Server information acquisition and updating method and device, system, medium, product and equipment

Abstract

The invention discloses a server information collecting and updating method and device, a server information collecting and updating system, a medium, a product and equipment. The method comprises the steps of obtaining a target mirror image; the obtained target mirror image is loaded in a target server, and the target server comprises a neural network processor NPU; the hardware information of the target server is acquired through the loaded target mirror image, and the hardware information comprises the network card information of the NPU, so that the acquisition efficiency and accuracy of the server hardware information can be improved, the construction efficiency and expandability of a large-scale cluster are improved, and the operation and maintenance complexity of the cluster is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and particularly relates to a server information acquisition and updating method and device, system, medium, product and equipment. BACKGROUND

[0002] In a cloud computing or computing power network, a large number of servers are usually deployed to form a large-scale cluster to provide computing resources, so as to facilitate users to use the computing resources. However, such a large-scale cluster often needs to update the servers contained therein (for example, replace, reduce or increase the servers), and after each update, it is necessary to acquire the relevant parameter information of the updated servers in the cluster to ensure the normal use of the cluster.

[0003] In the related art, the relevant parameter information of the updated servers is usually acquired by using manual work, which is low in efficiency and low in accuracy, thereby seriously reducing the construction efficiency and expandability of the large-scale cluster and increasing the operation and maintenance complexity of the cluster. SUMMARY

[0004] To solve the above technical problems, the embodiments of the present application provide a server information acquisition and updating method and device, system, medium, product and equipment, which can improve the acquisition efficiency and accuracy of server hardware information, thereby improving the construction efficiency and expandability of the large-scale cluster and reducing the operation and maintenance complexity of the cluster.

[0005] In a first aspect, the embodiments of the present application provide a server information acquisition method, comprising: acquiring a target image; loading the acquired target image in a target server, wherein the target server comprises a neural network processor (NPU); acquiring hardware information of the target server through the loaded target image, wherein the hardware information comprises network card information of the NPU.

[0006] Optionally, the acquiring the target image comprises: acquiring a temporary network address after the target server enters a specified startup mode; acquiring the target image based on the temporary network address.

[0007] Optionally, the acquiring the hardware information of the target server through the loaded target image comprises: acquiring the hardware information of the target server by running a preset proxy service in the loaded target image.

[0008] Optionally, the preset agent service is configured to invoke a network management tool to collect the hardware information of the target server during its running process, the network management tool corresponding to the NPU.

[0009] Optionally, the target server is a bare metal server.

[0010] Optionally, the hardware information further includes at least one of: CPU information of a central processing unit (CPU); memory information; disk information.

[0011] Optionally, the network card information includes link layer discovery protocol (LLDP) information of the NPU, the number of the NPU is multiple, and each NPU communicates via a parameter plane network, the method further comprising: determining network topology parameters related to the parameter plane network based on the LLDP information; sending the network topology parameters to a management and control node corresponding to the target server.

[0012] Optionally, the network topology parameters include at least one of: media access control (MAC) address of a network card of the NPU; parameter information of a switch associated with the network card of the NPU; interface parameter information of the network card of the NPU.

[0013] Optionally, the network topology parameters are used to instruct the management and control node to perform a preset operation on the target server, the preset operation including at least one of: node information management, port information management.

[0014] In a second aspect, an embodiment of the present application provides a server information updating method, applicable to a management and control node, the method comprising: receiving an inspection request for a target server, wherein the target server includes an NPU; determining a server self-check task in response to the inspection request; controlling the target server to enter a specified startup mode based on the server self-check task, wherein the target server is configured to: after entering the specified startup mode, obtain a target image; load the obtained target image; and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU; receive network topology parameters determined according to the hardware information from the target server, and perform a preset operation on the target server according to the network topology parameters, the preset operation including at least one of node information management and port information management.

[0015] In a third aspect, an embodiment of the present application provides a server information collection system, including: a management node configured to receive an inspection request for a target server, determine a server self-check task in response to the inspection request, control the target server to enter a specified startup mode based on the server self-check task, and the target server including an NPU, the target server being configured to, after entering the specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU.

[0016] In a fourth aspect, an embodiment of the present application provides a server information collection device, including: an image acquisition module configured to acquire a target image; an image loading module configured to load the acquired target image in a target server, wherein the target server includes an NPU. a collection module configured to collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU.

