Bare metal server deployment method and device

Abstract

The invention provides a bare metal server deployment method and device. The method comprises the following steps: for a bare metal server to be deployed, acquiring a basic mirror image corresponding to a CPU model of the bare metal server and installing the basic mirror image; discovering and identifying hardware components of the bare metal server to obtain hardware fingerprint information of the key hardware components; obtaining a drive program corresponding to the hardware fingerprint information from a preset drive warehouse; and installing the obtained drive program to a bare metal server. According to the method, the basic operating system can be automatically installed for the server according to the CPU model, then each hardware component of the server is automatically identified, and the corresponding driving program is installed, so that mirror image manufacturing for each hardware combination is avoided, the number of mirror images is greatly reduced, the labor cost and the time cost are remarkably reduced, and the storage space is saved; and meanwhile, the complexity of mirror image maintenance management is reduced, and the server deployment efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a method and apparatus for deploying bare metal servers. Background Technology

[0002] In current business scenarios centered on high-performance computing, artificial intelligence training, and computation, GPU bare-metal servers have become critical infrastructure due to their ability to provide dedicated hardware computing power and extreme I / O performance. The hardware architecture of these servers is highly heterogeneous. To adapt to different hardware combinations, traditional solutions generally adopt a static image pre-provisioning approach, that is, creating and storing a separate complete operating system image for each specific hardware configuration.

[0003] While this solution achieves basic system compatibility, it suffers from a rapid increase in the number of images. A large number of images not only consumes massive amounts of storage space and increases storage costs, but also makes image management complex and cumbersome, reducing the efficiency of server deployment and maintenance. Furthermore, when hardware is upgraded or new configurations are added, a complete set of corresponding images needs to be recreated. The frequent creation and maintenance of images for different hardware combinations also significantly consumes human and time resources. Summary of the Invention

[0004] This application provides a bare metal server deployment method and apparatus to solve the problems of redundant operating system images, wasted time and manpower costs, and low deployment efficiency of GPU bare metal servers.

[0005] Specifically, this application provides the following technical solution: In a first aspect, this application provides a bare-metal server deployment method, the method comprising: For the bare metal server to be deployed, obtain the base image corresponding to the CPU model of the bare metal server and install it; The hardware components of the bare metal server are discovered and identified, and the hardware fingerprint information of the key hardware components is obtained. Retrieve the driver program corresponding to the hardware fingerprint information from the preset driver repository; The obtained driver is installed on the bare metal server.

[0006] Secondly, this application provides a bare metal server deployment apparatus, the apparatus comprising: The first module is used to obtain and install a base image corresponding to the CPU model of the bare metal server to be deployed. The second module is used to discover and identify the hardware components of the bare metal server and obtain the hardware fingerprint information of the key hardware components. The third module is used to obtain the driver program corresponding to the hardware fingerprint information from a preset driver repository; The fourth module is used to install the acquired driver onto the bare metal server.

[0007] Thirdly, this application provides an electronic device, comprising: A memory, one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the method described above.

[0008] Fourthly, this application provides a computer-readable storage medium including computer instructions that, when executed on an electronic device, cause the electronic device to perform the method described above.

[0009] Fifthly, this application provides a computer program product that, when run on a computer, causes the computer to perform the method described above.

[0010] The technical solution provided in this application has the following beneficial effects: During server deployment, for the bare-metal server to be deployed, the system automatically retrieves and installs the corresponding base image based on the server's CPU model. After the base operating system is running, automated hardware discovery technology identifies the hardware fingerprint information of each key hardware component. Based on the identified hardware fingerprint information, the system matches and retrieves the corresponding driver from a pre-defined driver repository, and finally installs the driver into the server's base operating system, achieving dynamic and automated overlay of the driver layer.

[0011] Using the above method, the basic operating system is automatically installed on the server based on the CPU model. Then, the server's hardware components are automatically identified and the corresponding drivers are installed, avoiding the need to create images for each hardware combination. This significantly reduces the number of images, saves storage space, and lowers the complexity of image maintenance and management. When new hardware is deployed, there is no need to recreate a complete image, significantly reducing labor and time costs and improving server deployment efficiency.

[0012] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0014] Figure 1 A flowchart illustrating the bare metal server deployment method provided in this application embodiment; Figure 2 A schematic diagram of the framework of the bare metal server deployment device provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the embodiments of this application is only used to describe specific embodiments of this application and is not intended to limit this application.

[0016] GPU bare metal servers refer to physical servers equipped with graphics processing units (GPUs) that users can exclusively use in cloud platforms or data center environments. Their hardware configurations are highly customizable and heterogeneous, often including combinations of various CPUs, GPUs, RAID controllers, and high-speed network cards, and they run operating systems directly without virtualization layer intervention.

