Solid state disk accelerated life test method and device

Abstract

The invention discloses a solid state disk accelerated life testing method and device, and relates to the technical field of disk accelerated life testing, and the method comprises the steps: firstly, determining a target read-write instruction and a target incubator temperature of a to-be-tested solid state disk by using a Q learning model, setting an RDT equipment temperature, and issuing the determined target read-write instruction to the to-be-tested solid state disk, therefore, the PE value of the disc is ground to the end of life of the disc in the shortest time so as to carry out corresponding reliability and / or performance test, and the problems that in the related technology, the PE value of the disc is difficult to be ground to the end of life in the shortest time, the reliability and performance test cannot be effectively carried out on the end of life of the disc, and the reliability and performance of the disc are influenced are solved. The technical problem that the market is hard to know the performance condition of hard disks such as NVMe SSD in the end-of-life period is solved, and the technical effects that the performance and the reliability performance of the disk in the end-of-life period are tested in the limited project test schedule, and the product quality is further guaranteed are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of disk accelerated life testing, and in particular to a method and device for accelerated life testing of solid-state hard drives. Background Art

[0002] Currently, enterprise-class NVMe SSDs (Non-Volatile Memory Express Solid State Drives) have attracted significant attention due to their high performance and other advantages. However, some NVMe SSDs are now reaching the end of their lifecycles, and customers are increasingly prioritizing the stability and performance of NVMe SSDs at the end of their lifecycles. Therefore, after a new product is produced, it is crucial to ensure that its performance value (PE) is quickly polished to the end of its lifecycle.

[0003] Currently, existing technologies mostly test the lifespan of SSDs by constructing hot and cold data and writing them into the SSD under test. Specifically, existing technologies can write cold and hot data separately to the SSD under test, and store the cold data in another device for backup. Hot data is continuously written to the SSD, and the number of hot data writes is recorded. Based on the number of hot data writes, the cold data in the SSD is compared with the backed-up cold data at a preset interval to see if they are consistent. If so, the hot data is continuously written to the SSD; otherwise, the test ends.

[0004] However, related technologies make it difficult to reduce the PE value of the disk to the end of its life in the shortest possible time, and are unable to effectively perform performance tests on disks at the end of their life. This makes it difficult to meet the market's urgent need to understand the performance of NVMe SSDs at the end of their life, and a solution is urgently needed. Summary of the Invention

[0005] The present application provides a method and device for accelerating the life test of a solid-state drive, so as to at least solve the technical problems in the related art, such as the difficulty in grinding the PE value of the disk to the end of its life in the shortest possible time, the inability to effectively perform reliability and performance tests on the disk at the end of its life, and the difficulty in meeting the market's urgent need to understand the performance of NVMe SSDs at the end of their life.

[0006] The present application provides a method for accelerating the life test of a solid-state drive, including the following steps: sending multiple test tool parameters to the solid-state drive to be tested, obtaining the temperature of the temperature chamber of a preset aging test device, and performing permutation and combination on the multiple test tool parameters and the temperature of the temperature chamber to obtain multiple state values; using a pre-constructed Q-learning model to traverse the multiple state values to obtain a target state value that meets the requirements of the preset number of disk erasure and write cycles, and splitting the target state value to determine the target read / write instruction and the target temperature of the temperature chamber of the solid-state drive to be tested; based on the target read / write instruction and the target temperature of the temperature chamber, performing an accelerated life test operation on the solid-state drive to be tested, monitoring the number of disk erasure and write cycles during the accelerated life test operation, and determining whether the number of disk erasure and write cycles reaches a preset erasure and write cycle threshold. Among them, when the number of disk erasure and write cycles reaches the preset erasure and write cycle threshold, performing a reliability test and / or a performance test on the solid-state drive to be tested to obtain corresponding test data, and sending the test data to a target client.

[0007] The present application further provides a device for accelerating the life test of a solid-state drive, including: a permutation and combination module, configured to send multiple test tool parameters to the solid-state drive to be tested, obtain the temperature of the temperature chamber of a preset aging test device, and perform permutation and combination on the multiple test tool parameters and the temperature of the temperature chamber to obtain multiple state values; a traversal module, configured to use a pre-constructed Q-learning model to traverse the multiple state values to obtain a target state value that meets the requirements of the preset number of disk erasure and write cycles, and split the target state value to determine the target read / write instruction and the target temperature of the temperature chamber of the solid-state drive to be tested; a life test module, configured to perform an accelerated life test operation on the solid-state drive to be tested based on the target read / write instruction and the target temperature of the temperature chamber, monitor the number of disk erasure and write cycles during the accelerated life test operation, and determine whether the number of disk erasure and write cycles reaches a preset erasure and write cycle threshold. Among them, when the number of disk erasure and write cycles reaches the preset erasure and write cycle threshold, performing a reliability test and / or a performance test on the solid-state drive to be tested to obtain corresponding test data, and sending the test data to a target client.

[0008] The present application further provides an electronic device, including: a memory, configured to store a computer program; a processor, configured to implement the steps of any one of the above methods for accelerating the life test of a solid-state drive when executing the computer program.

[0009] The present application further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above methods for accelerating the life test of a solid-state drive are implemented.

[0010] The present application also provides a computer program product, including a computer program, which when executed by a processor implements the steps of any one of the above-mentioned solid-state drive accelerated life test methods.