[0017] In a fifth aspect, an embodiment of the present application provides a server information updating device suitable for a management node, including: a request receiving module configured to receive an inspection request for a target server, wherein the target server includes an NPU. a task determining module configured to determine a server self-check task in response to the inspection request. a mode control module configured to control the target server to enter a specified startup mode based on the server self-check task, wherein the target server is configured to, after entering the specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. a management module configured to receive network topology parameters determined according to the hardware information from the target server, and perform a preset operation on the target server according to the network topology parameters, the preset operation including at least one of node information management and port information management.

[0018] In a sixth aspect, an embodiment of the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, the computer program, when executed by a processor, implements the steps of the method according to any one of the preceding aspects.

[0019] In a seventh aspect, an embodiment of the present application provides a computer program product comprising computer instructions, the computer instructions, when executed by a processor, implement the steps of the method according to any one of the preceding aspects.

[0020] In an eighth aspect, an embodiment of the present application provides a computer device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, the processor, when executing the computer program, implements the steps of the method according to any one of the preceding aspects.

[0021] To sum up, the embodiments of the present application have at least the following beneficial effects: By using the embodiments of the present application, the target image obtained is loaded in the target server comprising an NPU, so as to realize automatic collection of hardware information of the target server through the target image loaded in the target server, thereby improving the collection efficiency of server hardware information, and in particular, the NPU network card information which has a greater impact on the server and large-scale cluster performance is also collected, thereby improving the collection accuracy of key server hardware information, and further improving the construction efficiency, scalability of the large-scale cluster, and reducing the operation and maintenance complexity of the cluster. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 FIG. 1 is a flowchart of a server information collection method provided by an embodiment of the present application; Figure 2 FIG. 2 is a schematic diagram of server information update provided by an embodiment of the present application; Figure 3 FIG. 3 is a flowchart of a server information update method provided by an embodiment of the present application; Figure 4 FIG. 4 is a schematic diagram of a server information collection system provided by an embodiment of the present application; Figure 5 FIG. 5 is a schematic diagram of a server information collection device provided by an embodiment of the present application; Figure 6 FIG. 6 is a schematic diagram of a server information update device provided by an embodiment of the present application; Figure 7 FIG. 7 is a schematic diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] With reference to the drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below. Obviously, the described embodiments / examples are only a part of the embodiments / examples of the present application, but not all the embodiments / examples of the present application. Based on the embodiments / examples in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort are within the scope of the present application.

[0024] In the description of the present application, the terms "first", "second", "third", etc. are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second", "third", etc. can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specified. In the description of the present application, the term "comprising" and its variants are open inclusion, i.e. "including but not limited to". The term "based on" is "at least partially based on". The term "according to" is "at least partially according to". The term "one embodiment / example" means "at least one embodiment / example"; the term "another embodiment / example" means "at least one other embodiment / example"; the term "some embodiments / examples" means "at least some embodiments / examples".

[0025] In the description of the present application, it should be noted that, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood in a broad sense, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through intermediate medium, or internal communication of two elements. For a person of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0026] In the description of the present application, it should be noted that, unless otherwise defined, all the technical and scientific terms used in the present application have the same meaning as that understood by a person skilled in the art. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments, and are not intended to limit the present application. For a person of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0027] Some terms and concepts related to the embodiments of the present application are explained below.

[0028] A neural processing unit (NPU) is a special-purpose hardware processor designed specifically to accelerate artificial intelligence (AI) (e.g., neural networks in deep learning) computations.

[0029] Link layer discovery protocol (LLDP) information refers to a set of data describing the identity and capabilities of a network device, which is actively sent and received by the network device through the link layer discovery protocol.

[0030] Ironic refers to a bare-metal management service in the OpenStack architecture. The control plane components of Ironic are generally deployed in the management nodes used to manage and control corresponding servers. For example, ironic-api is the external interface gateway of Ironic, which receives all bare-metal operation requests (such as inspect, deploy, and delete) and forwards them to the backend components. Ironic-conductor is the task scheduling core of Ironic, which is responsible for executing specific bare-metal operations (such as controlling server power on / off and coordinating the inspect process). Ironic-inspector is a dedicated service for detecting hardware information of bare-metal, which can be used to trigger PXE boot, receive hardware data reported by ironic-python-agent, and process and standardize the information.

