Method, device and equipment for testing read-write pressure of solid state disk and storage medium

Abstract

The invention discloses a solid state disk read-write pressure test method, device and equipment and a storage medium, and the method comprises the steps: loading predefined parameterized test configuration to apply a read-write pressure test to a to-be-tested solid state disk; a pressure test is executed, test parameters are adjusted in real time based on a preset dynamic change strategy in the test process, and multi-dimensional performance data and system health state data of the to-be-tested solid state disk are collected in real time; performing intelligent analysis based on the multi-dimensional performance data and the system health state data to obtain an analysis result; when the analysis result comprises an abnormal result, a predefined coping strategy is automatically triggered, and a diagnosis report containing the abnormal result is generated. According to the method, the test parameters are dynamically optimized, and the result is intelligently analyzed, so that the intelligence and the self-adaption of the test process are realized, the performance bottleneck and the potential fault of the solid state disk can be comprehensively and efficiently found, and the accuracy and the efficiency of the test are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of solid-state drive (SSD) testing technology, and in particular to a method, apparatus, device, and storage medium for testing the read and write stress of a solid-state drive. Background Technology

[0002] Solid-state drives (SSDs), as core storage components of computer systems, are widely used in consumer electronics, enterprise storage, and data centers due to their high performance, low latency, and shock resistance. The performance and reliability of SSDs directly impact the performance of the entire computing system; therefore, rigorous stress testing is crucial during their research, development, quality control, and selection. While traditional SSD performance testing methods can simulate certain load pressures through parameterized configuration, they have significant shortcomings in terms of test realism and problem diagnosis capabilities.

[0003] The current mainstream testing solutions have the following technical shortcomings: First, the load models used in the testing process are usually static and cannot effectively simulate the load fluctuations and sudden traffic characteristics that exist in the real business environment, resulting in a significant deviation between the test results and the actual operating performance; Second, when performance degradation or abnormal events occur during the testing process, the existing solutions lack intelligent diagnostic and response capabilities, and still rely heavily on the experience of engineers for manual processing, which is inefficient and difficult to guarantee accuracy. Summary of the Invention

[0004] This invention provides a method, apparatus, computer equipment, and storage medium for testing the read and write stress of a solid-state drive (SSD), in order to solve the technical problems of existing SSD testing modes being fixed and singular, and having low levels of automation and intelligence.

[0005] Firstly, a method for testing the read / write stress of a solid-state drive is provided, including: Load predefined parameterized test configurations to apply read and write stress tests to the solid-state drive under test; Perform stress tests and adjust test parameters in real time based on preset dynamic change strategies during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time; Intelligent analysis is performed based on multi-dimensional performance data and system health status data to obtain analysis results; When the analysis results include abnormal results, predefined response strategies are automatically triggered and a diagnostic report containing the abnormal results is generated.

[0006] Secondly, a solid-state drive read / write stress testing device is provided, comprising: The loading module is used to load predefined parameterized test configurations to apply read and write stress tests to the solid-state drive under test. The execution module is used to perform stress tests and adjust test parameters in real time based on a preset dynamic change strategy during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time. The analysis module is used to perform intelligent analysis based on multi-dimensional performance data and system health status data to obtain analysis results. The results output module is used to automatically trigger predefined response strategies and generate a diagnostic report containing the abnormal results when the analysis results include abnormal results.

[0007] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described solid-state drive read / write stress test method.

[0008] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described solid-state drive read / write stress test method.

[0009] The aforementioned solution for testing solid-state drive read / write stress, including its method, apparatus, equipment, and storage media, utilizes a dynamic adjustment strategy to adjust test parameters in real-time. This effectively simulates load fluctuations in a real-world business environment, making the test results more closely resemble actual application scenarios and significantly improving the realism and accuracy of the tests. Furthermore, the use of intelligent analysis methods to comprehensively analyze multi-dimensional performance data and system health status data enables automatic detection and diagnosis of anomalies, significantly enhancing the intelligence level of the tests and reducing reliance on human experience. In addition, by automatically triggering predefined response strategies and generating diagnostic reports, the entire testing process is automated, significantly improving testing efficiency and ensuring consistency and repeatability. Simultaneously, the comprehensive analysis of multi-dimensional performance data and system health status data allows for precise identification of performance bottlenecks and root causes of anomalies, providing a reliable basis for subsequent problem remediation and performance optimization. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating a solid-state drive read / write stress testing method according to an embodiment of the present invention.

