Method and device for identifying internal connection state of computing cluster, equipment and medium

Abstract

The invention provides a computing cluster internal connection state identification method and device, equipment and a medium, and the method comprises the steps: determining a standard topology connection state of a computing cluster according to the attribute information of each computing server and switch; obtaining identification information of each computing node connected to each switch; determining an actual topology connection state of the computing cluster according to the identification information of each computing node; and comparing the actual topology connection state with the standard topology connection state, and determining a connection state in the computing cluster according to a comparison result. According to the technical scheme, the actual topological connection state of the computing cluster is compared with the standard topological connection state, so that the connection state of each port in the computing cluster can be accurately identified, the problem root cause can be accurately and quickly positioned when the connection is abnormal from the angle of accuracy and convenience, and the reliability of the computing cluster is improved. Therefore, the convenience of operation and maintenance management of the computing cluster is improved.

Description

Technical Field

[0001] The present application relates to the field of computing cluster technology, and in particular to a method, device, equipment and medium for identifying the internal connection status of a computing cluster. Background Art

[0002] With the continuous growth of modern computing needs, especially in the fields of scientific computing, machine learning and data analysis, the computing power of a single GPU (Graphics Processing Unit) server is gradually unable to meet the needs of the scene. It has become a trend to interconnect multiple GPU servers to form a powerful computing cluster. For example, multiple GPU servers are connected through multiple PCIe (Peripheral Component Interconnect Express, a high-speed serial computer expansion bus standard) switches to enable them to communicate with each other through a large-capacity, high-density interconnection solution. Figure 1 The figure shows an interconnection diagram of a computing cluster consisting of multiple GPU servers and multiple PCIe switches.

[0003] Figure 1 The computing cluster shown in the figure may include GPUs from different manufacturers, and different GPU manufacturers have different requirements for the topology connection of the entire cluster, but almost all of them have relatively strict requirements for the internal connection of the cluster. Once the connection is wrong, the function of the cluster will be affected or even completely unable to operate normally. Although the connection of the cluster can be known by marking the cables, the possibility of wrong connection cannot be ruled out, and the impact of wrong connection is great, and even cluster function abnormality. Therefore, the detection function of the internal topology connection status of the computing cluster is particularly important. If the relevant abnormalities can be accurately identified and reported to the user, the convenience of cluster operation and maintenance management can be significantly improved.

[0004] Therefore, how to accurately identify the status of topological connections within a computing cluster is a technical problem that urgently needs to be solved in this field. Summary of the invention

[0005] The purpose of this application is to provide a method, device, equipment and storage medium for identifying the internal connection status of a computing cluster, which can accurately identify the connection status within a computing server interconnection cluster.

[0006] In a first aspect, the present application provides a method for identifying the internal connection status of a computing cluster, wherein the computing cluster includes a management server, multiple computing servers, and at least one switch, each computing server includes multiple computing nodes, and all computing nodes in the computing cluster are interconnected through the at least one switch. The method is executed by the management server, and the method includes:

[0007] Determining a standard topology connection state of the computing cluster according to the attribute information of each computing server and the switch;

[0008] Obtaining identification information of each computing node connected to each of the switches;

[0009] Determining the actual topological connection state of the computing cluster according to the identification information of each computing node;

[0010] The actual topology connection state is compared with the standard topology connection state, and the connection state inside the computing cluster is determined according to the comparison result.

[0011] In a possible implementation, the attribute information of the computing server includes: a node identifier of each computing node in the computing server, a port number for internal connection and a port number for external connection of each computing node, and grouping information corresponding to each computing node; the attribute information of the switch includes: the number of ports of the switch;

[0012] The standard topology connection state includes at least one of the following:

[0013] The number of computing nodes connected to each switch does not exceed the number of ports on the switch;

[0014] Each computing node is connected to the switch through its port for external connection;

[0015] Compute nodes belonging to the same group are connected to the same switch.

[0016] In a possible implementation, the identification information includes a server identification of a computing server to which the computing node belongs, a node identification of the computing node, and a port number of a connection between the computing node and a switch;

[0017] The determining, according to the identification information of each computing node, an actual topological connection state of the computing cluster includes:

[0018] According to the identification information of each computing node, the computing nodes actually connected to each switch and the number of computing nodes are counted, and the port numbers of the computing nodes actually connected to each switch are counted.

[0019] In a possible implementation, comparing the actual topology connection state with the standard topology connection state, and determining the connection state within the computing cluster according to the comparison result, includes:

[0020] If the number of computing nodes actually connected to each switch exceeds the number of ports of the switch, it is determined that the switch connection is abnormal.