[0031] In some example scenarios, with the continuous popularity of cloud computing, application to the cloud has become a trend. In the field of cloud computing, infrastructure as a service (IaaS) technology has been very mature and has been widely applied in the industry. In some cases, users may need more control, more hardware access, higher performance, higher security isolation, and / or the ability to select an operating environment. Bare-metal servers can provide tenants (users) with a close-to-native computing capability experience, which can be used to make up for the disadvantage of excessive loss of computing performance of virtualized instances in related technologies. Therefore, in cloud computing and computing power networks, bare-metal servers are provided as an important basic computing resource for users.

[0032] In the related art, in the deployment of bare metal servers, the configuration of the parameter plane network is heavily dependent on manual operation, and generally requires hardware integration (HI) personnel to manually collect the physical information of each NPU network card during the construction phase and form a document. Then, according to the document, software integration (SI) personnel manually create bare metal node-ports to enter the information into the Ironic bare metal management database one by one. When the bare metal server is opened, the Ironic service passes the pre-entered hardware information as parameters to the Neutron network service to complete the configuration of the parameter plane network. The above parameter plane network (also known as management network or control network) can be used to provide a stable, isolated and low-latency communication channel for servers and / or devices (such as NPUs) in the servers, and can also be used to manage and control the servers at various stages of their life cycle.

[0033] However, the above related technology is highly dependent on manual operation and is inefficient, and HI and SI personnel need to repeatedly collect and enter information every time a new device is deployed, which is a cumbersome process that seriously hinders the construction efficiency of large-scale clusters. Moreover, since a server may contain multiple NPU cards, any error in the entry of information for any NPU card can cause the entire machine parameter plane network configuration to fail, in other words, the accuracy of manual entry cannot be guaranteed, and the error rate is high, and the reliability is poor. In addition, if the corresponding server fails to open or the network is abnormal due to information errors, problem positioning at this time will be extremely difficult, requiring cross-team (HI / SI / operation and maintenance) repeated verification, which is time-consuming and labor-intensive, seriously affecting the construction period and operation and maintenance experience, increasing the operation and maintenance complexity. Further, since the manual method is difficult to adapt to the needs of rapid expansion and contraction and automated operation and maintenance, it affects the scalability of the cluster / server.

[0034] In a first aspect, see Figure 1 A flowchart of a server information collection method provided by an embodiment of the present application is shown, which can be applied to a target server. The method includes S101-S103, and the details are as follows.

[0035] S101, obtaining a target image.

[0036] In some examples, the target image can be obtained by downloading related files of the target image.

[0037] In some examples, the above target image can be a slim Linux system, which can be a micro operating system for hardware self-checking.

[0038] S102, loading the obtained target image in a target server, wherein the target server comprises a neural network processor (NPU).

[0039] In some examples, one or more NPUs can be integrated in the target server.

[0040] In some examples, the target server can be configured to automatically obtain and load the target image after a pre-boot execution environment (PXE) boot. The PXE boot can be implemented by a management node by invoking a related service. For example, the management node can control the target server to enter the PXE boot through an out-of-band management tool in response to an inspection request for the target server received via a related application programming interface (API) after receiving an administrator input.

[0041] S103, collecting hardware information of the target server through the loaded target image, wherein the hardware information comprises network card information of the NPU.

[0042] In some examples, the obtained target image can be directly loaded into the target server, so that the target server can perform self-inspection (hardware self-inspection) through the loaded target image, thereby completing automatic collection of hardware information of the target server. Here, the target image can be integrated with some self-inspection programs for controlling hardware self-inspection, which can be configured to automatically run in the loaded target image, thereby realizing automatic collection of hardware information of the target server.

[0043] In some examples, the loaded target image can run in the memory of the target server. In this embodiment, the target image runs in the memory, and thus does not depend on the local disk system of the target server, so as to be able to run on a brand-new server without deploying an operating system, and to avoid interference with a server with a deployed system. This feature also enables the target image to be applied in batch collection of hundreds or thousands of servers, to support rapid construction of large-scale clusters. Meanwhile, the collection process is fully automated, reducing operation and maintenance costs (such as fault diagnosis and configuration synchronization) caused by manual intervention, and reducing the operation and maintenance complexity of the cluster.

[0044] In an optional embodiment, the obtaining the target image comprises: obtaining a temporary network address after the target server enters a specified boot mode; obtaining the target image based on the temporary network address.