[0011] Through the present application, multiple test tool parameters can be sent to the solid-state drive to be tested by using a preset aging test device, and the temperature of the incubator of the preset aging test device can be obtained, and the multiple test tool parameters and the incubator temperature are arranged and combined to obtain multiple state values; the pre-constructed Q-learning model is used to traverse the multiple state values to obtain a target state value that meets the requirements of the preset number of disk erases, and the target state value is split to determine the target read / write instruction and the target incubator temperature of the solid-state drive to be tested; based on the target read / write instruction and the target incubator temperature, an accelerated life test operation is performed on the solid-state drive to be tested, and the number of disk erases during the accelerated life test operation is monitored, and it is determined whether the number of disk erases reaches a preset erase count threshold. Among them, when the number of disk erases reaches the preset erase count threshold, a reliability test and / or a performance test is performed on the solid-state drive to be tested to obtain corresponding test data, and the test data is sent to the target client. Therefore, the technical problem in the related art that it is difficult to grind the PE value of the disk to the end of its life in the shortest time, it is impossible to effectively perform reliability and performance tests on the disk at the end of its life, and it is difficult to meet the market's urgent need to understand the performance of NVMe SSDs at the end of their lives can be solved, and the technical effect of testing the performance and reliability of the disk at the end of its life during the limited project test schedule is achieved, and the bugs hidden at the end of the disk's life are dug out earlier, providing a more solid guarantee for product quality. Description of the Drawings

[0012] In order to more clearly illustrate the embodiments of the present application, the drawings required for the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0013] Figure 1 It is a flowchart of a solid-state drive accelerated life test method provided according to an embodiment of the present application;

[0014] Figure 2 It is a schematic execution logic diagram of a solid-state drive accelerated life test method provided by an embodiment of the present application;

[0015] Figure 3 It is an example diagram of a solid-state drive accelerated life test device according to an embodiment of the present application.

[0016] Among them, 10 - Solid State Drive Accelerated Life Test Device, 100 - Permutation and Combination Module, 200 - Traversal Module, 300 - Life Test Module. Detailed implementation manners

[0017] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0018] It should be noted that in the description of the present application, the terms "include", "comprise" or any other variant thereof are intended to cover non - exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device. The terms "first", "second", etc. in the present application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0019] In order to enable those skilled in the art of the present technology to better understand the solution of the present application, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0020] Combined with the specific application environment architecture or specific hardware architecture on which the execution of the solid - state drive accelerated life test method depends, the specific application environment architecture or specific hardware architecture is described herein.

[0021] An embodiment of the present application provides a solid - state drive accelerated life test method.

[0022] As Figure 1 shown, it is a flowchart of the solid - state drive accelerated life test method according to the embodiment of the present application. Among them, the solid - state drive accelerated life test method includes the following steps:

[0023] In step S101, a plurality of test tool parameters are sent to the solid - state drive to be tested, and the temperature of the incubator of the preset aging test device is obtained, and the plurality of test tool parameters and the incubator temperature are permuted and combined to obtain a plurality of state values.

[0024] In the embodiments of the present application, first, an aging test device, i.e., an RDT (Reliability Demonstration Testing) test device, can be used to send various FIO parameters (i.e., multiple test tool parameters) to the solid-state drive to be tested, and the RDT incubator temperature of the RDT test device can be obtained through the disk smart-log command, so as to perform permutation and combination operations on various FIO parameters and the RDT incubator temperature, thereby obtaining multiple state values corresponding to the Q-learning AI model (i.e., the Q-learning model).

[0025] Those skilled in the art should understand that in the embodiments of the present application, the FIO (Flexible I / O Tester) is an open-source disk performance test tool that can send read and write instructions to the disk, and can be well used to evaluate the read and write performance of storage devices (such as hard disks, SSDs) and file systems; the RDT test device is a professional device for verifying the reliability of solid-state drives, mainly used to simulate high-temperature environments to test their long-term stability and durability.

[0026] In addition, the embodiments of the present application can perform accelerated life tests on various solid-state drives such as NVMe SSDs. Taking the NVMe SSD as an example, the accelerated life test will be described in detail below.

[0027] Optionally, in an embodiment of the present application, before sending multiple test tool parameters to the solid-state drive to be tested, it further includes: inserting the solid-state drive to be tested into the aging test device and powering on the aging test device to deploy the test environment of the aging test device.

[0028] It should be noted that before performing the accelerated life test on the NVMe SSD, the embodiments of the present application need to deploy its physical test environment and software test environment.

[0029] During the process of deploying the physical test environment, the embodiments of the present application can prepare a professional test device that can adjust the environmental temperature and send instructions to the disk, i.e., the RDT test device, insert the NVMe SSD to be tested into the RDT test device, and power on the RDT test device.

[0030] Thus, the embodiments of the present application provide reliable hardware and environmental support for the smooth execution of the accelerated life test of the NVMe SSD by deploying the physical test environment.

[0031] Optionally, in an embodiment of the present application, multiple test tool parameters are sent to the solid-state drive to be tested, the temperature of the incubator of the preset aging test device is obtained, and the multiple test tool parameters and the incubator temperature are arranged and combined to obtain multiple state values, including: configuring a preset test software in the aging test device to send a test command corresponding to the test software to the solid-state drive to be tested through the test software; based on the test command, sending the multiple test tool parameters to the solid-state drive to be tested, and obtaining the temperature of the incubator of the aging test device through the solid-state drive to be tested; arranging and combining the multiple test tool parameters and the incubator temperature by running a preset script to obtain multiple state values of the Q-learning model.