[0021] In a possible implementation, comparing the actual topology connection state with the standard topology connection state, and determining the connection state within the computing cluster according to the comparison result, includes:

[0022] If the port number of the computing node actually connected to the switch is not the port number of the computing node used for external connection, it is determined that the connection of the computing node is abnormal.

[0023] In a possible implementation, comparing the actual topology connection state with the standard topology connection state, and determining the connection state within the computing cluster according to the comparison result, includes:

[0024] If the computing nodes of the same group are not connected to the same switch, it is determined that the computing nodes of the group are abnormally connected.

[0025] A second aspect of the present application provides a device for identifying the internal connection status of a computing cluster, wherein the computing cluster includes a management server, multiple computing servers, and at least one switch, each computing server includes multiple computing nodes, and all computing nodes in the computing cluster are interconnected through the at least one switch, and the device is executed by the management server, and the device includes:

[0026] A standard topology determination module, configured to determine a standard topology connection state of the computing cluster according to attribute information of each computing server and the switch;

[0027] An identification information acquisition module, used to acquire identification information of each computing node connected to each of the switches;

[0028] An actual topology determination module, used to determine the actual topological connection state of the computing cluster according to the identification information of each computing node;

[0029] The connection state identification module is used to compare the actual topological connection state with the standard topological connection state, and determine the connection state inside the computing cluster according to the comparison result.

[0030] In a possible implementation, the attribute information of the computing server includes: a node identifier of each computing node in the computing server, a port number for internal connection and a port number for external connection of each computing node, and grouping information corresponding to each computing node; the attribute information of the switch includes: the number of ports of the switch;

[0031] The standard topology connection state includes at least one of the following:

[0032] The number of computing nodes connected to each switch does not exceed the number of ports on the switch;

[0033] Each computing node is connected to the switch through its port for external connection;

[0034] Compute nodes belonging to the same group are connected to the same switch.

[0035] In a possible implementation, the identification information includes a server identification of a computing server to which the computing node belongs, a node identification of the computing node, and a port number of a connection between the computing node and a switch;

[0036] The actual topology determination module is specifically used to:

[0037] According to the identification information of each computing node, the computing nodes actually connected to each switch and the number of computing nodes are counted, and the port numbers of the computing nodes actually connected to each switch are counted.

[0038] In a possible implementation, the connection status identification module is specifically used to:

[0039] If the number of computing nodes actually connected to each switch exceeds the number of ports of the switch, it is determined that the switch connection is abnormal.

[0040] In a possible implementation, the connection status identification module is specifically used to:

[0041] If the port number of the computing node actually connected to the switch is not the port number of the computing node used for external connection, it is determined that the connection of the computing node is abnormal.

[0042] In a possible implementation, the connection status identification module is specifically used to:

[0043] If the computing nodes of the same group are not connected to the same switch, it is determined that the computing nodes of the group are abnormally connected.

[0044] A third aspect of the present application provides a management device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.

[0045] A fourth aspect of the present application provides a computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions can be executed by a processor to implement the method as described in the first aspect.

[0046] Compared with the prior art, the method, device, equipment and storage medium for identifying the internal connection status of a computing cluster provided by the present application determine the standard topological connection status of the computing cluster according to the attribute information of each computing server and switch; obtain the identification information of each computing node connected to each of the switches; determine the actual topological connection status of the computing cluster according to the identification information of each computing node; compare the actual topological connection status with the standard topological connection status, and determine the internal connection status of the computing cluster according to the comparison result. The technical solution of the present application can accurately identify the connection status of each port inside the computing cluster by comparing the actual topological connection status of the computing cluster with the standard topological connection status. From the perspective of accuracy and convenience, when a connection abnormality occurs, the root cause of the problem can be accurately and quickly located, thereby improving the convenience of computing cluster operation and maintenance management. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0048] Figure 1 A schematic diagram of the structure of an existing computing cluster is shown;

[0049] Figure 2 A schematic diagram of the structure of a computing cluster provided by the present application is shown;

[0050] Figure 3 A flowchart of a method for identifying the internal connection status of a computing cluster provided by the present application is shown;

[0051] Figure 4 A schematic diagram of the interaction within the computing cluster provided by the present application is shown;

[0052] Figure 5 A schematic diagram showing a device for identifying the internal connection status of a computing cluster provided by the present application is shown;

[0053] Figure 6 A schematic diagram of a management device provided by the present application is shown. DETAILED DESCRIPTION

[0054] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0055] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by technicians in the field to which this application belongs.