[0045] In some examples, the specified boot mode can include the PXE boot described above.

[0046] In some examples, the temporary network address can include a temporary Internet Protocol (IP) address, so that the target server can obtain the temporary IP address from a Dynamic Host Configuration Protocol Server (DHCP-Server) (see, for example, Figure 2 by sending an address acquisition request to the DHCP-Server to request to obtain the temporary network address, where the DHCP-Server can be used to provide DHCP service for the dnsmasq (a lightweight, open source software integrated with DHCP and TFTP functions) of ironic.

[0047] In some examples, see Figure 2 After obtaining the temporary network address, the target server can be regarded as a node in the network, and can then obtain the target image from a Trivial File Transfer Protocol (TFTP) server through the temporary network address.

[0048] In the present embodiment, the program code related to the specified boot mode set herein plays the role of an image boot file, and can automatically guide the target server to obtain and load the target image when the file is called, thereby realizing automated booting without human intervention, enabling the target server to automatically load the target image from the network without human operation, skipping the step of traditional manual booting, and completely automating the process to avoid the efficiency loss of manual operation in a large-scale cluster, thereby directly improving the collection efficiency. In addition, the obtained target image is pre-associated with defined startup parameters (such as image loading path, temporary network configuration, etc.), to ensure that all target servers enter the collection environment with the same process and parameters, which avoids configuration deviations (such as incorrect booting mode and missing parameters) that can occur in manual operation, and lays a foundation for the accuracy of subsequent hardware information collection.

[0049] In an optional implementation, the collecting hardware information of the target server through the loaded target image includes: collecting hardware information of the target server by running a preset agent service in the loaded target image.

[0050] In some examples, the default proxy service may include an Ironic Python Agent (IPA) service, which, when run, can automatically collect hardware information of the target server. The loaded target image can automatically run this default proxy service on it.

[0051] In one optional implementation, the preset proxy service is used to invoke a network management tool during its operation to collect hardware information of the target server, the network management tool corresponding to the NPU.

[0052] In some examples, the network management tool can interact with the network controller on the NPU, actively sending LLDP packets (i.e., LLDP requests) into the network, and receiving LLDP responses carrying the LLDP information from peer devices (such as the switch associated with / communicated with the aforementioned NPU), thereby collecting the NPU's network interface card (NIC) information. For example, the NPU may include an Ascend NPU card, such as the Ascend 910B. Thus, the network management tool invoked by the default proxy service may include the hccn_tool tool for the Ascend NPU card, which can interact with the network controller on the NPU card and actively send LLDP packets into the network. It is easy to understand that, as described above, the network management tool here can actually be determined by the NPU (e.g., by the type and model of the NPU).

[0053] In one alternative implementation, the target server is a bare metal server.

[0054] It should be noted that a bare metal server is a physical dedicated server that runs directly on the hardware, generally without relying on any virtualization layer (such as a hypervisor), and does not share underlying computing resources with other tenants. In this embodiment, compared to traditional physical servers, bare metal servers represent a modern evolution of traditional physical servers in the cloud computing era, significantly improving performance, security, and programmable management capabilities.

[0055] In one alternative implementation, the hardware information further includes at least one of the following: Central Processing Unit (CPU) information; in some examples, this CPU information may include key performance parameters such as CPU model, architecture, number of cores and / or frequency.

[0056] Memory information; in some examples, this memory information may include information such as the capacity, type, speed, and / or slot distribution of physical memory.

[0057] Disk information. In some examples, the disk information can include information such as type, capacity, interface, and / or health status of the local storage device.

[0058] In an optional implementation, the NIC information includes Link Layer Discovery Protocol (LLDP) information of the NPUs, the NPUs are multiple in number, and the NPUs communicate with each other via a parameter plane network. The method further includes: determining network topology parameters related to the parameter plane network based on the LLDP information; sending the network topology parameters to a management and control node corresponding to the target server.

[0059] In some examples, the LLDP information of the NPUs can be collected by actively sending an LLDP packet (i.e., an LLDP request) to the network and receiving an LLDP response carrying the LLDP information from a peer device (e.g., a switch associated with / communicatively connected to the NPU).

[0060] In some examples, the LLDP information can be parsed to specific and corresponding network topology parameters.