[0012] Figure 2This is a schematic diagram of a solid-state drive read / write stress testing device according to an embodiment of the present invention.

[0013] Figure 3 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Please see Figure 1 As shown, Figure 1 A flowchart illustrating a solid-state drive read / write stress testing method provided in an embodiment of the present invention includes the following steps: Step S1: Load the predefined parameterized test configuration to apply read and write stress tests to the solid-state drive under test.

[0016] Specifically, the parameterized test configuration includes predefined initial test parameters, which include: basic I / O parameters, concurrency control parameters, timing control parameters, and resource monitoring parameters. These parameters specifically include read / write mode (sequential / random / mixed), block size (4K-1M), queue depth (1-256), number of threads, number of processes, task distribution strategy, total test duration, ramp-up time, warm-up time, cooldown interval, performance data sampling frequency, health status check cycle, and log recording level. In a preferred embodiment, the test configuration is stored in JSON or YAML format.

[0017] Step S2: Perform stress test, and adjust test parameters in real time based on preset dynamic change strategy during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time.

[0018] Specifically, after loading the test configuration, read and write stress tests are performed on the solid-state drive under test according to the test configuration. During the test, test parameters are dynamically optimized and multi-dimensional data is collected synchronously. The dynamic adjustment of parameters can discover boundary conditions that cannot be triggered under fixed load, thereby enabling in-depth testing.

[0019] Furthermore, in some embodiments, the test parameters are adjusted in real time based on a preset dynamic change strategy, including: simulating periodic changes of the test parameters based on a preset time function, and / or simulating sudden fluctuations of the test parameters based on a preset random function.

[0020] It should be noted that the dynamic adjustment of test parameters can be achieved based on time functions and random functions.

[0021] The time-function-based periodic variation simulation aims to simulate regular and predictable business load fluctuations in the real world. The system pre-sets a time-related variation function, causing test parameters to automatically adjust according to specific periodic patterns. For example, daily cycle simulation can simulate the load characteristics of a typical workday. During a set "working period" (e.g., 9 AM to 6 PM), the system automatically increases queue depth and the number of concurrent threads to simulate high-concurrency user access, while during "non-working periods," the load level is correspondingly reduced to simulate the system's idle state. Long-term test cycle simulation, in endurance tests lasting several days, can define a test cycle including multiple stages such as "stress ramp-up," "peak hold," and "stress decay." For example, the load is slowly increased to a peak in the first few hours of the test and then stabilized at that peak for tens of hours before gradually decreasing the load, thus verifying the performance and stability of the SSD under different stress stages. This method can systematically evaluate the adaptability of the SSD to scenarios with regular load changes, especially verifying whether its performance can smoothly transition during load switching and whether there are firmware issues that only trigger during specific load stages.

[0022] The simulation of sudden fluctuations based on random functions aims to simulate unpredictable and sudden traffic surges in real-world business environments, bringing unexpected challenges to the storage system during stress testing. The system uses a pseudo-random algorithm to randomly generate sudden spikes on a preset baseline load. For example, the load's "queue depth" or "thread count" may suddenly surge to several times the normal value within a short period, simulating instantaneous high-concurrency requests, such as a "flash sale" on an e-commerce platform or a sudden surge in trending news. In addition to sudden changes in load intensity, key parameters such as the "read-write ratio" and "block size" of I / O will also fluctuate randomly within a set range. This random combination can cover a large number of atypical parameter combinations that are difficult to exhaustively cover in manual preset tests, making it easier to discover deeply hidden boundary condition defects. Random fluctuations greatly enhance the exploratory nature of the test, effectively exposing whether there are any flaws in the robustness, resource scheduling capabilities, and error handling mechanisms of the SSD controller and firmware when dealing with chaotic and unpredictable I / O request sequences.