[0017] When dealing with the diverse operating system deployment requirements of hardware configurations, traditional solutions rely on static pre-built images. The core of this approach is to create a separate operating system image for each hardware combination, with each image pre-installed with drivers and system configurations specific to the hardware. During server deployment, the image matching the target server hardware is directly used for system installation. When new hardware combinations emerge, operations personnel need to manually install the operating system, install the corresponding drivers for each hardware type, and then package the entire installed system into a new image for later use.

[0018] This solution has the following drawbacks: 1) Image quantity expansion: The number of images grows exponentially with the number of hardware types. In scenarios that only include common hardware components such as CPU, GPU, RAID card and network card models, the required number of images can reach hundreds or thousands, such as 5 types of CPU × 4 types of GPU × 3 types of RAID × 6 types of network cards = 360 images.

[0019] 2) Severe resource waste: Creating separate images for each hardware combination leads to a large amount of repetitive work, resulting in a significant waste of human and time resources. Furthermore, when hardware is iterated or new configurations are added, most existing images are incompatible with the new hardware combinations, causing idle and wasted resources. A single image is approximately 10-20GB, and a cluster of thousands of machines results in petabyte (PB) level storage redundancy.

[0020] 3) Difficult to manage and maintain: The large number of images makes image management extremely complex. When searching, updating and maintaining images, operation and maintenance personnel need to spend a lot of time and energy, which reduces work efficiency.

[0021] 4) Low deployment efficiency: During server deployment, it is necessary to select the corresponding image from many images for the target hardware combination. If the corresponding image is not found, the operating system installation, driver adaptation and image packaging process must be performed from scratch, which seriously affects the server deployment efficiency and makes it difficult to support application scenarios that require rapid elastic expansion.

[0022] To address the aforementioned issues, this application provides a bare metal server deployment solution.

[0023] The practical application of this application will be described in detail below through specific embodiments.

[0024] This application provides a bare-metal server deployment method for installing operating systems and hardware drivers on bare-metal servers. Figure 1 As shown, the method may include the following steps: Step 110: For the bare metal server to be deployed, obtain the base image corresponding to the CPU model of the bare metal server and install it; Specifically, in response to the power-on or deployment command of the bare metal server, the CPU model of the server is detected; based on the detected CPU model, the corresponding base image is obtained; and the base image is installed on the bare metal server.

[0025] In this embodiment, the base image is a pre-built bootable operating system template file adapted to a specific CPU model or architecture. It integrates an operating system kernel optimized for that CPU, necessary system configurations, and optional (can be added on demand) application software (such as databases, middleware, etc.), but does not include drivers for other hardware components (such as GPUs, RAID, network cards, etc.). Its core lies in hardware driver decoupling, thereby ensuring that the system can boot and run normally on the target CPU platform while maintaining the lightweight nature of the image file and the flexibility of hardware adaptation.

[0026] Optionally, a base image repository is pre-built and maintained to store base images corresponding to various CPU models. Step 110 specifically obtains the base image corresponding to the CPU model of the bare metal server in the following way: based on the CPU model of the bare metal server, the corresponding base image is matched and obtained from the pre-set base image repository.

[0027] Specifically, in this embodiment, the base image can be installed on the bare metal server in the following way: the base image is written to the server's system storage device (such as a hard drive), and the server is configured to boot from the system storage device. Then, the server boots the base operating system from the system storage device according to the boot configuration.

[0028] In summary, step 110 creates the runtime environment of the base operating system by installing the base image, thereby providing a unified software foundation for subsequent dynamic driver installation based on hardware identification.

[0029] Step 120: Discover and identify the hardware components of the bare metal server to obtain the hardware fingerprint information of the key hardware components; After the server completes booting, the base operating system enters the running state and performs the automatic hardware discovery and identification process. The base operating system either has a pre-installed hardware discovery and fingerprinting module or loads it automatically upon first boot. This module uses an automated hardware discovery mechanism to identify key hardware components of the server, such as the GPU, RAID card, and network card, obtaining information for each key hardware component, such as its model, manufacturer, and unique device identifier, thus obtaining the hardware fingerprint information for each key hardware component.

[0030] The aforementioned key hardware components can be any hardware component other than the CPU, or any specified hardware component; this embodiment does not limit this.

[0031] In this embodiment, hardware fingerprint information refers to the identity information that can uniquely identify a single hardware component. This information is usually a unique identifier inherent to the device, such as the hardware serial number, PCIe device ID, USB product ID, or other unique identification code provided by the manufacturer.

[0032] Optionally, one or more of the following methods can be used to discover and identify key hardware components of the server (such as GPU, RAID card, and network card) and obtain information about each key hardware component: PCIe device enumeration, USB device identification, system management BIOS information reading, and Redfish remote management interface access.