[0061] In an optional implementation, the network topology parameters include at least one of: a Media Access Control (MAC) address of a NIC of the NPU; in some examples, the MAC address is also referred to as a LAN Address, a MAC Address, an Ethernet Address, a Hardware Address, or a Physical Address. The MAC address actually refers to an adapter address or an adapter identifier EUI-48, which is used to indicate the address of a network device position. The MAC address can be used to uniquely identify a NIC in a network. If a device has one or more NICs, each NIC needs and will have a unique MAC address.

[0062] parameter information of a switch associated with the NIC of the NPU; in some examples, the switch can include an uplink switch associated with the NIC of the NPU, and the parameter information of the switch can include at least one of the following: information of a port where a patch cord of the switch is located, a name of the switch, and a management IP address of the switch.

[0063] interface parameter information of the network card of the NPU. In some examples, the interface parameter information can include Port VLAN ID (PVID) configuration information of the interface of the network card of the NPU, wherein VLAN refers to Virtual Local Area Network, and the PVID configuration is used to set PVID for the physical network port connected to the NPU (Neural Network Processor) network card, so as to ensure that the port can correctly transmit and receive data frames of a specified VLAN when accessing the switch.

[0064] In an optional implementation, the network topology parameter is used to instruct the management and control node to perform a preset operation on the target server, and the preset operation includes at least one of the following: node information management, port information management.

[0065] In some examples, after the network topology parameter is acquired, the network topology parameter can be reported to the management and control node, for example, refer to Figure 2 The IPA service (through API call) in the target server can be caused to call the interface of the ironic-inspector to implement information reporting of the network topology parameter, and the ironic-inspector in the management and control node can further call the ironic-api through the ironicclient to update or create the management database for the node (node) and / or port (port) and / or port group (portgroup), so as to complete the node information management and / or the port information management. It should be noted that the creation in the embodiment is a self-discovery process, and the node can not need to be created in advance, but only needs to be set to start the bare metal server from PXE.

[0066] In this way, the embodiment can realize the automatic registration / update of the parameter plane network related information in the management database.

[0067] In a second aspect, refer to Figure 3 , a flowchart of a server information updating method provided by an embodiment of the application is shown, and the server information updating method is applicable to a management and control node, and the method includes S301-S304, and specifically as follows.

[0068] S301, receiving a check request for a target server, wherein the target server includes an NPU; S302, in response to the check request, determining a server self-check task; S303, based on the server self-check task, control the target server to enter a specified startup mode, wherein the target server is configured to: after entering the specified startup mode, acquire a target image; load the acquired target image; and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. S304, receive network topology parameters determined according to the hardware information from the target server, and perform a preset operation on the target server according to the network topology parameters, the preset operation including at least one of the following: node information management and port information management.

[0069] In some examples, a user (administrator) can input a check request for a target server through an ironic client, and make the check request sent to an ironic-api, and the management node can receive the check request via the ironic-api.

[0070] In some examples, after receiving the check request, the ironic-api in the management node can determine a server self-check task corresponding to the check request (for example, according to the target server to which the check request is directed, determine the target server to be checked, and thus acquire the server self-check task for the target server), and the ironic-api can forward the server self-check task to the ironic-conductor in the management node through a remote procedure call (RPC, that is, RPC call in Figure 2 The ironic-conductor can further control the ironic-inspector in the management node to execute the server self-check task through an API call, and the ironic-inspector can be invoked by the ironic client to invoke the ironic-api, and control the target server to enter a specified startup mode (for example, PXE restart, that is, the restart operation) through an ipmitool out-of-band control command.

[0071] In some examples, the ironic-inspector of the management node can also control the target server to shut down through the ironic client invoking the ironic-api.

[0072] In an optional implementation, the acquiring the target image includes: After the target server enters the specified startup mode, acquiring a temporary network address; Based on the temporary network address, acquiring the target image.

[0073] In an optional implementation, the hardware information of the target server is collected by running a preset agent service in the loaded target image. The hardware information of the target server is collected by running a preset agent service in the loaded target image.

[0074] In an optional implementation, the preset agent service is configured to invoke a network management tool to collect the hardware information of the target server during the running of the preset agent service, and the network management tool corresponds to the NPU.

[0075] In an optional implementation, the target server is a bare-metal server.

[0076] In an optional implementation, the hardware information further includes at least one of the following: CPU (Central Processing Unit) information; memory information; disk information.