[0032] In addition, the embodiments of the present application also need to deploy a software test environment.

[0033] Specifically, in the embodiment of the present application, the test software such as fio and cli can be configured in the RDT test device first, and the fio command and cli command (i.e., test command) are sent to the NVMe SSD to be tested through the configured test software such as fio and cli, so as to send multiple test tool parameters such as rw, size, bs, iodepth, and numjobs to the NVMe SSD to be tested, so as to control the script to perform permutation and combination operations on variables such as rw, size, bs, iodepth, numjobs, and the RDT incubator temperature T, and use the permutation and combination results of the FIO parameters sent to the NVMe SSD and the RDT incubator temperature as multiple state values of the Q-learning AI model, and determine the actions and rewards corresponding to each state value; then, the embodiment of the present application can input the state value, action, and reward into the Q-learning AI model, so that the Q-learning AI model gradually obtains the accelerated life test method with the fastest PE value growth rate.

[0034] In the embodiment of the present application, the rw parameter in the above multiple test tool parameters is a set of write parameter, read parameter, randwrite parameter, randread parameter, rw parameter, and randrw parameter; the size parameter is a variable from 0 to 100% with an increment of 1%; the bs parameter represents a variable from 4k to 100M with an increment of 4k, the iodepth parameter represents a variable from 1 to 1024 with an increment of 1; the numjobs parameter represents a variable from 1 to 20 with an increment of 1; the RDT incubator temperature T represents a variable from 20 to 110 with an increment of 1.

[0035] Thus, the embodiment of the present application can realize the permutation and combination of multiple FIO parameters and the RDT incubator temperature through the deployment of the software test environment, so as to obtain multiple state values of the Q-learning AI model, providing reliable data guidance and basis for finally determining the required FIO workload and incubator temperature.

[0036] In step S102, a pre-constructed Q-learning model is used to traverse multiple state values to obtain a target state value that meets the requirements of the preset number of disk write / erase cycles, and the target state value is split to determine the target read / write instructions and the target incubator temperature of the solid-state drive to be tested.

[0037] Furthermore, an embodiment of the present application can also traverse each state s through a pre-constructed Q-learning AI model (i.e., the Q-learning model) to obtain the required state value (i.e., the target state value) that meets the requirements of the preset number of disk write / erase cycles (PE, P / E Cycles) (i.e., meets the disk life cycle requirements), and split it. The number of disk write / erase cycles (i.e., the number of erase cycles) is an important indicator for evaluating the write / erase life of storage devices such as SSDs, so as to obtain the required disk temperature (i.e., the target incubator temperature) and the read / write instruction Workload sent to the NVMe SSD through the FIO tool.

[0038] Thus, the embodiment of the present application combines the disk temperature of the NVMe SSD and the issued FIO workload, and applies the Q-learning AI model, so as to efficiently and accurately obtain the required disk temperature and FIO workload.

[0039] Optionally, in an embodiment of the present application, using a pre-constructed Q-learning model to traverse multiple state values to obtain a target state value that meets the requirements of the preset number of disk write / erase cycles, and splitting the target state value to determine the target read / write instructions and the target incubator temperature of the solid-state drive to be tested includes: initializing the Q function in the Q-learning model to obtain the corresponding initialized Q function value, and determining the current action corresponding to the current state value according to the preset ε-greedy policy; executing the current action and obtaining the reward corresponding to the current action, so as to obtain the disk write / erase cycle growth rate corresponding to the current state value according to the current action and the reward, and adjusting each parameter of the current state value through the disk write / erase cycle growth rate; determining the next state value through the adjusted current state value, and updating the initialized Q function value through the next state value and the reward to obtain a new Q function value, and iteratively executing the update operation on the new Q function value to determine the target state value corresponding to the maximum disk write / erase cycle growth rate during multiple iterations; splitting the target state value to obtain the corresponding target read / write instructions and target incubator temperature.

[0040] During the actual execution process, the embodiment of the present application can traverse all combination methods of the permuted and combined state values through multiple rounds of iterative calculations of the Q-learning AI model, and obtain the final optimal state value (i.e., the target state value) by real-time monitoring of the feedback of the PE value; afterwards, the embodiment of the present application can further split the target state value, so as to clarify the target read / write instructions and the target incubator temperature of the optimal accelerated life test method.

[0041] Those skilled in the art should understand that the core of the Q-learning AI model in the embodiments of the present application mainly includes the Markov Decision Process (MDP), Q-function, Bellman equation, etc.

[0042] In the specific implementation process, the Markov Decision Process mainly involves State, Action, Reward, Transition Probability, and Discount Factor, etc. Among them, the state represents the current state of the environment, which can be denoted as s; the action represents the action that the agent can take, which can be denoted as a; the reward is the immediate reward obtained by the agent after executing the action a in the state s, which can be denoted as r; the transition probability is the probability of transferring from the state s to the state s' after executing the action a, which can be denoted as P(s'|s,a); the discount factor can be used to weigh the importance of the current reward and future rewards, which can be denoted as γ (generally, 0 ≤ γ < 1).

