A method and system for out-of-band management of CXL memory devices

Abstract

The present invention discloses an out-of-band management method and system for a CXL memory device, and relates to the technical field of computer system hardware management and maintenance, including real-time judgment of the insertion status of a CXL memory device through a BMC and initialization of the inserted CXL memory device; real-time monitoring of the status of the initialized CXL memory device and fault repair, and device reset and verification of the CXL memory device after the fault repair; combining multiple CXL memory devices to form a resource pool and performing real-time monitoring and hot maintenance through a CXL memory device management system; recording and storing all data of the CXL memory device. The CXL memory device management method proposed in the present invention can effectively manage the connected CXL memory device, more intuitively understand the current CXL memory device status or CXL resource pool utilization and allocation, and the hot maintenance method can avoid the occurrence of server deactivation, the maintenance method is more convenient, supports remote monitoring and automatic management, and greatly improves the stability and availability of the server.

Description

Technical Field

[0001] The present invention relates to the technical field of computer system hardware management and maintenance, and in particular to an out-of-band management method and system for a CXL memory device. Background Art

[0002] With the continuous development of computer architecture, the demand for high-performance computing and data centers is growing. The traditional PCIe bus technology has gradually exposed its limitations in resource expansion and management. In order to cope with the complexity of computing tasks and the demand for large-scale data processing, CXL technology came into being. CXL is a high-speed interconnection protocol based on the PCIe physical layer. It aims to efficiently connect CPU, memory, accelerators and other devices through a unified interface to achieve low-latency, high-bandwidth memory access and device communication. The emergence of CXL technology has brought revolutionary changes to server architecture, enabling multiple devices to share memory resources, enhancing the scalability and computing performance of the system. At present, CXL technology is mainly divided into three sub-protocols: CXL.io, CXL.cache and CXL.memory, which support I / O expansion, cache consistency and memory expansion functions respectively. However, with the increasing application of CXL memory devices in large-scale data centers and high-performance computing environments, how to efficiently manage these devices, especially when the OS fails or the device fails, timely out-of-band management has become an urgent problem to be solved.

[0003] The existing CXL memory device management method mainly relies on management at the operating system level, that is, monitoring and managing the device status through the operating system. However, this management method has many shortcomings. The management method that relies on the operating system will fail when the operating system fails, resulting in the inability to obtain the device status in real time, affecting the stability and reliability of the system. The existing management method is difficult to achieve unified management of multiple CXL memory devices, especially in the case of device resource pooling. There is a lack of dynamic monitoring and resource allocation mechanism for each device in the resource pool, which makes the system perform poorly in load balancing and fault handling. The existing technology has limited support for hot maintenance of CXL memory devices. When the device needs to be replaced or maintained, it is often necessary to interrupt the operation of the server, thereby affecting the service continuity of the entire system. Summary of the invention

[0004] In view of the problems existing in the above-mentioned existing out-of-band management method and system for CXL memory devices, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that the existing CXL memory device management technology has problems such as failure to obtain device status in real time when the operating system fails, insufficient dynamic management of resource pooling devices, and limited hot maintenance support.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: an out-of-band management method for a CXL memory device, comprising:

[0007] Use BMC to determine the insertion status of CXL memory devices in real time and initialize the inserted CXL memory devices;

[0008] Monitor the status of the initialized CXL memory device in real time and perform fault repair, and reset and verify the CXL memory device after the fault repair;

[0009] Combine multiple CXL memory devices into a resource pool and perform real-time monitoring and thermal maintenance through the CXL memory device management system;

[0010] Records all data of CXL memory devices and stores them.

[0011] As a preferred solution of the out-of-band management method of the CXL memory device of the present invention, wherein: the real-time determination of the insertion status of the CXL memory device by the BMC refers to real-time monitoring of the voltage and signal integrity of the PCle bus by the BMC in the CXL memory device management system using an electrical signal detection method;

[0012] When the signal of the PCle bus changes and meets the standard for device insertion, the BMC uses the time domain analysis method to detect the change rate of the rising and falling edges of the signal. When a CXL memory device is inserted, the signal level change on the PCIe bus shows a fast rising and falling edge. A change rate threshold is set. If the change rate exceeds the change rate threshold, it is determined that a CXL memory device is inserted;

[0013] According to the analysis results, if it is determined that no CXL memory device is inserted, the CXL memory device management system prompts that there is no device at present, and displays the "IDEL" prompt message on the LCD display. If it is determined that a CXL memory device is inserted, the CXL memory device management system enters the device insertion state, and the BMC broadcasts the insertion event to other system components, including the operation and maintenance system, CPLD and OS, through the PCle bus.