[0056] In addition, the terms "first" and "second" etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0057] For the scenario where the connection between the GPU server port and the switch port in the computing cluster is abnormal, an existing solution is that the internal software of the GPU determines the communication status of each port and the rationality of PCIe resource allocation, sets the fault register corresponding to the unreasonable port, collects and identifies faults through the BMC (Baseboard Management Controller) in the GPU server, and reports the alarm log of the abnormal GPU port connection.

[0058] In the above solution, first of all, users need to pay attention to whether the BMC in each GPU server in the cluster reports relevant faults, which increases the complexity of operation and maintenance work; secondly, the port connection status is identified inside the GPU. Once a problem occurs, the P2P interconnection between GPUs cannot be established. Each GPU can only be identified individually through communication status and resource allocation rationality. It may not be possible to accurately identify the specific location of the connection abnormality, which is not conducive to problem analysis and resolution.

[0059] P2P: Peer to Peer, different PCIe devices can communicate with each other directly without going through the central processor.

[0060] In view of this, embodiments of the present application provide a method and apparatus for identifying the internal connection status of a computing cluster, a management device, and a computer-readable storage medium, which are described below in conjunction with the accompanying drawings.

[0061] Please refer to Figure 2, which shows a schematic diagram of the structure of a computing cluster provided by some embodiments of the present application. Figure 2 As shown, the computing cluster includes a management server, multiple computing servers and at least one switch, each computing server includes multiple computing nodes, and all computing nodes in the computing cluster are interconnected through the at least one switch.

[0062] Computing servers include but are not limited to GPU servers, such as an 8-card GPU server. Usually, the models of GPU cards in the same GPU server are the same, but the models of GPU cards in different GPU servers in the cluster may be different. Due to different models, the port connection methods of GPU cards interconnected by switches may be different.

[0063] The switch can be a switch that supports PCIe. All GPU cards in the cluster can be connected to the PCIe switch to achieve efficient interconnection.

[0064] Please refer to Figure 3 , which shows a flowchart of a method for identifying the internal connection status of a computing cluster provided by some embodiments of the present application, the method is executed by a management server, such as Figure 3 As shown, the following steps S101-S104 may be specifically included:

[0065] S101, determining a standard topology connection state of the computing cluster according to attribute information of each computing server and the switch;

[0066] The attribute information of the computing server includes: the node identification of each computing node in the computing server, the port number used for internal connection and the port number used for external connection of each computing node, and the grouping information corresponding to each computing node.

[0067] The attribute information of the switch includes: the number of ports of the switch.

[0068] Specifically, the management server establishes a communication connection with the computing server and the switch, obtains the attribute information of the computing server and the switch, and obtains the standard topological connection state of the computing cluster according to the attribute information, that is, the correct topological connection state that meets the P2P interconnection requirements. For example, according to the number of ports of the switch, the standard topological connection state can be obtained that the number of computing nodes connected to the switch cannot exceed the number of ports of the switch; according to the port number for internal connection and the port number for external connection of the computing node, the standard topological connection state can be obtained that the port number connected to the switch cannot be the port number for internal connection, but should be the port number for external connection, and so on.

[0069] In some implementations of the present application, the standard topology connection state determined according to the attribute information of each of the computing servers and the switch includes at least one of the following:

[0070] The number of computing nodes connected to each switch does not exceed the number of ports on the switch;

[0071] Each computing node is connected to the switch through its port for external connection;

[0072] Compute nodes belonging to the same group are connected to the same switch.

[0073] S102, obtaining identification information of each computing node connected to each of the switches, the identification information including a server identification of a computing server to which the computing node belongs, a node identification of the computing node, and a port number through which the computing node is connected to the switch;

[0074] Each computing server has a server ID, which can be sent to the computing server by the management server. The computing node has a corresponding node ID, and each port of the computing node has a corresponding port number. Therefore, when the computing node is connected to the switch, the switch can obtain the various identification information of the computing node, thereby knowing the computing server to which the computing node belongs and the connected port number. The number of computing nodes connected to the switch can also be known by counting the node IDs.

[0075] S103, determining the actual topological connection state of the computing cluster according to the identification information of each computing node;

[0076] Specifically, according to the identification information of each computing node, the computing nodes actually connected to each switch and the number of computing nodes can be obtained by counting, and the port numbers of the computing nodes actually connected to each switch can be obtained by counting.