[0043] Secondly, in the embodiments of the present application, the Q-function Q(s,a) represents the expected value of the future cumulative reward after executing the action a in the state s; the goal of Q-learning is to find the optimal Q-function Q(s,a) so as to determine the optimal policy π(s).

[0044] In addition, it can be understood that the theoretical basis of Q-learning is the Bellman equation, and the mathematical expression of this Bellman equation is:

[0045] Q(s,a) = E[r + γmaxa'Q(s',a')]

[0046] Where r represents the immediate reward; s' represents the next state; maxa'Q(s',a') represents the Q value of selecting the optimal action in the next state s'.

[0047] It should be noted that in the embodiments of the present application, Q-learning can approximate the optimal Q-function by iteratively updating the Q-function, and its update formula is:

[0048] Q(s,a) ← Q(s,a) + α[r + γmaxa'Q(s',a') - Q(s,a)]

[0049] Where α represents the learning rate (0 < α ≤ 1), which is used to control the update step size; r + γmaxa'Q(s',a') is the target Q value; Q(s,a) represents the current Q value.

[0050] In the specific execution process, the execution process of Q-learning in the embodiments of the present application is as follows:

[0051] 1. Initialize the Q - function Q(s,a) with arbitrary values (usually set to 0).

[0052] 2. For each episode:

[0053] Initialize the state s.

[0054] For each step:

[0055] Select an action a according to the current policy (such as the ε - greedy policy).

[0056] Execute the action a, observe the reward r and the next state s′.

[0057] Update the Q - value:

[0058] Q(s,a)←Q(s,a)+α[r + γmaxa′Q(s′,a′)-Q(s,a)]

[0059] Update the state s←s′.

[0060] Until the termination state is reached.

[0061] Among them, in order to balance exploration and exploitation, in the embodiments of the present application, Q - learning usually uses the ε - greedy policy to select actions; as a specific implementable way, in the embodiments of the present application, an action is randomly selected with probability ∈ (exploration); an action with the largest current Q - value is selected with probability 1 - ∈ (exploitation), and this probability ∈ usually gradually decreases over time (such as ∈ = 0.1).

[0062] In summary, Q - learning mainly has characteristics such as model - free, off - policy, and convergence. Specifically, in the embodiments of the present application, Q - learning can directly learn through interaction without knowing the environment transition probability P(s′∣s,a) and the reward function R(s,a); secondly, when updating the Q - value, Q - learning can use maxa′Q(s′,a′), rather than the action selected by the current policy, so as to explore more possibilities; in addition, under certain conditions (such as each state - action pair being visited infinitely many times, and the learning rate gradually decreasing), Q - learning can converge to the optimal Q - function.

[0063] It can be understood that the Q - learning in the embodiments of the present application does not require an environment model, is applicable to unknown environments, and can perform flexible exploration through off - policy; in addition, the Q - learning algorithm is relatively simple and is conducive to implementation.

[0064] In the actual execution process, the embodiments of the present application can apply the Q - learning AI model to determine the target read - write instruction and the target incubator temperature to implement the best accelerated life test method.

[0065] As an implementable approach, embodiments of the present application can obtain the vendor-log and smart-log of the disk every second to obtain the PE value of the disk and the disk operating temperature value in real time by using the vendor-log and smart-log. The PE value represents the value of the disk life cycle. Currently, the PE value of newly produced NVMe SSDs starts increasing from 0. When the PE value increases to 9500, embodiments of the present application can determine that the disk has entered the end-of-life stage. The increase rate of the PE value is strongly correlated with the FIO read / write instruction workload sent to the NVMe SSD and the operating temperature of the disk. The FIO read / write instruction can achieve load switching by changing its various parameters. Secondly, the disk operating temperature value can adjust the ambient temperature through the interface of the RDT test device to adjust the temperature, and then adjust it to the disk operating temperature. By adjusting the above-defined state value s, the optimal accelerated life test method corresponding to the NVMe SSD can be determined.

[0066] Secondly, embodiments of the present application also need to define the action a of the Q-learning AI model. Specifically, embodiments of the present application can adjust the various parameters of the state s according to the growth rate of the PE value until the fastest PE increase, that is, the maximum growth rate of the PE value, is obtained. After that, embodiments of the present application can determine the target state value s in the case of the fastest PE growth rate and split the target state value s to determine what requirements the FIO read / write instruction and operating temperature of the NVMe SSD need to meet in order to achieve the optimal accelerated life test method.

[0067] In addition, in embodiments of the present application, the reward r is the rapid increase of the PE value. Since embodiments of the present application need to obtain the test method with the fastest increase of the PE value, in the Q-learning AI test model of embodiments of the present application, the reward r can gradually determine the optimal accelerated life test method by obtaining the PE value in real time and adjusting the FIO read / write instruction and the temperature of the temperature-controlled chamber T in real time.

[0068] Thus, embodiments of the present application use the Q-learning AI model to perform iterative calculations on the permutations and combinations of the generated multiple FIO read / write instructions and the RDT temperature-controlled chamber temperature to determine the NVMe SSD based on the FIO workload and the RDT device temperature value, thereby determining the optimal method for accelerating the life test of the NVMe SSD, and finally obtaining the read / write instruction and the disk temperature that cause the PE value of the NVMe SSD disk to increase to the end-of-life stage of the disk in the shortest time.