[0023] In practical testing, the two strategies described above can be used in conjunction, using a periodic load defined by a time function as the basic framework, and then superimposing sudden fluctuations generated by a random function on this framework. This hybrid strategy ensures that the test covers known and regular business scenarios, while also giving the test the ability to discover unknown and abnormal issues, thereby greatly improving the realism and depth of the test. This makes the test results more reflective of the real performance of SSDs in complex and ever-changing real-world application environments, thus more effectively identifying potential performance bottlenecks and reliability issues.

[0024] Furthermore, in other embodiments, test parameters are adjusted in real time based on a preset dynamic change strategy, including: 1. Retrieve historical test data of solid-state drives of the same model as the solid-state drive under test in the current test scenario from the knowledge base.

[0025] Specifically, during the testing process, historical test data of SSDs of the same model as the one currently under test are retrieved in real time from the knowledge base. The knowledge base stores a large number of complete test records for the same model of SSDs, including performance data, health status data, abnormal event records, and corresponding test parameters. The historical test data is cleaned and feature extracted to form a structured dataset, which is used as input for the subsequent policy network.

[0026] 2. Input historical test data, multi-dimensional performance data, and system health status data into the pre-trained policy network to generate the next set of test parameters. The policy network is built based on the reinforcement learning agent algorithm and is used to generate test parameters that can maximize the discovery of performance bottlenecks or potential faults.

[0027] Specifically, historical test data, real-time multi-dimensional performance data (such as IOPS, latency, bandwidth, etc.) collected during the current test, and system health status data (such as SMART parameters, temperature, error count) are all input into a pre-trained policy network. The policy network is built based on reinforcement learning agent algorithms (such as Deep Q-Network or Proximal Policy Optimization), and learns the mapping relationship between test parameters and performance bottlenecks / fault detection through joint analysis of historical and real-time data. The goal of the policy network is to generate the next set of test parameters (such as temperature change rate, load intensity, power interruption timing, etc.) to maximize the exposure of potential defects or performance degradation of the solid-state drive. The training process of the policy network adopts a reward mechanism. The reward function is designed based on indicators such as the number of abnormal events discovered during the test, the magnitude of performance fluctuations, and the rate of health decline, ensuring that the generated test parameters can effectively cover high-risk scenarios. Based on the test parameters output by the policy network, the configuration of the composite stress test is adjusted in real time, such as: dynamically coupling the rate of change of temperature stress and capacity occupancy stress; adaptively triggering the timing of power interruption events, prioritizing high-risk phases such as garbage collection or metadata updates; and adjusting the read / write ratio and queue depth of the mixed load to simulate more extreme business scenarios.

[0028] This embodiment achieves intelligent dynamic adjustment of test parameters by introducing a knowledge base and a reinforcement learning-driven policy network. It overcomes the limitations of fixed parameters and insufficient scenario coverage in traditional testing, and can more efficiently and accurately discover the performance bottlenecks and potential faults of solid-state drives under complex stress, providing deeper data support for reliability assessment.

[0029] Furthermore, the stress test process also includes injecting controllable abnormal events into the solid-state drive under test. These controllable abnormal events include at least one of the following: simulating PCIe link bandwidth degradation or training errors, simulating insufficient system memory resources, forcibly triggering the solid-state drive's internal garbage collection mechanism, and simulating a sudden power outage.

[0030] Specifically, in order to proactively and purposefully simulate various abnormal scenarios that may occur in the real world, and thus systematically verify the fault tolerance, data integrity, and fault recovery mechanism of the solid-state drive (SSD), this embodiment proactively injects controllable abnormal events into the SSD under test. These controllable abnormal events include: simulating PCIe link bandwidth degradation or training errors, simulating insufficient system memory resources, forcibly triggering the SSD's internal garbage collection mechanism, and simulating sudden power outages.

[0031] Simulating PCIe link anomalies aims to test the stability of the hardware connection between the SSD and the host. In practice, this involves artificially triggering bandwidth degradation of the PCIe link (e.g., forcing a drop from PCIe 4.0 x4 speed to PCIe 2.0 x1 speed) or periodically retraining the link using specialized tools or driver commands. This verifies whether the SSD can maintain basic data read / write functions or trigger error recovery procedures normally under conditions of unstable communication between the controller and the host or degraded signal quality, without leading to system crashes or data corruption.