[0077] S104: Compare the actual topology connection state with the standard topology connection state, and determine the connection state inside the computing cluster according to the comparison result.

[0078] Specifically, the process of determining the internal connection status of the computing cluster includes:

[0079] Compare whether the number of computing nodes actually connected to each switch exceeds the number of ports of the switch. If the number of computing nodes actually connected to each switch exceeds the number of ports of the switch, it is determined that the switch connection is abnormal.

[0080] Whether each computing node is actually connected to the switch through its port for external connection, if the port number of the computing node actually connected to the switch is not the port number of the computing node for external connection, it is determined that the connection of the computing node is abnormal.

[0081] The computing nodes belonging to the same group should be connected to the same switch. If the computing nodes of the same group are not connected to the same switch, it is determined that the connection of the computing nodes of the group is abnormal.

[0082] For ease of understanding, the method for identifying the internal connection status of a computing cluster of the present application is described below by way of example.

[0083] Take the computing cluster consisting of management servers, GPU servers, and PCIe switches as an example. Figure 4 The figure shows the interaction diagram inside the computing cluster.

[0084] 1. The definition of cluster topology identification code is as follows:

[0085] GPU ID: The serial number used to identify the GPU card connected to the corresponding port.

[0086] PORT ID: Used to identify the GPU port number connected to the corresponding port, such as the port number for internal connection and the port number for external connection. Some GPU models do not need to distinguish between the ports for internal connection and the ports for external connection.

[0087] SERVER ID: The serial number used to identify the GPU server connected to the corresponding port.

[0088] GROUP ID: This is used to identify the group to which the GPU connected to the corresponding port belongs. Some GPU models do not need to identify whether they are in the same group or have other ways to allow the GPU to distinguish them internally, so there is no need to pass the group ID.

[0089] 2. Cluster operation software is introduced as follows:

[0090] System Manager: runs in a separate management server node and is responsible for sorting and issuing SERVER IDs.

[0091] Server Manager: runs in the system of each GPU server as single server management software, used to set the SERVER ID.

[0092] GPU Driver: runs in the system of each GPU server and is used to interact with the GPU card.

[0093] Switch Manager: runs in the PCIe switch of the cluster and is used to collect all the identification codes of each GPU connected to this switch.

[0094] 3. The cluster connection status identification process is as follows:

[0095] (1) Generation of identification code

[0096] GPU ID, PORT ID, GROUP ID: generated by the GPU card itself and stored in its PCIe configuration space. The configuration space is one of the four address spaces defined by the PCIe architecture, the other three being memory space, read-write space, and message space.

[0097] SERVER ID: Assigned uniformly by System Manager.

[0098] (2) Transmission and judgment of identification code

[0099] When the cluster is powered on, the IP address of the System Manager needs to be fixed. First, the Server Manager and Switch Manager transmit their own IP addresses and other information to the System Manager through the network to establish a network communication connection.

[0100] System Manager identifies each connected GPU server and sends the SERVER ID to ServerManager. Server Manager sends the SERVER ID to each GPU card through the GPU Driver and stores it in the GPU's PCIe configuration space. System Manager determines the standard topology connection status of the computing cluster based on the attribute information of each GPU server and switch.

[0101] Switch Manager reads each identification code from the PCIe configuration space of the GPU and makes simple judgments as much as possible, such as identifying the GPU servers connected to the switch and whether the total number of GPU cards meets expectations.

[0102] System Manager interacts with Switch Manager to obtain the GPU cards and related identification codes connected to each switch, and performs integration verification of the identification codes based on the standard topology connection status, such as determining whether the identification codes belong to the same cluster, and whether the GPU cards and port numbers connected to each switch meet the topology requirements of the GPU manufacturer's P2P interconnection. In this way, the topology connection status within the entire cluster is ultimately determined, and abnormal switch ports are reported to facilitate unified rectification by operation and maintenance personnel.

[0103] It can be seen that, different from the existing technical solutions, this application defines various identification codes within the cluster, subdivides the responsibilities of each software in the cluster, clarifies the interaction process of each software, and finally completes the accurate identification of the internal connection status of the computing cluster. By defining various identification codes, this solution can be compatible with most GPU cards to the greatest extent. Even if GPUs from other manufacturers require additional judgment conditions in the future, the corresponding identification codes can be directly modified or added, and then the standard topology connection status can be updated, and the overall process is almost unaffected. The division of labor of each management software and driver is further clarified, and the entire computing cluster is managed through the collaboration of multiple software, laying the foundation for the enrichment of subsequent cluster management functions.