[0069] In step S103, an accelerated life test operation is performed on the solid state drive to be tested based on the target read / write instruction and the target incubator temperature, and the number of disk writes and erases during the accelerated life test operation is monitored, and it is determined whether the number of disk writes and erases reaches a preset number-of-writes threshold. Among them, when the number of disk writes and erases reaches the preset number-of-writes threshold, a reliability test and / or a performance test is performed on the solid state drive to be tested to obtain corresponding test data, and the test data is sent to the target client.

[0070] Those skilled in the art should understand that currently, NVMe SSDs have attracted much attention mainly for their high performance. However, as the number of NVMe SSD disks in the market continues to increase, customers are gradually paying more attention to the reliability and performance of the disks at the end of their life cycle. However, due to the extremely limited time for testing during the development and launch of new NVMe SSDs.

[0071] Therefore, the embodiment of the present application can, through the above-mentioned disk accelerated life test method, wear the PE value of the NVMe SSD to the end of its life cycle in the shortest time, and then, in combination with the actual needs of the customer, perform a reliability test or a performance test on the disk at the end of its life cycle, or perform tests such as reliability and performance simultaneously. Among them, the disk wearing process is a relatively severe test for a disk, and performing a stress and performance test on the disk after wearing is extremely important for verifying the stability and performance of the disk at the end of its life cycle, so as to be able to obtain very valuable and persuasive actual measurement data for the customer, and also provide sufficient actual measurement basis for judging whether the NVMe SSD product can be used without defects until the warranty period.

[0072] Optionally, in an embodiment of the present application, an accelerated life test operation is performed on the solid state drive to be tested based on the target read / write instruction and the target incubator temperature, the number of disk erasures and writes during the accelerated life test operation is monitored, and it is determined whether the number of disk erasures and writes reaches a preset erasure and write times threshold. Wherein, when the number of disk erasures and writes reaches the preset erasure and write times threshold, a reliability test and / or a performance test is performed on the solid state drive to be tested to obtain corresponding test data, and the test data is sent to the target client, including: sending the target read / write instruction to the solid state drive to be tested, and determining the environmental temperature of the aging test device according to the target incubator temperature, so as to perform an accelerated life test on the solid state drive to be tested through the target read / write instruction and the environmental temperature; monitoring the number of disk erasures and writes during the accelerated life test, and determining whether the number of disk erasures and writes is less than the preset erasure and write times threshold; if the number of disk erasures and writes is less than the preset erasure and write times threshold, continue to perform the accelerated life test on the solid state drive to be tested; if the number of disk erasures and writes is greater than or equal to the preset erasure and write times threshold, stop performing the accelerated life test on the solid state drive to be tested, and determine the actual disk test requirements, so as to perform a reliability test and / or a performance test on the solid state drive to be tested according to the actual disk test requirements to obtain corresponding test data, and send the test data to the target client.

[0073] It should be noted that after obtaining the target read / write instruction and the target incubator temperature, further, an embodiment of the present application can send the target read / write instruction to the NVMe SSD to be tested, and set the environmental temperature of the RDT test device according to the target incubator temperature T, so as to perform an accelerated life test on the NVMe SSD by using the environmental temperature of the RDT test device and the target read / write instruction, and monitor the test process in real time.

[0074] If the disk PE value is less than the preset erasure and write times threshold (such as 9500), continue to perform the accelerated life test on the NVMe SSD to be tested; when it is monitored that the disk PE value reaches 9500, the embodiment of the present application stops the accelerated life test, and flexibly switches the reliability test mode or the performance test mode in combination with the actual test requirements, so as to further verify various scenarios at the end of the disk life to obtain corresponding test data, and send the test data to the client; in addition, the embodiment of the present application can also archive the target read / write instruction and the target incubator temperature determined by the Q learning AI model to apply them to multiple NVMe SSD projects.

[0075] As an implementable approach, when conducting reliability and performance tests on disks at the end of their lifespan, embodiments of the present application can systematically perform tests on reliability (evaluating data storage stability of disks at the end of their lifespan, such as bad block rate, read / write error rate, etc.), performance (testing key indicators such as read / write speed, latency, throughput, etc.), and data integrity. The specific test process is as follows:

[0076] 1. Reliability test method:

[0077] (1) Bad block and error detection

[0078] Full disk scan: Embodiments of the present application can use badblocks or hdparm to detect physical bad blocks and logical errors;

[0079] SMART data analysis: Embodiments of the present application can read SMART (Self-Monitoring Analysis and Reporting Technology) parameters of the hard disk, such as the number of reallocated sectors (Reallocated Sectors Count, RSC), the number of uncorrectable sectors (Uncorrectable Sector Count, USC), and the seek error rate (Seek Error Rate, SER), etc., and perform data analysis on them to determine whether there are potential faults in the disk based on the analysis results, and when there are potential faults, prompt the user in a timely manner through an alarm or an attribute error code.

[0080] 2. Data integrity test:

[0081] (1) Write-read verification:

[0082] 1) Write a known data pattern (such as all 0s, all 1s, random data);

[0083] 2) Compare the checksum (such as SHA-256) after reading.

[0084] (2) Long-term stability test:

[0085] Embodiments of the present application can read and write data periodically to observe the change of the error rate over time, so as to achieve long-term stability testing.

[0086] 3. Performance test method:

[0087] (1) Basic performance indicators

[0088] 1) Sequential read / write speed: Embodiments of the present application can use dd or fio to test the continuous read / write performance of large files;

[0089] 2) Random read and write speed: The embodiments of the present application can simulate random I / O loads (such as 4K random read and write) to test the random read and write performance;

[0090] 3) Latency test: The embodiments of the present application can measure the average response time (IOPS) to test the latency performance.