[0014] As a preferred solution of the out-of-band management method of the CXL memory device of the present invention, wherein: the initialization of the inserted CXL memory device refers to that after the BMC completes the insertion event broadcast, it sends a query command to the OS through the PCIe bus to request to obtain the current status information of the CXL memory device, and the OS returns the current status information of the CXL memory device according to the query command, including temperature, hardware status, power status, memory usage and network connection status;

[0015] BMC compares the difference between each current status information and the safety range. If each current status information of the CXL memory device is within the safety range, BMC will send an initialization signal to CPLD. After receiving the initialization signal, CPLD performs the initialization operation and displays the "SETUP" prompt message on the LCD display, indicating that the device is initializing.

[0016] After BMC transmits the initialization signal, it further checks the status information of the CXL memory device after initialization through the PCIe bus and analyzes the difference between the status information after initialization and the safety range. If the status information after initialization is all within the safety range, the LED indicator turns green and the LCD screen displays the prompt message "NORMAL", indicating that the CXL memory device can work normally and proceed to the next step. Otherwise, re-initialize the inserted CXL memory device.

[0017] As a preferred solution of the out-of-band management method of the CXL memory device of the present invention, wherein: the real-time monitoring of the initialized CXL memory device status and fault repair means that after the CXL memory device is initialized, the CXL memory device status is periodically updated and monitored by the BMC, the acquired periodic device information is compared with the safety range, and the comparison result is transmitted to the CPLD, and the CPLD lights up the relevant LED signal lamp according to the comparison result and displays the prompt information on the LCD display screen;

[0018] When the periodic status information in the CXL memory device is out of the safe range, it indicates that the CXL memory device has failed. The BMC receives the failure information from the OS and controls the PCIe bus to put the data channel of the failed CXL memory device into read-only mode for isolation operation.

[0019] Fault information includes expected fault information and unexpected fault information;

[0020] Expected fault information includes temperature exceeding the limit, memory usage exceeding the limit, and network connection error. If any of the expected fault information occurs, the BMC will determine that the CXL memory device is in a fault state and send a fault signal to the CPLD. The LED indicator turns yellow and the "WARNING" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to perform maintenance.

[0021] Unexpected fault information includes hardware status fault and power status fault. If any of the unexpected fault information occurs, the BMC will determine that the CXL memory device is in an error state and send an error signal to the CPLD. The LED indicator turns red and the "ERROR" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to repair it immediately.

[0022] The operation and maintenance personnel repair the device according to the prompt information. The repair process includes physical hardware replacement and software reset. After the repair, the BMC re-detects the status of the CXL memory device to confirm the repair of the fault.

[0023] As a preferred solution of the out-of-band management method of the CXL memory device of the present invention, wherein: the device resetting and verification of the CXL memory device after the fault is repaired means that after the repair is completed, the BMC releases the isolation operation of the device and sends a reset signal to the CPLD;

[0024] After receiving the reset signal, CPLD controls the CXL memory device management system to reinitialize, BMC resumes normal monitoring of CXL memory devices, and displays the "RESETUP" status information on the LCD display;

[0025] After the CXL memory device is reset, the BMC verifies the status of the CXL memory device to ensure that all status information is within the normal range;

[0026] After verification is complete, the LED indicator turns green and the LCD display updates to "NORMAL", indicating that the CXL memory device has resumed normal operation.

[0027] As a preferred solution of the out-of-band management method of the CXL memory device described in the present invention, wherein: the combining of multiple CXL memory devices to form a resource pool and performing real-time monitoring and hot maintenance through the CXL memory device management system refers to combining multiple CXL memory devices to form a resource pool, and monitoring the total utilization of the resource pool and the task status of each CXL memory device in real time through the CXL memory device management system. The total utilization of the resource pool evaluates the overall load and operating status of the resource pool by integrating the memory utilization and network bandwidth occupancy indicators of each CXL memory device. If a CXL memory device fails, the CXL memory device management system migrates the tasks being executed on the failed device to the low-load CXL memory device through hot maintenance.