[0032] Simulating insufficient system memory resources aims to test the robustness of SSD drivers under conditions of limited host system resources. By rapidly exhausting available system memory using memory stress tools or creating memory contention by adjusting kernel parameters, the system observes whether SSD I / O operations experience timeouts, failures, or deadlocks. This helps identify design flaws in the driver layer regarding resource allocation and I / O retry mechanisms.

[0033] Forced triggering of the SSD's internal garbage collection mechanism aims to evaluate the SSD's foreground performance under high background maintenance loads. By first performing large-scale sequential writes to fill the remaining space on the drive, and then immediately switching to random small-block writes, a large amount of fragmented and invalid data can be artificially created, thereby forcing the controller to initiate a high-intensity garbage collection operation. During this process, monitoring the SSD's read / write latency and IOPS stability can verify the effectiveness of its quality of service control mechanism.

[0034] Simulating sudden power outages is a crucial test for verifying data reliability. This involves simulating abnormal power failure scenarios by triggering a momentary power outage on the device via a programmable power controller or at the software level. After power is restored, the integrity of the file system, the correctness of user data, and the information in the SSD's SMART log regarding unsafe shutdown counts and various error statistics are rigorously checked. This directly tests the completeness of the SSD's capacitor protection circuitry, firmware data write strategies, and metadata protection mechanisms.

[0035] By systematically injecting the aforementioned controllable anomalies, this embodiment can proactively and efficiently expose the potential defects of solid-state drives under extreme operating conditions, providing an indispensable verification method for comprehensively evaluating their enterprise-level reliability and data security.

[0036] Step S3: Perform intelligent analysis based on multi-dimensional performance data and system health status data to obtain analysis results.

[0037] Specifically, after performing stress tests and obtaining real-time test data, the data is intelligently analyzed to obtain analysis results. This intelligent analysis process can be implemented through an analysis model or a pre-set analysis strategy.

[0038] Furthermore, in some embodiments, step S3 specifically includes: 1.1 Extract IOPS, latency data, throughput, SSD SMART attributes, CPU utilization, memory utilization, and disk queue length from multi-dimensional performance data and system health status data, and construct multi-dimensional indicators.

[0039] Specifically, key indicators are systematically extracted from multi-dimensional performance data and system health status data to form multi-dimensional indicators for in-depth analysis. These indicators include: Performance metrics include IOPS (operations per second) which measures I / O processing capacity, latency data (average latency, tail latency, etc.) which characterize response speed, throughput (bandwidth) which reflects data transfer efficiency, and disk queue length which reflects I / O request queuing. System health metrics: These include SMART attributes (such as wear level, number of bad blocks, and error count) that reflect the wear and internal condition of the solid-state drive, CPU utilization that represents the pressure on computing resources, and memory usage that indicates the degree of memory resource stress.

[0040] By cleaning, aligning, and standardizing these indicators, differences in measurement units are eliminated, resulting in a unified and well-organized multidimensional indicator dataset.

[0041] 1.2. Input multidimensional indicators as input features into a pre-trained anomaly diagnosis model for anomaly diagnosis. The anomaly diagnosis model is used to confirm whether the solid-state drive under test has anomaly results, and outputs the anomaly category when anomaly results occur. The anomaly category includes at least one of the following: solid-state drive controller bottleneck, NAND flash memory performance degradation, firmware queue management defects, system memory resource contention, driver compatibility issues, PCIe link training errors, and overheating and frequency reduction caused by insufficient heat dissipation.

[0042] Specifically, the constructed multidimensional indicators are used as input features and fed in batches into a pre-trained anomaly diagnosis model. This model is a typical machine learning classifier that performs the following key tasks: (1) Abnormal state confirmation: The model first performs a comprehensive analysis of the input features and outputs a binary judgment about the current health status of the system, that is, confirms whether the solid-state drive under test has an abnormal result.