[0091] (2) Stress test

[0092] 1) High-load test: The embodiments of the present application can perform high-load tests through continuous high-concurrency read and write (such as running at full load for 72 hours);

[0093] 2) Temperature impact test: The embodiments of the present application can observe performance fluctuations in high-temperature / low-temperature environments.

[0094] After that, the embodiments of the present application can convert the above reliability data (bad block rate, error rate, SMART status) and performance data (read and write speed, latency, IOPS) into structured data (such as CSV or JSON, etc.) and construct visual charts (such as line charts, heat maps) corresponding to this data to generate corresponding data reports through the structured data and visual charts and send the reports to the target users.

[0095] It should be noted that during the reliability and performance testing process, the embodiments of the present application need to avoid data leakage during the testing process (sensitive data needs to be erased), need to customize test items according to the customer's usage scenario (such as enterprise-level hard drives need to focus on testing concurrent performance), and need to comply with industry standards (such as ISO / IEC 27040 data storage security standard).

[0096] Therefore, through systematic testing (reliability scanning, performance benchmarking, life prediction) and standardized reports, the embodiments of the present application can comprehensively evaluate the state of the end-of-life disk platter and provide a decision-making basis for customers; if further automation is required, batch execution of the test process can be achieved by combining scripts (such as Python and Shell).

[0097] Thus, the embodiments of the present application perform accelerated life testing on the NVMe SSD by accelerating the wear of the PE value of the disk platter according to the obtained target read and write instructions and target incubator temperature, so that the disk platter quickly reaches the end of its life, and then verify the performance and reliability of the NVMe SSD at the end of its life.

[0098] In summary, the present application first uses the Q-learning AI model to determine the target read / write instructions and the target incubator temperature of the NVMe SSD, then sets the incubator temperature to the T value, and issues the target read / write instructions to the NVMe SSD to be tested, so that the PE value of the disk is worn to the end of the disk life in the shortest time, thereby testing the performance and reliability performance at the end of the disk life in a limited project test schedule, so as to discover the defects hidden at the end of the disk life and provide solid technical support and guarantee for the improvement of product quality.

[0099] The following further explains and introduces the execution logic of the solid-state drive accelerated life test method of the present application by combining the accompanying drawings.

[0100] Figure 2 It is a schematic diagram of the execution logic of the solid-state drive accelerated life test method of the present application. As Figure 2 shown, the execution steps of the solid-state drive accelerated life test method of the present application are as follows:

[0101] S201: Insert the solid-state drive to be tested into the aging test device;

[0102] S202: Real-time monitor the number of erase / write cycles and the incubator temperature of the solid-state drive to be tested;

[0103] S203: Arrange and combine multiple parameters of the test tool and the incubator temperature to generate multiple state values of the Q-learning model;

[0104] S204: Input the number of erase / write cycles and multiple state values into the Q-learning model, and perform multiple iterations through the Q-learning model to traverse the multiple state values;

[0105] S205: Determine the target read / write instructions and the target incubator temperature of the solid-state drive to be tested;

[0106] [[ID=D27]]S206: Set the incubator temperature according to the target incubator temperature, and issue the target read / write instructions to the solid-state drive to be tested to perform an accelerated life test;

[0107] S207: After the solid-state drive to be tested reaches the end of its life, perform reliability and performance tests on the solid-state drive to be tested.

[0108] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0109] The embodiment of the present application also provides a solid-state drive accelerated life test device.

[0110] As Figure 3As shown in the figure, the solid-state drive accelerated life test device 10 includes: a permutation and combination module 100, a traversal module 200, and a life test module 300.

[0111] Among them, the permutation and combination module 100 is used to send multiple test tool parameters to the solid-state drive to be tested, obtain the temperature of the incubator of the preset aging test device, and perform permutation and combination on the multiple test tool parameters and the incubator temperature to obtain multiple state values.

[0112] The traversal module 200 is used to traverse multiple state values using a pre-constructed Q-learning model to obtain a target state value that meets the requirements of the preset number of disk erasure and write cycles, and split the target state value to determine the target read and write instructions and the target incubator temperature of the solid-state drive to be tested.

[0113] The life test module 300 is used to perform an accelerated life test operation on the solid-state drive to be tested based on the target read and write instructions and the target incubator temperature, monitor the number of disk erasure and write cycles during the accelerated life test operation, and determine whether the number of disk erasure and write cycles reaches the preset erasure and write cycle threshold. Among them, when the number of disk erasure and write cycles reaches the preset erasure and write cycle threshold, perform a reliability test and / or a performance test on the solid-state drive to be tested to obtain corresponding test data, and send the test data to the target client.

[0114] Optionally, in an embodiment of the present application, the solid-state drive accelerated life test device of the embodiment of the present application further includes: a deployment module, which is used to insert the solid-state drive to be tested into the aging test device and power on the aging test device before sending multiple test tool parameters to the solid-state drive to be tested, so as to deploy the test environment of the aging test device.

[0115] Optionally, in an embodiment of the present application, the permutation and combination module 100 includes: a sending unit, an obtaining unit, and a combining unit.

[0116] Among them, the sending unit is used to configure a preset test software in the aging test device to send a test command corresponding to the test software to the solid-state drive to be tested through the test software.