[0028] As a preferred solution of the out-of-band management method of the CXL memory device of the present invention, the recording and storing of all data of the CXL memory device refers to recording the insertion timestamp, PCIe bus signal change and initial status information of the CXL memory device during the insertion of the CXL memory device, and periodically saving the status information of each CXL memory device in real-time monitoring, recording the fault information when a CXL memory device fails, recording the entire process of task migration, device repair and reset when performing hot maintenance operations, and saving the total utilization data of the resource pool and the task status data of each CXL memory device.

[0029] Another object of the present invention is to provide an out-of-band management system for CXL memory devices, comprising:

[0030] An initialization module is used to determine the insertion status of a CXL memory device in real time through the BMC and initialize the inserted CXL memory device;

[0031] Real-time management module, used to monitor the status of initialized CXL memory devices in real time and perform fault repair;

[0032] The device reset module is used to reset and verify the CXL memory device after the fault is repaired;

[0033] Resource pool module, which is used to combine multiple CXL memory devices into a resource pool and perform real-time monitoring and thermal maintenance through the CXL memory device management system;

[0034] The data storage module is used to record and store all data of the CXL memory device.

[0035] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of an out-of-band management method for a CXL memory device when executing the computer program.

[0036] A computer-readable storage medium stores a computer program, which implements the steps of an out-of-band management method for a CXL memory device when executed by a processor.

[0037] The beneficial effects of the present invention are as follows: the CXL memory device management method proposed in the present invention can effectively manage the connected CXL memory devices, more intuitively understand the current CXL memory device status or CXL resource pool utilization and allocation, and the hot maintenance method can avoid the server being disabled. The maintenance method is more convenient, supports remote monitoring and automatic management, and greatly improves the stability and availability of the server. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0039] Figure 1 A flowchart of an out-of-band management method for a CXL memory device is shown.

[0040] Figure 2 Schematic diagram of the CXL memory device management workflow.

[0041] Figure 3 A schematic diagram of the out-of-band management system for CXL memory devices. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments. Example

[0045] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and which provides an out-of-band management method for a CXL memory device. The out-of-band management method for a CXL memory device includes:

[0046] S1, judging the insertion status of the CXL memory device in real time through the BMC and initializing the inserted CXL memory device;

[0047] Specifically, judging the insertion status of the CXL memory device in real time through the BMC means that the voltage and signal integrity of the PCle bus are monitored in real time through the BMC in the CXL memory device management system by using the electrical signal detection method. The BMC is a baseboard management controller, which ensures that the system is in a normal operating state by monitoring the power supply and temperature of the CXL memory device management system. In the scenario of CXL memory device insertion detection, the electrical signal detection method can quickly respond to the event of device insertion or disconnection, and can effectively reduce false detection and improve detection accuracy through signal verification and denoising processing;

[0048] The CXL memory device refers to a memory device based on the CXL interface specification, which can increase the storage capacity of each server by up to 10 times. The advantage of CXL memory lies in its high bandwidth and low latency, which enables it to meet the needs of applications such as high-performance computing and artificial intelligence. Through CXL technology, the system can achieve higher memory utilization and better resource management, thereby improving overall system performance and efficiency;

[0049] Traditional memory is directly connected to the CPU and belongs to the data access of a single system. CXL technology can expand memory through the CXL bus. When used with a CXL switch, memory pooling can be achieved for data access of multiple systems.

[0050] Traditional memory management is done in-band, that is, through the CPU. The CXL memory pooling system is a decoupled system, and it becomes difficult to manage through a certain CPU. Therefore, the present invention mainly introduces an out-of-band management method, which can be used to achieve unified management of CXL memory devices.