[0043] (2) Root Cause Classification: When the model determines that the system is in an abnormal state, it will further output a specific abnormality category. This category is not a general "poor performance", but points to a clear and actionable root cause of the fault. These predefined categories include, but are not limited to: solid-state drive controller bottleneck (insufficient computing resources), NAND flash memory performance degradation (slower read and write speeds), firmware queue management defects (I / O scheduling algorithm problems), system memory resource contention (insufficient host memory affects the driver), driver compatibility issues (driver is incompatible with the system or hardware), PCIe link training errors (unstable hardware connection), and overheating throttling caused by insufficient heat dissipation (protection mechanism triggered due to excessive temperature).

[0044] This embodiment upgrades fault diagnosis from manual judgment relying on expert experience to pattern-based automatic identification through a data-driven approach. This not only significantly improves analysis efficiency but, more importantly, enables precise localization of complex and coupled faults, providing the most direct decision-making basis for subsequent targeted response strategies. The model is trained on a large number of historical fault cases with known root causes, giving it a powerful ability to extract essential characteristics from complex phenomena.

[0045] Furthermore, in some other embodiments, step S3 specifically includes: 2.1. Based on a combination of predefined monitoring indicators, analyze multi-dimensional performance data and system health status data for complex anomaly scenarios.

[0046] 2.2. An abnormal result is determined to exist in the analysis results if and only if multiple related indicators simultaneously meet the preset conditions.

[0047] Specifically, intelligent analysis is achieved through a rule-based mechanism for judging complex anomalies. Its core lies in establishing an intelligent rule system that uses multiple indicators for joint judgment, accurately identifying complex anomalies by monitoring the correlations between these indicators. The implementation includes the following two key steps: (1) Predefined composite anomaly scenarios: A series of composite anomaly scenarios based on combinations of multiple monitoring indicators need to be predefined. Each scenario is not based on a simple threshold of a single indicator, but describes a specific failure mode when multiple indicators with causal or correlational relationships occur abnormally at the same time. The system matches these predefined scenarios by performing correlation analysis on multi-dimensional performance data and system health status data.

[0048] (2) Multi-condition joint triggering judgment mechanism: The anomaly judgment is performed by strictly adhering to the logic of "if and only if multiple related indicators simultaneously meet the preset conditions". This design ensures the accuracy and reliability of anomaly judgment and effectively avoids false alarms caused by fluctuations in a single indicator. Its judgment logic has the following characteristics: Correlation: It requires that multiple technically related indicators are abnormal at the same time. For example, "increased I / O latency" must occur at the same time as "increased SSD temperature" to be judged as an overheating problem; Simultaneity: It requires that these related abnormal indicators must be detected within the same time window; Strictness: The "if and only if" condition setting ensures that only anomalies that fully meet the preset mode will be triggered. Typical examples of compound anomaly scenarios include: Overheating degradation scenario: performance degradation due to overheating is only identified when both "I / O latency exceeds threshold A" and "SSD storage media temperature exceeds threshold B" are met simultaneously; Hardware failure scenario: hardware failure is only identified when both "uncorrectable error count exceeds threshold C" and "system experiences a specific PCIe link error" occur simultaneously; Flash memory lifespan exhaustion scenario: flash memory lifespan exhaustion is only identified when both "write bandwidth continuously decreases exceeding threshold D" and "bad block count rapidly increases exceeding threshold E" occur simultaneously; Resource contention scenario: system resource contention is only identified when both "system CPU utilization exceeds threshold F" and "disk queue length remains saturated" are present simultaneously.

[0049] This embodiment, based on a composite scenario analysis method, significantly improves the accuracy of anomaly diagnosis and reduces the false alarm rate through multi-index cross-validation. Simultaneously, it effectively encodes expert domain knowledge (i.e., understanding of various fault modes) into the automated system, enabling the system to stably and reliably execute complex diagnostic tasks, thus effectively complementing and validating machine learning-based methods.

[0050] Step S4: When the analysis results include abnormal results, the predefined response strategy is automatically triggered and a diagnostic report containing the abnormal results is generated.