[0104] The method for identifying the internal connection status of the computing cluster provided in the embodiment of the present application determines the standard topological connection status of the computing cluster based on the attribute information of each computing server and switch; obtains the identification information of each computing node connected to each of the switches; determines the actual topological connection status of the computing cluster based on the identification information of each computing node; compares the actual topological connection status with the standard topological connection status, and determines the internal connection status of the computing cluster based on the comparison result. The technical solution of the present application can accurately identify the connection status of each port inside the computing cluster by comparing the actual topological connection status of the computing cluster with the standard topological connection status. From the perspective of accuracy and convenience, when a connection abnormality occurs, the root cause of the problem can be accurately and quickly located, thereby improving the convenience of computing cluster operation and maintenance management.

[0105] In the above embodiment, a method for identifying the internal connection status of a computing cluster is provided. Correspondingly, the present application also provides a device for identifying the internal connection status of a computing cluster. The device for identifying the internal connection status of a computing cluster provided in the embodiment of the present application can implement the above method for identifying the internal connection status of a computing cluster. Figure 5 , which shows a schematic diagram of a device for identifying the internal connection status of a computing cluster provided by some embodiments of the present application. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described below is only illustrative.

[0106] The computing cluster includes a management server, multiple computing servers and at least one switch, each computing server includes multiple computing nodes, all computing nodes in the computing cluster are interconnected through the at least one switch, and the device is executed by the management server, such as Figure 5 As shown, the device 10 for identifying the internal connection status of the computing cluster includes:

[0107] The standard topology determination module 101 is used to determine the standard topology connection state of the computing cluster according to the attribute information of each computing server and the switch; the attribute information of the computing server includes: the node identification of each computing node in the computing server, the port number for internal connection and the port number for external connection of each computing node, and the grouping information corresponding to each computing node; the attribute information of the switch includes: the number of ports of the switch;

[0108] An identification information acquisition module 102 is used to acquire identification information of each computing node connected to each of the switches, wherein the identification information includes a server identification of a computing server to which the computing node belongs, a node identification of the computing node, and a port number of the computing node connected to the switch;

[0109] The actual topology determination module 103 is used to determine the actual topological connection state of the computing cluster according to the identification information of each computing node;

[0110] The connection state identification module 104 is used to compare the actual topological connection state with the standard topological connection state, and determine the connection state inside the computing cluster according to the comparison result.

[0111] In a possible implementation, the standard topology connection state includes at least one of the following:

[0112] The number of computing nodes connected to each switch does not exceed the number of ports on the switch;

[0113] Each computing node is connected to the switch through its port for external connection;

[0114] Compute nodes belonging to the same group are connected to the same switch.

[0115] In a possible implementation, the actual topology determination module 103 is specifically configured to:

[0116] According to the identification information of each computing node, the computing nodes actually connected to each switch and the number of computing nodes are counted, and the port numbers of the computing nodes actually connected to each switch are counted.

[0117] In a possible implementation, the connection status identification module 104 is specifically used to:

[0118] If the number of computing nodes actually connected to each switch exceeds the number of ports of the switch, it is determined that the switch connection is abnormal.

[0119] In a possible implementation, the connection status identification module 104 is specifically used to:

[0120] If the port number of the computing node actually connected to the switch is not the port number of the computing node used for external connection, it is determined that the connection of the computing node is abnormal.

[0121] In a possible implementation, the connection status identification module 104 is specifically used to:

[0122] If the computing nodes of the same group are not connected to the same switch, it is determined that the computing nodes of the group are abnormally connected.

[0123] The apparatus for identifying the internal connection status of a computing cluster provided in the above-mentioned embodiment of the present application and the method for identifying the internal connection status of a computing cluster provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects.

[0124] An embodiment of the present application also provides a management device corresponding to the method for identifying the internal connection status of a computing cluster provided in the aforementioned embodiment. The device may be an electronic device for a server, such as a laptop computer, a tablet computer, a desktop computer, etc., to execute the aforementioned method for identifying the internal connection status of a computing cluster.

[0125] Please refer to Figure 6 , which shows a schematic diagram of a management device provided by some embodiments of the present application. Figure 6 As shown, the management device 20 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, and the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and when the processor 200 runs the computer program, it executes the method for identifying the internal connection status of the computing cluster provided in any of the aforementioned embodiments of the present application.

[0126] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0127] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving the execution instruction. The method for identifying the internal connection state of the computing cluster disclosed in any implementation of the embodiment of the present application may be applied to the processor 200, or implemented by the processor 200.