[0051] When the signal of the PCle bus changes and meets the standard for device insertion, the BMC uses the time domain analysis method to detect the change rate of the rising and falling edges of the signal. When a CXL memory device is inserted, the signal level change on the PCIe bus shows a fast rising and falling edge. A change rate threshold is set. If the change rate exceeds the change rate threshold, it is determined that a CXL memory device is inserted. The signal changes of the PCle bus include changes in electrical signals, changes in the impedance of the signal line, changes in the link state, handshake signals, and changes in the bus idle state;

[0052] According to the analysis results, if it is determined that no CXL memory device is inserted, the CXL memory device management system prompts that there is no device at present, and displays the "IDEL" prompt message on the LCD display. If it is determined that a CXL memory device is inserted, the CXL memory device management system enters the device insertion state, and the BMC broadcasts the insertion event to other system components, including the operation and maintenance system, CPLD and OS, through the PCle bus.

[0053] The BMC can determine the insertion status of the CXL memory device in real time, and use the electrical signal detection method and time domain analysis method to accurately identify the changes in PCIe bus signals, effectively reduce false detections, improve detection accuracy, and ensure rapid response during device insertion and stable operation of the system. By recording insertion events and timestamps, it provides data support for subsequent system management and troubleshooting, thereby improving the maintainability and reliability of the overall system.

[0054] Furthermore, initializing the inserted CXL memory device means that after the BMC completes the insertion event broadcast, it sends a query command to the OS through the PCIe bus to request the current status information of the CXL memory device. The OS returns the current status information of the CXL memory device according to the query command, including temperature, hardware status, power status, memory usage and network connection status;

[0055] The BMC compares the difference between each current state information and the safety range. The safety range refers to the safe working range of each current state information of the CXL memory device during operation. The safe working range is defined by the design specifications, operation manuals and industry standards of the CXL memory device to ensure that the device will not fail or perform poorly when operating within this range.

[0056] If each current status information of the CXL memory device is within the safe range, the BMC will send an initialization signal to the CPLD. After receiving the initialization signal, the CPLD performs the initialization operation and displays the "SETUP" prompt message on the LCD display, indicating that the device is initializing. The CPLD is a digital integrated circuit used to process signals from the BMC and control external hardware to display the status of the device.

[0057] After BMC transmits the initialization signal, it further checks the status information of the CXL memory device after initialization through the PCIe bus and analyzes the difference between the status information after initialization and the safety range. If the status information after initialization is all within the safety range, the LED indicator turns green and the LCD screen displays the prompt message "NORMAL", indicating that the CXL memory device can work normally and proceed to the next step. Otherwise, re-initialize the inserted CXL memory device.

[0058] The inserted CXL memory device is initialized through BMC to ensure that the current status information of the device is consistent with the safety range and to prevent the device from running in an abnormal state. If the status information of the device after initialization meets the safety standards, the system sets the device status to normal and proceeds to the next step. Through this process, the system effectively improves the safety and stability of the device, ensures that the device can run in the best state, and promptly corrects abnormal conditions in initialization to prevent potential failures.

[0059] S2, monitor the status of the initialized CXL memory device in real time and perform fault repair, and reset and verify the CXL memory device after the fault repair;

[0060] Specifically, real-time monitoring of the initialized CXL memory device status and fault repair means that after the CXL memory device is initialized, the CXL memory device status is periodically updated and monitored through the BMC, the acquired periodic device information is compared with the safety range, and the comparison result is transmitted to the CPLD, and the CPLD lights up the relevant LED signal lamp according to the comparison result and displays the prompt information on the LCD display screen;

[0061] When the periodic status information in the CXL memory device is out of the safe range, it indicates that the CXL memory device has failed. The BMC receives the failure information from the OS and controls the PCIe bus to put the data channel of the failed CXL memory device into read-only mode for isolation operation.

[0062] Fault information includes expected fault information and unexpected fault information;

[0063] Expected fault information includes temperature exceeding the limit, memory usage exceeding the limit, and network connection error. If any of the expected fault information occurs, the BMC will determine that the CXL memory device is in a fault state and send a fault signal to the CPLD. The LED indicator turns yellow and the "WARNING" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to perform maintenance.

[0064] Unexpected fault information includes hardware status fault and power status fault. If any of the unexpected fault information occurs, the BMC will determine that the CXL memory device is in an error state and send an error signal to the CPLD. The LED indicator turns red and the "ERROR" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to repair it immediately.