[0051] Specifically, this embodiment realizes closed-loop management from "intelligent diagnosis" to "automatic handling", ensuring the safety and controllability of the testing process and providing the richest on-site information for problem localization.

[0052] Furthermore, predefined response strategies are automatically triggered, including: performing corresponding response operations based on abnormal results in the analysis results, such as: automatically reducing the test load intensity, triggering a deep dump of system logs, saving a snapshot of the device driver status, or executing a safe shutdown process.

[0053] 1. The purpose of automatically reducing the test load intensity is to prevent permanent damage to the equipment due to overheating or excessive wear, ensure the safety of the test platform, and observe whether the equipment can return to normal after the load is reduced. The trigger condition is set to: when anomalies strongly correlated with sustained high loads, such as overheating and frequency throttling caused by insufficient heat dissipation or NAND flash memory performance degradation, are diagnosed. The execution action is set to: the central control unit immediately sends instructions to the I / O load generation tool (such as fio) to dynamically reduce the queue depth, reduce the number of concurrent threads, or switch to a lower-pressure I / O mode.

[0054] The purpose of triggering a deep system log dump is to capture the system's hardware and software state information at the moment the anomaly occurs, providing developers with crucial first-hand information for subsequent root cause analysis. The trigger condition is set to: when an anomaly related to the system software or driver layer is diagnosed, such as driver compatibility issues or firmware queue management defects. The execution action is set to: automatically execute a script to collect all log information in memory, including the operating system kernel log (dmesg), system logs (syslog / journalctl), and SSD vendor-specific driver logs, and save it to a specified file.

[0055] The purpose of saving a snapshot of the device driver state is to freeze the internal state of the driver at the time of the anomaly, helping to reproduce and debug complex driver-level or link-level faults. The trigger condition is set to: when a PCIe link training error is diagnosed or a driver state anomaly is suspected. The action to be performed is: through the Device Manager interface or a dedicated command-line tool, export the current configuration parameters, internal state machine information, buffer states, etc., of the SSD device driver and save them as a snapshot file.

[0056] The purpose of executing a safe shutdown procedure is to maximize the protection of test data and hardware assets, prevent the escalation of faults, and serve as a final safety measure. The trigger condition is set when an extremely serious anomaly is diagnosed, such as data consistency verification failure after multiple power interruptions, or the device becoming completely unresponsive, and continued testing may lead to data corruption or hardware damage. The execution action is set as follows: the system automatically terminates all test loads, waits for ongoing I / O operations to complete, and then safely shuts down the test platform power supply via operating system commands, or performs a controlled power-off on the SSD under test via a power circulator.

[0057] This embodiment adjusts test parameters in real time through a dynamic change strategy, effectively simulating load fluctuations in a real business environment. This makes the test results more closely resemble actual application scenarios, significantly improving the realism and accuracy of the test. Furthermore, it employs intelligent analysis methods to comprehensively analyze multi-dimensional performance data and system health status data, enabling automatic detection and diagnosis of anomalies. This significantly enhances the intelligence level of the test and reduces reliance on human experience. In addition, by automatically triggering predefined response strategies and generating diagnostic reports, the entire testing process is automated, significantly improving testing efficiency and ensuring consistency and repeatability. Simultaneously, the comprehensive analysis of multi-dimensional performance data and system health status data accurately pinpoints performance bottlenecks and root causes of anomalies, providing a reliable basis for subsequent problem remediation and performance optimization.

[0058] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0059] In one embodiment, a read / write stress testing device for a solid-state drive (SSD) is provided, which corresponds one-to-one with the read / write stress testing method for the SSD described in the above embodiments. For example... Figure 2 As shown, the read / write stress testing device for the solid-state drive includes a loading module 11, an execution module 12, an analysis module 13, and a result output module 14.

[0060] Loading module 11 is used to load predefined parameterized test configurations to apply read and write stress tests to the solid-state drive under test; The execution module 12 is used to perform stress tests and adjust test parameters in real time based on a preset dynamic change strategy during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time. Analysis module 13 is used to perform intelligent analysis based on multi-dimensional performance data and system health status data to obtain analysis results; The results output module 14 is used to automatically trigger predefined response strategies and generate a diagnostic report containing the abnormal results when the analysis results include abnormal results.