[0077] In an optional implementation, the network card information includes LLDP (Link Layer Discovery Protocol) information of the NPU, the number of the NPU is multiple, the NPU communicates via a parameter plane network, and the network topology parameter is related to the parameter plane network and is determined by the LLDP information.

[0078] In an optional implementation, the network topology parameter includes at least one of the following: a MAC (Media Access Control) address of a network card of the NPU; parameter information of a switch associated with the network card of the NPU; interface parameter information of the network card of the NPU.

[0079] In a third aspect, refer to Figure 4 , a structural schematic diagram of a server information collection system is shown, the server information collection system includes: a management and control node 401 configured to receive an inspection request for a target server 402, determine a server self-check task in response to the inspection request, control the target server 402 to enter a specified startup mode based on the server self-check task, and the target server 402 including an NPU, the target server 402 being configured to obtain a target image after entering the specified startup mode, load the obtained target image, and collect hardware information of the target server 402 by using the loaded target image, wherein the hardware information includes network card information of the NPU.

[0080] In an optional implementation, the obtaining the target image comprises: After the target server 402 enters the specified starting mode, a temporary network address is obtained; Based on the temporary network address, the target image is obtained.

[0081] In an optional implementation, the collecting the hardware information of the target server 402 through the loaded target image comprises: The hardware information of the target server 402 is collected by running a preset agent service in the loaded target image.

[0082] In an optional implementation, the preset agent service is configured to invoke a network management tool to collect the hardware information of the target server 402 during the running of the preset agent service, and the network management tool corresponds to the NPU.

[0083] In an optional implementation, the target server 402 is a bare-metal server.

[0084] In an optional implementation, the hardware information further comprises at least one of: Central processing unit (CPU) information; Memory information; Disk information.

[0085] In an optional implementation, the network card information comprises link layer discovery protocol (LLDP) information of the NPU, the number of the NPU is multiple, and the NPU communicates via a parameter plane network, and the target server 402 is further configured to: Based on the LLDP information, a network topology parameter related to the parameter plane network is determined; The network topology parameter is sent to the management and control node 401.

[0086] In an optional implementation, the network topology parameter comprises at least one of: Media access control (MAC) address of a network card of the NPU; Parameter information of a switch associated with the network card of the NPU; Interface parameter information of the network card of the NPU.

[0087] In an optional implementation, the management and control node 401 is further configured to perform a preset operation on the target server 402 according to the network topology parameter, and the preset operation comprises at least one of node information management and port information management.

[0088] In a fourth aspect, correspondingly, the embodiments of the present application further provide a server information collection apparatus, which can implement all processes of the server information collection method provided in any of the embodiments of the first aspect.

[0089] Referring to Figure 5 , a structure schematic diagram of a server information collection apparatus provided by the embodiments of the present application is shown, the server information collection apparatus 500 comprises: a mirror image acquisition module 501, configured to acquire a target mirror image; a mirror image loading module 502, configured to load the acquired target mirror image in a target server, wherein the target server comprises a neural network processor NPU; a collection module 503, configured to collect hardware information of the target server through the loaded target mirror image, wherein the hardware information comprises network card information of the NPU.

[0090] In an optional implementation, the acquiring of the target mirror image comprises: acquiring a temporary network address after the target server enters a specified startup mode; acquiring the target mirror image based on the temporary network address.

[0091] In an optional implementation, the collecting of the hardware information of the target server through the loaded target mirror image comprises: collecting the hardware information of the target server by running a preset proxy service in the loaded target mirror image.

[0092] In an optional implementation, the preset proxy service is configured to invoke a network management tool to collect the hardware information of the target server in the process of running of the preset proxy service, and the network management tool corresponds to the NPU.

[0093] In an optional implementation, the target server is a bare metal server.

[0094] In an optional implementation, the hardware information further comprises at least one of the following: central processing unit CPU information; memory information; disk information.

[0095] In an optional implementation, the network card information comprises link layer discovery protocol LLDP information of the NPU, the number of the NPU is multiple, the NPU communicates via a parameter plane network, and the apparatus further comprises: a parameter determination module, configured to determine a network topology parameter related to the parameter plane network based on the LLDP information; The parameter sending module is configured to send the network topology parameter to a management and control node corresponding to the target server.

[0096] In an optional implementation, the network topology parameter comprises at least one of: a media access control (MAC) address of a network card of the NPU; parameter information of a switch associated with the network card of the NPU; interface parameter information of the network card of the NPU.