[0065] The operation and maintenance personnel repair the device according to the prompt information. The repair process includes physical hardware replacement and software reset. After the repair, the BMC re-detects the status of the CXL memory device to confirm the repair of the fault.

[0066] By monitoring the status of CXL memory devices in real time and performing fault repair, the reliability and operation and maintenance efficiency of the system are improved. After the device is initialized, the BMC periodically updates and monitors the device status, compares the actual operating parameters of the device with the preset safety range, and promptly discovers potential problems. In this way, when the device's temperature, memory usage, or network connection status information exceeds the safety range, the system can quickly identify the fault and isolate the affected device to prevent the fault from spreading. The BMC transmits the fault information to the CPLD, displays it on the LCD screen in real time, and intuitively prompts the operation and maintenance personnel to take appropriate actions through the color change of the LED indicator. For different types of faults, the system can distinguish between serious faults that need to be handled immediately and issue a red alarm to ensure that the problem is quickly responded to and resolved. This process not only ensures the normal operation of the equipment, but also reduces the need for manual intervention through automated monitoring and prompts, thereby improving the operation and maintenance efficiency and the overall stability of the system.

[0067] Furthermore, resetting and verifying the CXL memory device after the fault is repaired means that after the repair is completed, the BMC releases the isolation operation of the device and sends a reset signal to the CPLD;

[0068] After receiving the reset signal, CPLD controls the CXL memory device management system to reinitialize, BMC resumes normal monitoring of CXL memory devices, and displays the "RESETUP" status information on the LCD display;

[0069] After the CXL memory device is reset, the BMC verifies the status of the CXL memory device to ensure that all status information is within the normal range;

[0070] After verification is complete, the LED indicator turns green and the LCD display updates to "NORMAL", indicating that the CXL memory device has resumed normal operation.

[0071] By resetting and verifying the status of the CXL memory device after the fault is repaired, the device is ensured to return to normal working status after repair. The BMC releases the isolation and sends a reset signal. The CPLD control system is reinitialized. After completion, the device status is verified to ensure that all parameters are within the normal range. The system confirms that the device has resumed normal operation, avoiding potential problems caused by incomplete repair and significantly improving the reliability and operational stability of the system.

[0072] S3, combining multiple CXL memory devices into a resource pool and performing real-time monitoring and thermal maintenance through the CXL memory device management system;

[0073] Specifically, combining multiple CXL memory devices to form a resource pool and performing real-time monitoring and hot maintenance through the CXL memory device management system means combining multiple CXL memory devices to form a resource pool, and monitoring the total utilization of the resource pool and the task status of each CXL memory device in real time through the CXL memory device management system. The total utilization of the resource pool evaluates the overall load and operating status of the resource pool by integrating the memory utilization and network bandwidth occupancy indicators of each CXL memory device. If a CXL memory device fails, the CXL memory device management system migrates the tasks being executed on the failed device to a low-load CXL memory device through hot maintenance.

[0074] By combining multiple CXL memory devices to form a resource pool and using the CXL memory device management system for real-time monitoring and hot maintenance, efficient resource management and reliable system operation are achieved. The management system can monitor the total utilization of the resource pool and the task status of each CXL memory device in real time. By integrating the memory utilization, temperature and network bandwidth occupancy of each device, the overall load and operating status of the resource pool can be accurately evaluated. This monitoring mechanism enables the system to promptly discover and identify any possible fault conditions, ensuring that the resource pool can remain stable and efficient when running under high load. When a CXL memory device fails, the management system seamlessly migrates the tasks being executed on the failed device to other low-load CXL memory devices through hot maintenance, avoiding task interruption and system downtime. This not only ensures the continuity of the system and the uninterrupted operation of the business, but also improves the utilization efficiency of resources and reduces the risk of performance degradation due to device failure. Through this intelligent resource management and fault handling mechanism, the overall reliability, maintainability and resource utilization efficiency of the system have been significantly improved.

[0075] S4, record and store all data of the CXL memory device;

[0076] Specifically, recording and storing all data of the CXL memory device means recording the insertion timestamp, PCIe bus signal changes and initial status information of the CXL memory device during the CXL memory device insertion process, and periodically saving the status information of each CXL memory device in real-time monitoring. When a CXL memory device fails, the failure information is recorded. When performing hot maintenance operations, the entire process of task migration, device repair and reset is recorded, and the total utilization data of the resource pool and the task status data of each CXL memory device are saved.