[0061] Optionally, the execution module 12 performs operations to adjust the test parameters in real time based on a preset dynamic change strategy, specifically including: simulating the periodic changes of the test parameters based on a preset time function, and / or simulating the sudden fluctuations of the test parameters based on a preset random function.

[0062] Optionally, the execution module 12 performs operations to adjust the test parameters in real time based on a preset dynamic change strategy. Specifically, this includes: obtaining historical test data of solid-state drives of the same model as the solid-state drive under test in the current test scenario from the knowledge base; inputting the historical test data, multi-dimensional performance data, and system health status data into a pre-trained strategy network to generate the next set of test parameters. The strategy network is constructed based on a reinforcement learning agent algorithm and is used to generate test parameters that can maximize the discovery of performance bottlenecks or potential faults.

[0063] Optionally, the operation of performing stress testing by the execution module 12 further includes: injecting a controllable abnormal event into the solid-state drive under test, the controllable abnormal event including at least one of: simulating PCIe link bandwidth degradation or training error, simulating insufficient system memory resources, forcibly triggering the solid-state drive's internal garbage collection mechanism, and simulating a sudden power outage.

[0064] Optionally, the analysis module 13 performs intelligent analysis based on multi-dimensional performance data and system health status data to obtain analysis results. Specifically, this includes: extracting IOPS, latency data, throughput, SSD SMART attributes, CPU utilization, memory utilization, and disk queue length from the multi-dimensional performance data and system health status data to form multi-dimensional indicators; inputting the multi-dimensional indicators as input features into a pre-trained anomaly diagnosis model for anomaly diagnosis. The anomaly diagnosis model is used to confirm whether the SSD under test has abnormal results, and outputs the anomaly category when abnormal results occur. The anomaly category includes at least one of the following: SSD controller bottleneck, NAND flash memory performance degradation, firmware queue management defects, system memory resource contention, driver compatibility issues, PCIe link training errors, and overheating and frequency reduction due to insufficient heat dissipation.

[0065] Optionally, the analysis module 13 performs intelligent analysis based on multi-dimensional performance data and system health status data to obtain analysis results. Specifically, it includes: analyzing multi-dimensional performance data and system health status data based on a composite abnormal scenario of multiple predefined monitoring indicators; and determining that there are abnormal results in the analysis results if and only if multiple related indicators simultaneously meet preset conditions.

[0066] Optionally, the result output module 14 performs operations that automatically trigger predefined response strategies, specifically including: performing corresponding response operations based on abnormal results in the analysis results, such as: automatically reducing the test load intensity, triggering a deep dump of system logs, saving a snapshot of the device driver status, or performing a safe shutdown process.

[0067] Specific limitations regarding the read / write stress testing device for solid-state drives (SSDs) can be found in the limitations of the SSD read / write stress testing method described above, and will not be repeated here. Each module in the aforementioned SSD read / write stress testing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0068] In one embodiment, a computer device is provided, the internal structure of which can be shown in the following diagram. Figure 3 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it performs the following steps: Load predefined parameterized test configurations to apply read and write stress tests to the solid-state drive under test; Perform stress tests and adjust test parameters in real time based on preset dynamic change strategies during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time; Intelligent analysis is performed based on multi-dimensional performance data and system health status data to obtain analysis results; When the analysis results include abnormal results, predefined response strategies are automatically triggered and a diagnostic report containing the abnormal results is generated.

[0069] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Load predefined parameterized test configurations to apply read and write stress tests to the solid-state drive under test; Perform stress tests and adjust test parameters in real time based on preset dynamic change strategies during the test, and collect multi-dimensional performance data and system health status data of the solid-state drive under test in real time; Intelligent analysis is performed based on multi-dimensional performance data and system health status data to obtain analysis results; When the analysis results include abnormal results, predefined response strategies are automatically triggered and a diagnostic report containing the abnormal results is generated.