[0097] In an optional implementation, the network topology parameter is used to instruct the management and control node to perform a preset operation on the target server, and the preset operation comprises at least one of node information management and port information management.

[0098] In a fifth aspect, correspondingly, the embodiments of the present application further provide a server information updating apparatus, which can implement all processes of the server information updating method provided in any of the embodiments of the second aspect.

[0099] Referring to Figure 6 , a structural schematic diagram of a server information updating apparatus provided by the embodiments of the present application is shown. The server information updating apparatus 600 is suitable for a management and control node, and comprises: a request receiving module 601 configured to receive a check request for a target server, wherein the target server comprises an NPU; a task determining module 602 configured to determine a server self-check task in response to the check request; a mode control module 603 configured to control the target server to enter a specified startup mode based on the server self-check task, wherein the target server is configured to: after entering the specified startup mode, acquire a target image; load the acquired target image; and collect hardware information of the target server through the loaded target image, wherein the hardware information comprises network card information of the NPU; a management module 604 configured to receive a network topology parameter determined according to the hardware information from the target server, and perform a preset operation on the target server according to the network topology parameter, wherein the preset operation comprises at least one of node information management and port information management.

[0100] In an optional implementation, the acquiring of the target image comprises: acquiring a temporary network address after the target server enters the specified startup mode; acquiring the target image based on the temporary network address.

[0101] In an optional implementation, the hardware information of the target server is collected by running a preset agent service in the loaded target image. The hardware information of the target server is collected by running a preset agent service in the loaded target image.

[0102] In an optional implementation, the preset agent service is configured to invoke a network management tool to collect the hardware information of the target server during the running of the preset agent service, and the network management tool corresponds to the NPU.

[0103] In an optional implementation, the target server is a bare-metal server.

[0104] In an optional implementation, the hardware information further includes at least one of the following: CPU (Central Processing Unit) information; memory information; disk information.

[0105] In an optional implementation, the network card information includes LLDP (Link Layer Discovery Protocol) information of the NPU, the number of the NPU is multiple, the NPU communicates via a parameter plane network, and the network topology parameter is related to the parameter plane network and is determined by the LLDP information.

[0106] In an optional implementation, the network topology parameter includes at least one of the following: a MAC (Media Access Control) address of a network card of the NPU; parameter information of a switch associated with the network card of the NPU; interface parameter information of the network card of the NPU.

[0107] In a sixth aspect, an embodiment of the present application provides a non-transitory computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement steps of the method in any of the preceding aspects.

[0108] In a seventh aspect, an embodiment of the present application provides a computer program product, which includes computer instructions, and the computer instructions are executed by a processor to implement steps of the method in any of the preceding aspects.

[0109] In an eighth aspect, an embodiment of the present application provides a computer device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor executes the computer program to implement steps of the method in any of the preceding aspects.

[0110] Reference is made to Figure 7The computer device of this embodiment comprises a processor 701, a memory 702, and a computer program stored in the memory 702 and capable of running on the processor 701, for example, a server information collection program or a server information update program. The processor 701 implements the steps in each of the above server information collection method embodiments or server information update method embodiments when executing the computer program.

[0111] For example, the computer program can be divided into one or more modules / units, which are stored in the memory 702 and executed by the processor 701 to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the computer device.

[0112] The computer device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The computer device can comprise, but is not limited to, the processor 701 and the memory 702. Those skilled in the art can understand that the schematic diagram is only an example of the computer device, and does not limit the computer device, which can comprise more or fewer components than the diagram, or combine certain components, or different components, for example, the computer device can also comprise an input / output device, a network access device, a bus, and the like.

[0113] The processor 701 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, and the like. The general-purpose processor can be a microprocessor or the processor 701 can also be any conventional processor and the like, which is the control center of the computer device, and connects all parts of the computer device through various interfaces and lines.