[0077] By recording and storing all the data of CXL memory devices, the system can achieve comprehensive status tracking and historical data management, greatly improving the reliability and management efficiency of the system. During the device insertion process, the system records the insertion timestamp, PCIe bus signal changes and initial status information in detail. The preservation of this data ensures the system's accurate monitoring of the device status and subsequent event backtracking capabilities. In real-time monitoring, the system periodically saves the status information of each CXL memory device, such as memory utilization, temperature and bandwidth occupancy, providing a complete historical record of the device operation. When a CXL memory device fails, the system not only records the type of failure, but also saves the entire process of fault handling to ensure that each fault handling link is documented. When performing hot maintenance operations, the system records every step of task migration, device repair and reset to ensure the transparency and traceability of the device maintenance process. The system also saves the total utilization of the resource pool and the task status of each CXL memory device, allowing operation and maintenance personnel to use these historical data for analysis and diagnosis, thereby optimizing system operation and improving maintenance efficiency. Example

[0078] Reference Figure 3 , which is a second embodiment of the present invention, and which is different from the previous embodiment, provides an out-of-band management system for a CXL memory device, including:

[0079] An initialization module is used to determine the insertion status of a CXL memory device in real time through the BMC and initialize the inserted CXL memory device;

[0080] Real-time management module, used to monitor the status of initialized CXL memory devices in real time and perform fault repair;

[0081] The device reset module is used to reset and verify the CXL memory device after the fault is repaired;

[0082] Resource pool module, which is used to combine multiple CXL memory devices into a resource pool and perform real-time monitoring and thermal maintenance through the CXL memory device management system;

[0083] The data storage module is used to record and store all data of the CXL memory device.

[0084] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0085] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0086] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0087] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0088] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An out-of-band management method for a CXL memory device, characterized in that: include, Use BMC to determine the insertion status of CXL memory devices in real time and initialize the inserted CXL memory devices; Monitor the status of the initialized CXL memory device in real time and perform fault repair, and reset and verify the CXL memory device after the fault repair; Combine multiple CXL memory devices into a resource pool and perform real-time monitoring and thermal maintenance through the CXL memory device management system; Record and store all data of CXL memory devices; The real-time monitoring of the initialized CXL memory device status and fault repair means that after the CXL memory device is initialized, the CXL memory device status is periodically updated and monitored through the BMC, the acquired periodic device information is compared with the safety range, and the comparison result is transmitted to the CPLD, and the CPLD lights up the relevant LED signal lights according to the comparison result and displays the prompt information on the LCD display screen; When the periodic status information in the CXL memory device is out of the safe range, it indicates that the CXL memory device has failed. The BMC receives the failure information from the OS and controls the PCIe bus to put the data channel of the failed CXL memory device into read-only mode for isolation operation. Fault information includes expected fault information and unexpected fault information; Expected fault information includes temperature exceeding the limit, memory usage exceeding the limit, and network connection error. If any of the expected fault information occurs, the BMC will determine that the CXL memory device is in a fault state and send a fault signal to the CPLD. The LED indicator turns yellow and the "WARNING" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to perform maintenance. Unexpected fault information includes hardware status fault and power status fault. If any of the unexpected fault information occurs, the BMC will determine that the CXL memory device is in an error state and send an error signal to the CPLD. The LED indicator turns red and the "ERROR" prompt message is displayed on the LCD screen, prompting the operation and maintenance personnel to repair it immediately. The operation and maintenance personnel repair the device according to the prompt information. The repair process includes physical hardware replacement and software reset. After the repair, the BMC re-detects the status of the CXL memory device to confirm the repair of the fault; Combining multiple CXL memory devices to form a resource pool and performing real-time monitoring and hot maintenance through the CXL memory device management system refers to combining multiple CXL memory devices to form a resource pool, and monitoring the total utilization of the resource pool and the task status of each CXL memory device in real time through the CXL memory device management system. The total utilization of the resource pool evaluates the overall load and operating status of the resource pool by integrating the indicators of memory utilization and network bandwidth occupancy of each CXL memory device. If a CXL memory device fails, the CXL memory device management system migrates the tasks being executed on the failed device to a low-load CXL memory device through hot maintenance.