[0128] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 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 readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the above method in combination with its hardware.

[0129] The management device provided in the embodiment of the present application and the method for identifying the internal connection status of the computing cluster provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented therein.

[0130] An embodiment of the present application also provides a computer-readable storage medium corresponding to the method for identifying the internal connection status of a computing cluster provided in the aforementioned embodiment. The computer-readable storage medium may be a CD on which a computer program (i.e., a program product) is stored. When the computer program is executed by a processor, the method for identifying the internal connection status of a computing cluster provided in any of the aforementioned embodiments will be executed.

[0131] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0132] The computer-readable storage medium provided in the above-mentioned embodiment of the present application and the method for identifying the internal connection status of a computing cluster provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0133] Finally, it should be noted that the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0134] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0135] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0136] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application.

Claims

1. A method for identifying the internal connection status of a computing cluster, characterized in that: The computing cluster includes a management server, multiple computing servers and at least one switch, each computing server includes multiple computing nodes, all computing nodes in the computing cluster are interconnected through the at least one switch, and the method is executed by the management server, and the method includes: Determining a standard topology connection state of the computing cluster according to the attribute information of each computing server and the switch; Obtaining identification information of each computing node connected to each of the switches; Determining the actual topological connection state of the computing cluster according to the identification information of each computing node; The actual topology connection state is compared with the standard topology connection state, and the connection state inside the computing cluster is determined according to the comparison result.

2. The method according to claim 1, characterized in that The attribute information of the computing server includes: the node identification of each computing node in the computing server, the port number for internal connection and the port number for external connection of each computing node, and the grouping information corresponding to each computing node; the attribute information of the switch includes: the number of ports of the switch; The standard topology connection state includes at least one of the following: The number of computing nodes connected to each switch does not exceed the number of ports on the switch; Each computing node is connected to the switch through its port for external connection; Compute nodes belonging to the same group are connected to the same switch.

3. The method according to claim 2, characterized in that The identification information includes a server identification of a computing server to which the computing node belongs, a node identification of the computing node, and a port number of a connection between the computing node and the switch; The determining, according to the identification information of each computing node, an actual topological connection state of the computing cluster includes: According to the identification information of each computing node, the computing nodes actually connected to each switch and the number of computing nodes are counted, and the port numbers of the computing nodes actually connected to each switch are counted.

4. The method according to claim 3, characterized in that The comparing the actual topology connection state with the standard topology connection state and determining the connection state inside the computing cluster according to the comparison result includes: If the number of computing nodes actually connected to each switch exceeds the number of ports of the switch, it is determined that the switch connection is abnormal.

5. The method according to claim 3, characterized in that: The comparing the actual topology connection state with the standard topology connection state and determining the connection state inside the computing cluster according to the comparison result includes: If the port number of the computing node actually connected to the switch is not the port number of the computing node used for external connection, it is determined that the connection of the computing node is abnormal.

6. The method according to claim 3, characterized in that The comparing the actual topology connection state with the standard topology connection state and determining the connection state inside the computing cluster according to the comparison result includes: If the computing nodes of the same group are not connected to the same switch, it is determined that the computing nodes of the group are abnormally connected.

7. A device for identifying the internal connection status of a computing cluster, characterized in that: The computing cluster includes a management server, multiple computing servers and at least one switch, each computing server includes multiple computing nodes, all computing nodes in the computing cluster are interconnected through the at least one switch, and the device is executed by the management server, and the device includes: A standard topology determination module, configured to determine a standard topology connection state of the computing cluster according to attribute information of each computing server and the switch; An identification information acquisition module, used to acquire identification information of each computing node connected to each of the switches; An actual topology determination module, used to determine the actual topological connection state of the computing cluster according to the identification information of each computing node; The connection state identification module is used to compare the actual topological connection state with the standard topological connection state, and determine the connection state inside the computing cluster according to the comparison result.

8. The device according to claim 7, characterized in that The attribute information of the computing server includes: the node identification of each computing node in the computing server, the port number for internal connection and the port number for external connection of each computing node, and the grouping information corresponding to each computing node; the attribute information of the switch includes: the number of ports of the switch; The standard topology connection state includes at least one of the following: The number of computing nodes connected to each switch does not exceed the number of ports on the switch; Each computing node is connected to the switch through its port for external connection; Compute nodes belonging to the same group are connected to the same switch.

9. A management device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 6 when executing the computer program.

10. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and the computer-readable instructions can be executed by a processor to implement the method according to any one of claims 1 to 6.

Images