[0117] The obtaining unit is used to send multiple test tool parameters to the solid-state drive to be tested based on the test command, and obtain the temperature of the incubator of the aging test device through the solid-state drive to be tested.

[0118] The combining unit is used to perform permutation and combination on multiple test tool parameters and the incubator temperature by running a preset script to obtain multiple state values of the Q-learning model.

[0119] Optionally, in an embodiment of the present application, the traversal module 200 includes: a monitoring unit, an adjustment unit, an iteration unit, and a splitting unit.

[0120] Among them, the monitoring unit is used to monitor the number of times the disk of the solid-state drive to be tested is erased and written and the temperature of the incubator in real time, initialize the Q function in the Q-learning model to obtain the corresponding initialized Q function value, and determine the current action corresponding to the current state value according to the preset ε-greedy strategy.

[0121] The adjustment unit is used to execute the current action and obtain the reward corresponding to the current action, so as to obtain the growth rate of the number of times the disk is erased and written corresponding to the current state value according to the current action and the reward, and adjust each parameter of the current state value through the growth rate of the number of times the disk is erased and written.

[0122] The iteration unit is used to determine the next state value through the adjusted current state value, and update the initialized Q function value through the next state value and the reward to obtain a new Q function value, and perform an update operation on the new Q function value iteratively to determine the target state value corresponding to the maximum growth rate of the number of times the disk is erased and written during multiple iterations.

[0123] The splitting unit is used to split the target state value to obtain the corresponding target read / write instruction and target incubator temperature.

[0124] Optionally, in an embodiment of the present application, the life test 300 includes: a determination unit, a judgment unit, a first analysis unit, and a second analysis unit.

[0125] Among them, the determination unit is used to send the target read / write instruction to the solid-state drive to be tested, and determine the environmental temperature of the aging test device according to the target incubator temperature, so as to perform an accelerated life test on the solid-state drive to be tested through the target read / write instruction and the environmental temperature;

[0126] The judgment unit is used to monitor the number of times the disk is erased and written during the accelerated life test, and judge whether the number of times the disk is erased and written is less than a preset erasure count threshold;

[0127] The first analysis unit is used to continue to perform an accelerated life test on the solid-state drive to be tested if the number of times the disk is erased and written is less than the preset erasure count threshold;

[0128] The second analysis unit is used to stop the accelerated life test on the solid-state drive to be tested if the number of times the disk is erased and written is greater than or equal to the preset erasure count threshold, and determine the actual disk test requirements, so as to perform a reliability test and / or a performance test on the solid-state drive to be tested according to the actual disk test requirements to obtain corresponding test data, and send the test data to the target client.

[0129] For the description of the features in the embodiments corresponding to the solid-state drive accelerated life test device, reference can be made to the relevant description of the embodiments corresponding to the solid-state drive accelerated life test method, which will not be elaborated here one by one.

[0130] Embodiments of the present application further provide an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any of the embodiments of the above-mentioned solid-state drive accelerated life test method.

[0131] Embodiments of the present application further provide a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps in any of the embodiments of the above-mentioned solid-state drive accelerated life test method when running.

[0132] In an exemplary embodiment, the above-mentioned computer-readable storage medium may include, but is not limited to: various media that can store computer programs such as USB flash drives, read-only memories (ROM for short), random access memories (RAM for short), mobile hard disks, magnetic disks, or optical discs.

[0133] Embodiments of the present application further provide a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, it implements the steps in any of the embodiments of the above-mentioned solid-state drive accelerated life test method.

[0134] Embodiments of the present application further provide another computer program product, including a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps in any of the embodiments of the above-mentioned solid-state drive accelerated life test method.

[0135] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0136] The above has introduced in detail a method, device, equipment and medium for accelerating the life test of a solid-state drive provided by this application. Specific examples are used in this article to elaborate on the principle and implementation manner of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of this application, several improvements and modifications can still be made to this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A method for accelerating the life test of a solid-state drive, characterized in that, Including the following steps: Issuing multiple test tool parameters to the solid state drive to be tested, obtaining the temperature of the incubator of the preset aging test device, and performing permutations and combinations on the multiple test tool parameters and the incubator temperature to obtain multiple state values; Using a pre-constructed Q-learning model to traverse the multiple state values to obtain a target state value that meets the requirements of the preset number of disk erasure and write cycles, and splitting the target state value to determine the target read / write instruction and the target incubator temperature of the solid state drive to be tested; Based on the target read / write instruction and the target incubator temperature, performing an accelerated life test operation on the solid state drive to be tested, monitoring the number of disk erasure and write cycles during the accelerated life test operation, and determining whether the number of disk erasure and write cycles reaches a preset erasure and write cycle threshold. Among them, when the number of disk erasure and write cycles reaches the preset erasure and write cycle threshold, performing a reliability test and / or a performance test on the solid state drive to be tested to obtain corresponding test data, and sending the test data to the target client.

2. The method according to claim 1, characterized in that, Before issuing the multiple test tool parameters to the solid state drive to be tested through the preset aging test device, it further includes: Inserting the solid state drive to be tested into the aging test device and powering on the aging test device to deploy the test environment of the aging test device.