[0033] Step 130: Obtain the driver program corresponding to the hardware fingerprint information from the preset driver repository; Specifically, one or more trusted, version-managed driver repositories are pre-built and maintained, storing drivers for various hardware components. Each driver repository maintains a mapping table of "hardware fingerprint information—driver," recording the association between each hardware fingerprint and its corresponding driver. Additionally, each driver includes its version identifier and a security check value (such as a digital signature or hash value) for integrity verification.

[0034] Based on the hardware fingerprint information identified in step 120, the driver repository is accessed through the software driver download module. The driver corresponding to the hardware fingerprint information is determined by querying the mapping table, and then the driver corresponding to the hardware fingerprint information is obtained from the repository.

[0035] Optionally, the driver can be used to perform integrity checks and digital signature verification using the security check value and digital signature included with the driver, to ensure its source is secure and its content is intact.

[0036] Step 140: Install the obtained driver onto the bare metal server.

[0037] Specifically, the driver is installed into the base operating system of the bare metal server. This process includes: copying the driver files to a specific directory of the operating system, registering it in the configuration database of the base operating system, and loading and managing it by the operating system's device manager or service manager.

[0038] Optionally, during the installation process, the installation status, version information, and installation logs of each driver are recorded to generate a traceable initial deployment record for system auditing or fault diagnosis. Specifically, the installation progress of each driver, configured system environment variables, and updated system registry or service configurations are monitored and recorded in real time. If an anomaly or failure is detected during installation, the system automatically records the error code and context information and generates an alarm for troubleshooting and repair.

[0039] Optionally, after all installation operations are completed, the system generates a structured, timestamped, hardware fingerprint-linked, and traceable second deployment record, which is used for subsequent system audits, compliance checks, or fault diagnosis and recovery.

[0040] After driver installation and system deployment are complete, continuous driver maintenance and updates can be performed. Specifically, the driver repository is automatically polled at preset intervals or when preset trigger conditions are met to check for updated releases of each driver. When an available update is detected, the new driver version is automatically downloaded, verified, and installed on the server according to preset upgrade strategies and parameter configurations, without changing the original deployment script and system base image. The corresponding "hardware fingerprint information—driver" mapping and system configuration are also updated synchronously. This ensures that drivers deployed on the server are continuously maintained and always synchronized with the latest stable versions, thereby improving the overall system security, compatibility, and automation.

[0041] As can be seen from the above technical solutions, this embodiment provides a bare metal server deployment method, which constructs a lightweight base image that only contains the operating system kernel and basic environment that match the CPU model and does not contain dedicated drivers for other hardware components. This base image is used as a general deployment template, thus getting rid of the static binding of heterogeneous hardware such as GPUs, RAID cards, and network cards.

[0042] During server deployment, for the bare-metal server to be deployed, the system automatically obtains and installs the corresponding base image based on the server's CPU model. After the base operating system is running, automated hardware discovery technology identifies the hardware fingerprint information of each key hardware component. A driver repository is pre-created, and a mapping table is used in the driver repository to maintain the mapping relationship between hardware fingerprint information and drivers. Therefore, during deployment, the system can obtain the corresponding drivers based on the identified hardware fingerprint information and install them into the server's base operating system, achieving dynamic and automated overlay of the driver layer. Furthermore, the system supports version monitoring of the driver repository and automatically updates server drivers when an update is detected.

[0043] By employing the above methods, this embodiment reduces the number of images from a product of hardware combinations (hundreds) to a linear number based on CPU models (several), significantly reducing storage space usage and simplifying image maintenance and management. It achieves automatic driver identification and adaptation, eliminating the need to create images for each hardware combination. Furthermore, it avoids recreating complete images when new hardware is deployed, significantly reducing labor and time costs and improving server deployment efficiency. Finally, the driver update mechanism ensures that the production environment drivers are up-to-date, compatible, stable, and secure, mitigating the risk of system failures due to outdated drivers.

[0044] Based on the same inventive concept, this application also provides a bare metal server deployment device, the structural schematic diagram of which is shown below. Figure 2 As shown, it specifically includes: The first module 210 is used to obtain and install a base image corresponding to the CPU model of the bare metal server to be deployed. The second module 220 is used to discover and identify the hardware components of the bare metal server and obtain the hardware fingerprint information of the key hardware components. The third module 230 is used to obtain the driver program corresponding to the hardware fingerprint information from a preset driver repository; The fourth module 240 is used to install the acquired driver to the bare metal server.

[0045] Optionally, the device further includes: The fifth module is used to obtain and install the updated version of the driver from the driver repository when an updated version of the installed driver is detected in the driver repository.

[0046] Optionally, the device further includes: The sixth module is used to detect the CPU model of the bare metal server in response to the power-on command or deployment command of the bare metal server.

[0047] Optionally, the device further includes: The seventh module is used to determine the driver corresponding to the hardware fingerprint information by querying the mapping table of the driver repository, wherein the mapping table is used to record the one-to-one mapping relationship between the hardware fingerprint information and the driver.