[0114] The memory 702 can be used to store the computer programs and / or modules, and the processor 701 realizes various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 702, and calling the data stored in the memory 702. The memory 702 can mainly include a program storage area and a data storage area, wherein the program storage area can store operating systems, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), and the like. In addition, the memory 702 can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0115] The modules / units integrated in the computer device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a non-transitory computer readable storage medium. When the processor 701 executes the computer program, the steps of the above-mentioned various method embodiments can be realized. The computer program includes computer program code, which can be in the form of source code, object code, an executable file, or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0116] In summary, the embodiments of the present application have at least the following beneficial effects: By adopting the embodiment of the present application, the target image obtained is loaded in the target server containing the NPU, so as to realize automatic collection of hardware information of the target server through the target image loaded in the target server, thereby improving the collection efficiency of the server hardware information, in particular, the NPU network card information which has a greater impact on the server and large-scale cluster performance is also collected, thereby improving the collection accuracy of the key hardware information of the server, and further improving the construction efficiency, scalability of the large-scale cluster, and reducing the operation and maintenance complexity of the cluster.

[0117] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and necessary hardware platforms, and of course, it can also be implemented entirely by hardware. Based on such understanding, all or part of the technical solutions of the present application which contribute to the background art can be embodied in the form of a software product. The computer software product can be stored in a storage medium, such as a ROM (Read-Only Memory), a RAM (Random Access Memory), a magnetic disk, an optical disk, etc., and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present application.

[0118] The above is the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements are also considered to be within the protection scope of the present application.

Claims

1. A server information collection method characterized by comprising: The method comprises: acquiring a target image; loading the acquired target image in a target server, wherein the target server comprises a neural network processor (NPU); collecting hardware information of the target server through the loaded target image, wherein the hardware information comprises network card information of the NPU.

2. The method of claim 1, wherein, The acquiring of the target image comprises: acquiring a temporary network address after the target server enters a specified startup mode; acquiring the target image based on the temporary network address.

3. The method of claim 1, wherein, The collecting of the hardware information of the target server through the loaded target image comprises: collecting the hardware information of the target server by running a preset proxy service in the loaded target image.

4. The method of claim 3, wherein, The preset proxy service is configured to call a network management tool to collect the hardware information of the target server during the running of the preset proxy service, and the network management tool corresponds to the NPU.

5. The method of claim 1, wherein, The target server is a bare-metal server.

6. The method of claim 1, wherein, The hardware information further comprises at least one of: central processing unit (CPU) information; memory information; disk information.

7. The method according to any one of claims 1 to 6, characterized in that, The network card information comprises link layer discovery protocol (LLDP) information of the NPU, the number of the NPUs is multiple, the NPUs communicate via a parameter plane network, and the method further comprises: determining network topology parameters related to the parameter plane network based on the LLDP information; sending the network topology parameters to a management and control node corresponding to the target server.

8. The method of claim 7, wherein, The network topology parameters comprise at least one of: media access control (MAC) addresses of network cards of the NPUs; parameter information of switches associated with the network cards of the NPUs; interface parameter information of the network cards of the NPUs.

9. The method of claim 7, wherein, The network topology parameters are used to instruct the management and control node to perform a preset operation on the target server, and the preset operation comprises at least one of: node information management and port information management.

10. A server information update method characterized by comprising: The method is suitable for a management and control node, and the method comprises: receiving an inspection request for a target server, wherein the target server comprises an NPU; determining a server self-check task in response to the inspection request; controlling the target server to enter a specified startup mode based on the server self-check task, wherein the target server is configured to: acquire a target image after entering the specified startup mode; load the acquired target image; and collect hardware information of the target server through the loaded target image, wherein the hardware information comprises network card information of the NPU; receiving network topology parameters determined based on the hardware information from the target server, and performing a preset operation on the target server based on the network topology parameters, wherein the preset operation comprises at least one of: node information management and port information management.

11. A server information collection system characterized by comprising: The method comprises: a management and control node configured to receive an inspection request for a target server; determining a server self-check task in response to the inspection request; controlling the target server to enter a specified startup mode based on the server self-check task; and ​ The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU.

12. A server information collecting apparatus characterized by comprising: The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. The target server includes an NPU, and the target server is configured to, after entering a specified startup mode, acquire a target image, load the acquired target image, and collect hardware information of the target server through the loaded target image, wherein the hardware information includes network card information of the NPU. The computer program is executed by a processor to implement the method of any one of claims 1-10.

13. A server information updating apparatus characterized by comprising: The computer program is executed by a processor to implement the method of any one of claims 1-10. The computer program is executed by a processor to implement the method of any one of claims 1-10. The computer program is executed by a processor to implement the method of any one of claims 1-10. ​ ​ 14. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, ​ 15. A computer program product comprising computer instructions, characterized in that, ​ 16. A computer device, comprising: ​

Images