3. The method according to claim 1, wherein The issuing of the multiple test tool parameters to the solid state drive to be tested, obtaining the temperature of the incubator of the preset aging test device, and performing permutations and combinations on the multiple test tool parameters and the incubator temperature to obtain multiple state values includes: Configuring a preset test software in the aging test device to issue a test command corresponding to the test software to the solid state drive to be tested through the test software; Based on the test command, issuing the multiple test tool parameters to the solid state drive to be tested, and obtaining the temperature of the incubator of the aging test device through the solid state drive to be tested; Performing permutations and combinations on the multiple test tool parameters and the incubator temperature by running a preset script to obtain multiple state values of the Q-learning model.

4. The method according to claim 3, characterized in that, The using of the pre-constructed Q-learning model to traverse the multiple state values to obtain a target state value that meets the requirements of the preset number of disk erasure and write cycles, and splitting the target state value to determine the target read / write instruction and the target incubator temperature of the solid state drive to be tested includes: Initializing the Q function in the Q-learning model to obtain a corresponding initialized Q function value, and determining the current action corresponding to the current state value according to a preset ε-greedy strategy; Executing the current action and obtaining the reward corresponding to the current action, to obtain the growth rate of the number of disk erasure and write cycles corresponding to the current state value according to the current action and the reward, and adjusting each parameter of the current state value through the growth rate of the number of disk erasure and write cycles; Determining the next state value through the adjusted current state value, and updating the initialized Q function value through the next state value and the reward to obtain a new Q function value, and iteratively executing the update operation on the new Q function value to determine the target state value corresponding to the maximum growth rate of the number of disk erasure and write cycles during multiple iterations; Split the target status value to obtain the corresponding target read / write instruction and the target incubator temperature.

5. The method according to claim 4, wherein Based on the target read / write instruction and the target incubator temperature, perform an accelerated life test operation on the solid-state drive to be tested, monitor the number of disk writes and erases during the accelerated life test operation, and determine whether the number of disk writes and erases reaches a preset number-of-writes threshold. Among them, when the number of disk writes and erases reaches the preset number-of-writes threshold, perform a reliability test and / or a performance test on the solid-state drive to be tested to obtain corresponding test data, and send the test data to the target client, including: Send the target read / write instruction to the solid-state drive to be tested, and determine the environmental temperature of the aging test device according to the target incubator temperature, so as to perform an accelerated life test on the solid-state drive to be tested through the target read / write instruction and the environmental temperature; Monitor the number of disk writes and erases during the accelerated life test, and determine whether the number of disk writes and erases is less than the preset number-of-writes threshold; If the number of disk writes and erases is less than the preset number-of-writes threshold, continue to perform an accelerated life test on the solid-state drive to be tested; If the number of disk writes and erases is greater than or equal to the preset number-of-writes threshold, stop performing the accelerated life test on the solid-state drive to be tested, and determine the actual disk test requirements, so as to perform a reliability test and / or a performance test on the solid-state drive to be tested according to the actual disk test requirements to obtain corresponding test data, and send the test data to the target client.

6. A solid-state drive accelerated life test device, characterized in that, Including: A permutation and combination module for sending multiple test tool parameters to the solid-state drive to be tested, obtaining the incubator temperature of the preset aging test device, and performing permutation and combination on the multiple test tool parameters and the incubator temperature to obtain multiple status values; A traversal module for using a pre-constructed Q-learning model to traverse the multiple status values to obtain a target status value that meets the preset number-of-disk-writes requirement, and splitting the target status value to determine the target read / write instruction and the target incubator temperature of the solid-state drive to be tested; A life test module for performing an accelerated life test operation on the solid-state drive to be tested based on the target read / write instruction and the target incubator temperature, monitoring the number of disk writes and erases during the accelerated life test operation, and determining whether the number of disk writes and erases reaches a preset number-of-writes threshold. Among them, when the number of disk writes and erases reaches the preset number-of-writes threshold, perform a reliability test and / or a performance test on the solid-state drive to be tested to obtain corresponding test data, and send the test data to the target client.

7. The device according to claim 6, characterized in that, The traversal module includes: A monitoring unit for real-time monitoring of the number of disk writes and erases and the incubator temperature of the solid-state drive to be tested, initializing the Q function in the Q-learning model to obtain a corresponding initialized Q function value, and determining the current action corresponding to the current status value according to the preset ε-greedy strategy; An adjustment unit, configured to execute the current action, obtain a reward corresponding to the current action, obtain a disk rewrite times growth rate corresponding to the current state value according to the current action and the reward, and adjust each parameter of the current state value through the disk rewrite times growth rate; An iteration unit, configured to determine a next state value through the adjusted current state value, update the initialized Q function value through the next state value and the reward to obtain a new Q function value, and iteratively execute an update operation on the new Q function value to determine a target state value corresponding to the maximum disk rewrite times growth rate in multiple iteration processes; A splitting unit, configured to split the target state value to obtain the corresponding target read / write instruction and the target incubator temperature.

8. An electronic device, characterized in that, Comprising: A memory, configured to store a computer program; A processor, configured to implement the steps of the solid-state drive accelerated life test method according to any one of claims 1 to 5 when executing the computer program.

9. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, wherein the computer program implements the steps of the solid-state drive accelerated life test method according to any one of claims 1 to 5 when executed by a processor.

10. A computer program product, comprising a computer program, characterized in that, The computer program implements the steps of the solid-state drive accelerated life test method according to any one of claims 1 to 5 when executed by a processor.

Images