[0070] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0071] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other storage media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0072] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0073] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A read-write stress test method of a solid state disk, characterized by, The method comprises the following steps: loading a predefined parameterized test configuration to apply read-write stress testing to a solid state disk to be tested; performing stress testing and adjusting test parameters in real time based on a preset dynamic change strategy during testing, and collecting multi-dimensional performance data and system health state data of the solid state disk to be tested in real time; intelligently analyzing the multi-dimensional performance data and the system health state data to obtain an analysis result; when the analysis result includes an abnormal result, automatically triggering a predefined response strategy and generating a diagnostic report containing the abnormal result. 2.The read-write stress test method of a solid state drive according to claim 1, wherein, The real-time adjustment of test parameters based on the preset dynamic change strategy comprises: simulating periodic changes in test parameters based on a preset time function, and / or simulating sudden fluctuations in test parameters based on a preset random function. 3.The read-write stress test method of a solid state drive according to claim 1, wherein, The real-time adjustment of test parameters based on the preset dynamic change strategy comprises: obtaining historical test data of a solid state disk of the same model as the solid state disk to be tested under the current test scenario from a knowledge base; inputting the historical test data, the multi-dimensional performance data and the system health state data into a pre-trained strategy network to generate a next set of test parameters, the strategy network being constructed based on a reinforcement learning agent algorithm, and the strategy network being used to generate test parameters that can maximize the discovery of performance bottlenecks or potential faults. 4.The read-write stress test method of a solid state drive according to claim 1, wherein, The stress testing further comprises: injecting controllable abnormal events into the solid state disk to be tested, the controllable abnormal events including at least one of simulating PCIe link bandwidth degradation or training errors, simulating insufficient system memory resources, forcibly triggering a solid state disk internal garbage collection mechanism, and simulating sudden power failure. 5.The read-write stress test method of a solid state drive according to claim 1, wherein, The intelligent analysis of the multi-dimensional performance data and the system health state data to obtain an analysis result comprises: extracting IOPS, delay data, throughput, solid state disk SMART attributes, CPU occupancy, memory usage and disk queue length from the multi-dimensional performance data and the system health state data, and forming multi-dimensional indicators; inputting the multi-dimensional indicators as input features into a pre-trained abnormal diagnosis model for abnormal diagnosis, the abnormal diagnosis model being used to determine whether the solid state disk to be tested has an abnormal result, and outputting an abnormal category when an abnormal result is present, the abnormal category including at least one of solid state disk controller bottleneck, NAND flash performance degradation, firmware queue management defect, system memory resource competition, driver compatibility problem, PCIe link training error and overheating caused by insufficient heat dissipation.

6. The read-write stress test method of a solid state drive according to claim 1 or 5, wherein, The intelligent analysis of the multi-dimensional performance data and the system health state data to obtain an analysis result comprises: analyzing the multi-dimensional performance data and the system health state data based on a plurality of predefined composite abnormal scenarios of monitoring indicator combinations; determining that there is an abnormal result in the analysis result only when a plurality of associated indicators simultaneously satisfy a preset condition. 7.The read-write stress test method of a solid state drive according to claim 1, wherein, The automatic triggering of a predefined response strategy comprises: According to the abnormal result in the analysis result, a corresponding coping operation is performed, and the coping operation includes automatically reducing a test load intensity, triggering a system log deep dump, saving a device driver state snapshot, or performing a safe shutdown process.

8. A read-write stress testing device for a solid state drive, characterized by, The method comprises the following steps: a loading module is configured to load a predefined parameterized test configuration to apply read-write stress testing to a solid state disk under test; an execution module is configured to execute the stress testing, and to adjust test parameters in real time based on a preset dynamic change strategy during the testing, and to collect multi-dimensional performance data and system health state data of the solid state disk under test in real time; an analysis module is configured to perform intelligent analysis based on the multi-dimensional performance data and the system health state data to obtain an analysis result; a result output module is configured to automatically trigger a predefined coping strategy and generate a diagnostic report containing the abnormal result when the analysis result includes an abnormal result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the read-write stress testing method of the solid state disk according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the steps of the read-write stress testing method of the solid state disk according to any one of claims 1 to 7.

Images