2. The out-of-band management method of a CXL memory device according to claim 1, characterized in that: The real-time determination of the insertion status of the CXL memory device through the BMC refers to real-time monitoring of the voltage and signal integrity of the PCle bus by the BMC in the CXL memory device management system using an electrical signal detection method; When the signal of the PCle bus changes and meets the standard for device insertion, the BMC uses the time domain analysis method to detect the change rate of the rising and falling edges of the signal. When a CXL memory device is inserted, the signal level change on the PCIe bus shows a fast rising and falling edge. A change rate threshold is set. If the change rate exceeds the change rate threshold, it is determined that a CXL memory device is inserted; According to the analysis results, if it is determined that no CXL memory device is inserted, the CXL memory device management system prompts that there is no device at present, and displays the "IDEL" prompt message on the LCD display. If it is determined that a CXL memory device is inserted, the CXL memory device management system enters the device insertion state, and the BMC broadcasts the insertion event to other system components, including the operation and maintenance system, CPLD and OS, through the PCle bus.

3. The out-of-band management method of a CXL memory device according to claim 2, characterized in that: Initializing the inserted CXL memory device means that after the BMC completes the insertion event broadcast, it sends a query command to the OS through the PCIe bus to request to obtain the current status information of the CXL memory device, and the OS returns the current status information of the CXL memory device according to the query command, including temperature, hardware status, power status, memory usage and network connection status; BMC compares the difference between each current status information and the safety range. If each current status information of the CXL memory device is within the safety range, BMC will send an initialization signal to CPLD. After receiving the initialization signal, CPLD performs the initialization operation and displays the "SETUP" prompt message on the LCD display, indicating that the device is initializing. After BMC transmits the initialization signal, it further checks the status information of the CXL memory device after initialization through the PCIe bus and analyzes the difference between the status information after initialization and the safety range. If the status information after initialization is all within the safety range, the LED indicator turns green and the "NORMAL" prompt message is displayed on the LCD display, indicating that the CXL memory device can work normally and proceed to the next step. Otherwise, re-initialize the inserted CXL memory device.

4. The out-of-band management method of a CXL memory device according to claim 3, characterized in that: Resetting and verifying the CXL memory device after the fault is repaired means that after the repair is completed, the BMC releases the isolation operation of the device and sends a reset signal to the CPLD; After receiving the reset signal, CPLD controls the CXL memory device management system to reinitialize, BMC resumes normal monitoring of CXL memory devices, and displays "RESETUP" status information on the LCD display; After the CXL memory device is reset, the BMC verifies the status of the CXL memory device to ensure that all status information is within the normal range; After verification is complete, the LED indicator turns green and the LCD display updates to "NORMAL", indicating that the CXL memory device has resumed normal operation.

5. The out-of-band management method of a CXL memory device according to claim 4, characterized in that: The recording and storing of all data of the CXL memory device refers to recording the insertion timestamp, PCIe bus signal change and initial status information of the CXL memory device during the CXL memory device insertion process, and periodically saving the status information of each CXL memory device in real-time monitoring, recording the fault information when a CXL memory device fails, recording the entire process of task migration, device repair and reset during hot maintenance operations, and saving the total utilization data of the resource pool and the task status data of each CXL memory device.

6. An out-of-band management system for a CXL memory device based on the out-of-band management method for a CXL memory device according to any one of claims 1 to 5, characterized in that: include, An initialization module is used to determine the insertion status of a CXL memory device in real time through the BMC and initialize the inserted CXL memory device; Real-time management module, used to monitor the status of initialized CXL memory devices in real time and perform fault repair; The device reset module is used to reset and verify the CXL memory device after the fault is repaired; Resource pool module, which is used to combine multiple CXL memory devices into a resource pool and perform real-time monitoring and thermal maintenance through the CXL memory device management system; The data storage module is used to record and store all data of the CXL memory device.

7. A computer device comprising: Memory and processor; The memory stores a computer program, wherein the processor implements the steps of the out-of-band management method of a CXL memory device according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the out-of-band management method of a CXL memory device according to any one of claims 1 to 5 are implemented.

Images