instance and switch the running state of the management instance to the shutdown state; Mount the system disk and network module corresponding to the user instance, and start the user instance.

22. The management device according to claim 14, characterized in that, The second control module is further configured to, if it is determined that the management operation is completed, control the operation mode to switch from the management mode to the user mode.

23. The management device according to claim 14, characterized in that, The second control module is also used to unload the user instance and start the management instance.

24. The management device according to claim 13, characterized in that, The management device further includes: The second acquisition module is used to acquire maintenance instruction information for bare metal equipment in the running state; The second control module is further configured to, in response to the maintenance instruction information, perform maintenance operations on the bare metal equipment based on the management instance, wherein the maintenance operations include at least one of the following: Hardware quality inspection operations, hardware repair operations, and configuration operations within the operating system.

25. A management system for bare metal equipment, the bare metal equipment including a smart network interface card, characterized in that, The management system includes: a management device and a management instance component, the management device being connected to the management instance component, the management device being configured to perform the method as described in any one of claims 1 to 12, and the management instance component comprising: The system disk component is used to load the disk image of the management instance, wherein the disk image of the management instance is a trimmed remote disk image of the smart network card; The network components establish internal network connections with the internal network interface of the smart network card and external network connections with the external network module, respectively. The execution component is further configured to invoke the system disk component to load the disk image in order to start the management instance.

26. The management system according to claim 25, characterized in that, The bare metal device further includes a bare metal physical machine, which is connected to the internal network interface and includes multiple local disks.

27. The management system according to claim 25, characterized in that, The smart network interface card includes: A first physical function module, which is connected to the internal network interface; The second physical function module is connected to the system disk.

28. The management system according to claim 27, characterized in that, The smart network interface card (NIC) is a first smart NIC with the control channel not enabled, and the network components include: The virtual switching network module is connected to both the first physical functional module and the external network module.

29. The management system according to claim 28, characterized in that, The management instance component also includes: A first backup storage module is connected to the system disk component, and the first backup storage module includes a snapshot module or a block storage module.

30. The management system according to claim 27, characterized in that, The smart network card is a second smart network card with an enabled control channel. The external network module is a virtualization tool module. The network component is a third physical function module on the second smart network card. The third physical function module is used to form the control channel between the first physical function module and the second physical function module. The control channel is used to establish an internal connection with the internal interface and an external network connection with the virtualization tool module.

31. The management system according to claim 30, characterized in that, The management instance component includes: The second backup storage module is connected to the system disk component, and the second backup storage module is a block storage module.

32. A bare metal device, characterized in that, include: A processor and memory for storing computer programs that can run on the processor, wherein, The processor, when running a computer program, performs the steps of the method according to any one of claims 1 to 12.

33. A storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

34. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

Images