[0048] Optionally, the device further includes: The eighth module is used to perform integrity verification and digital signature verification on the driver using the security check value included with the driver.

[0049] This embodiment provides a bare metal server deployment device, which corresponds to the aforementioned method. For specific implementation details, please refer to the aforementioned method embodiment, which will not be repeated here.

[0050] This application also provides a bare metal server deployment system. The system includes: a deployment server, a driver repository, a base image repository, and at least one bare metal server to be deployed.

[0051] The deployment server communicates with the driver repository, base image repository, and bare metal server. The deployment server further includes: a hardware auto-discovery and fingerprinting module, a software driver download module, a software driver installation module, and a periodic monitoring module.

[0052] The aforementioned deployment server can obtain a base image corresponding to the CPU model of the bare metal server to be deployed and install it onto the bare metal server.

[0053] The aforementioned hardware auto-discovery and fingerprinting module runs after the base image is installed and started. It is used to automatically identify the hardware components of the server and generate hardware fingerprint information for each hardware component.

[0054] The aforementioned software driver download module connects to a pre-set driver repository, which stores drivers corresponding to different hardware fingerprint information. Based on the identified hardware fingerprint information, this module retrieves the corresponding driver from the repository and automatically performs integrity and security checks upon download completion to ensure the driver is reliable and has not been tampered with.

[0055] The aforementioned software driver installation module is used to install verified drivers onto the server's base operating system. This module monitors the installation process in real time; if any abnormalities occur, it records and outputs an error log for maintenance personnel to troubleshoot and repair.

[0056] The aforementioned periodic monitoring module is used to periodically check the version updates of drivers in the driver repository. This module supports automatic or manual driver upgrades via configuration parameters, ensuring that server drivers are up-to-date and stable without requiring modifications to deployment scripts.

[0057] This application provides an electronic device that may include a memory and one or more processors. The memory stores computer program code, including computer instructions. When the processor executes the computer instructions, the electronic device can perform various functions or steps of the above-described method embodiments.

[0058] The structure of this electronic device can be referenced. Figure 3 The structure of the electronic device 100 shown.

[0059] For example, the processor mentioned above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0060] This application also provides a computer-readable storage medium including computer instructions that, when executed on an electronic device, cause the electronic device to perform the various functions or steps of the above method embodiments.

[0061] The aforementioned computer-readable storage media include, but are not limited to, any of the following: USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and other media capable of storing program code.

[0062] This application also provides a computer program product that, when run on a computer, causes the computer to perform various functions or steps of the above method embodiments.

[0063] The electronic devices, computer-readable storage media, and computer program products provided in this application embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0064] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various variations or substitutions can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for deploying a bare-metal server, characterized in that, The method includes: For the bare metal server to be deployed, obtain the base image corresponding to the CPU model of the bare metal server and install it; The hardware components of the bare metal server are discovered and identified, and the hardware fingerprint information of the key hardware components is obtained. Retrieve the driver program corresponding to the hardware fingerprint information from the preset driver repository; The obtained driver is installed on the bare metal server.

2. The method according to claim 1, characterized in that, The method further includes: In response to power-on or deployment commands from the bare metal server, detect the CPU model of the bare metal server.

3. The method according to claim 1, characterized in that, The method further includes: By querying the mapping table of the driver repository, the driver corresponding to the hardware fingerprint information is determined. The mapping table is used to record the one-to-one mapping relationship between hardware fingerprint information and driver programs.

4. The method according to claim 1, characterized in that, The method further includes: Use the security check value included with the driver to perform integrity verification and digital signature verification on the driver.

5. The method according to claim 1, characterized in that, The method further includes: When an updated version of an installed driver is detected in the driver repository, the updated driver is retrieved from the driver repository and installed.

6. A bare-metal server deployment device, characterized in that, The device includes: The first module is used to obtain and install a base image corresponding to the CPU model of the bare metal server to be deployed. The second module is used to discover and identify the hardware components of the bare metal server and obtain the hardware fingerprint information of the key hardware components. The third module is used to obtain the driver program corresponding to the hardware fingerprint information from a preset driver repository; The fourth module is used to install the acquired driver onto the bare metal server.

7. The apparatus according to claim 6, characterized in that, The device further includes: The fifth module is used to obtain and install the updated driver from the driver repository when an updated version of the installed driver is detected in the driver repository.

8. An electronic device, characterized in that, include: A memory, one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the method as described in any one of claims 1-5.

9. A computer-readable storage medium comprising computer instructions, characterized in that, When the computer instructions are executed on the electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-5.

10. A computer program product, characterized in that, When the computer program product is run on a computer, it causes the computer to perform the method as described in any one of claims 1-5.

Images