A storage pool management method, system, device and storage medium

Abstract

The application discloses a kind of management methods of storage pool, comprising the following steps: predicting the state of each hard disk in the storage pool, and the hard disk with the state obtained by prediction as risk state is as risk disk;Obtain the redundancy protection strategy of storage pool;Based on the redundancy protection strategy of the storage pool, the number or position of the risk disk is counted and the risk level is determined according to the number or position of the risk disk;Based on the risk level, the strategy for processing the risk disk is determined.The application also discloses a system, a computer device and a readable storage medium.The scheme proposed by the application not only depends on the prediction result of hard disk level, but also combines the prediction result of hard disk with redundancy protection strategy, and takes different processing logic according to different redundancy protection strategies.

Description

Technical Field

[0001] This invention relates to the field of storage, and more specifically to a method, system, device, and storage medium for managing a storage pool. Background Technology

[0002] To improve hard drive reliability, the common practice is to collect historical indicator data (such as SMART) of storage devices, analyze the historical indicator data to understand the failure trend of storage devices, and then use machine learning algorithms to build a hard drive failure prediction model and deploy the model to the production environment to monitor the device in real time, predict whether the hard drive will fail, and provide early warning for risky drives.

[0003] While hard drive failure prediction technology can anticipate potential hard drive failures, it still suffers from a high false alarm rate. Furthermore, hard drive failures do not directly impact data availability. Therefore, simply obtaining the health status of a single hard drive is insufficient to fully determine the safety of the storage system's data. Relying solely on the failure prediction results of a single hard drive to determine whether to replace it can easily lead to wasted maintenance resources and increased maintenance costs. Summary of the Invention

[0004] In view of this, in order to overcome at least one aspect of the above problems, embodiments of the present invention propose a storage pool management method, comprising the following steps:

[0005] Predict the state of each hard drive in the storage pool, and designate the hard drives whose predicted state is risky as risky drives.

[0006] Obtain the redundancy protection strategy for the storage pool;

[0007] Based on the redundancy protection strategy of the storage pool, the number or location of the risk disks is counted and the risk level is determined according to the number or location of the risk disks.

[0008] The strategy for handling the risk panel is determined based on the risk level.

[0009] In some embodiments, the method of counting the number or location of the risky disks based on the redundancy protection strategy of the storage pool and determining the risk level based on the number or location of the risky disks further includes:

[0010] In response to the redundancy protection strategy being an erasure coding strategy, the number of risk disks is counted, and the risk level is determined based on the number of risk disks and the number of check blocks set in the erasure coding strategy.

[0011] In some embodiments, determining the risk level based on the number of risk disks and the number of verification blocks set in the erasure coding strategy further includes:

[0012] If the number of risk disks is greater than the number of verification blocks, the risk level is determined to be high risk;

[0013] In response to the fact that the number of risk disks equals the number of verification blocks, the risk level is determined to be medium risk;

[0014] In response to the fact that the number of risk disks is less than the number of verification blocks, the risk level is determined to be low risk.

[0015] In some embodiments, the method of counting the number or location of the risky disks based on the redundancy protection strategy of the storage pool and determining the risk level based on the number or location of the risky disks further includes:

[0016] In response to the redundancy protection strategy being a multi-replica strategy, the location and sum of each of the risk disks are obtained;

[0017] The risk level is determined based on the location of each risk disk and the number of replicas set in the multi-replica strategy.

[0018] In some embodiments, determining the risk level based on the number of risk disks, the location of each risk disk, and the number of replicas set in the multi-replica strategy further includes:

[0019] Obtain the fault domain type;

[0020] In response to the fault domain type being a node or a cabinet, the nodes or cabinets containing the risk disk are counted based on the location of each risk disk;

[0021] If the number of nodes or cabinets with the risk disk is greater than the number of replicas, the risk level is determined to be high risk.

[0022] In response to the number of nodes or cabinets containing the risk disk being equal to the number of replicas, the risk level is determined to be high risk;

[0023] If the number of nodes or cabinets with the risk disk is less than the number of replicas, the risk level is determined to be high risk.

[0024] In some embodiments, predicting the state of each hard drive in the storage pool further includes:

[0025] The system polls to obtain multiple SMART features for each of the hard drives, wherein the correlation coefficient between any two of the multiple SMART features is less than a threshold, and the correlation coefficient between each of the SMART features and the hard drive failure is greater than the threshold.

[0026] The state is predicted using multiple SMART features of each of the hard drives.

[0027] In some embodiments, predicting the state using a plurality of SMART features of each of the hard disks further includes:

[0028] The multiple SMART features are input into the encoder to obtain the first sample features;

[0029] The features of the first sample are input into the decoder to obtain the reconstructed sample;

[0030] The reconstructed sample is then input into the encoder again to obtain the second sample feature;

[0031] Calculate the norm between the first sample feature and the second sample feature and compare it with a preset threshold;

[0032] In response to the norm being greater than the preset threshold, the corresponding hard drive is marked as being in a risky state.

[0033] Based on the same inventive concept, according to another aspect of the present invention, embodiments of the present invention also provide a storage pool management system, comprising:

[0034] The prediction module is configured to predict the state of each hard drive in the storage pool and identify hard drives with a predicted state of risk as risk drives.

[0035] The acquisition module is configured to acquire the redundancy protection strategy of the storage pool;

[0036] The risk determination module is configured to count the number or location of the risk disks based on the redundancy protection strategy of the storage pool and determine the risk level based on the number or location of the risk disks.

[0037] The strategy module is configured to determine the strategy for handling the risk panel based on the risk level.

[0038] Based on the same inventive concept, according to another aspect of the present invention, embodiments of the present invention also provide a computer device, comprising:

[0039] At least one processor; and

[0040] The memory stores a computer program that can run on the processor, and when the processor executes the program, it performs the steps of any of the memory pool management methods described above.

[0041] Based on the same inventive concept, according to another aspect of the present invention, embodiments of the present invention also provide a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of any of the storage pool management methods described above.

[0042] This invention offers one of the following beneficial technical effects: The proposed solution does not solely rely on hard drive-level prediction results, but combines these results with redundancy protection strategies. Different processing logics are employed based on different redundancy protection strategies to comprehensively assess data security, thus avoiding the influence of the randomness of hard drive failure prediction. This approach utilizes both hard drive performance data and distribution data to assess risks from a storage pool perspective, thereby determining whether to perform a disk replacement. This not only provides timely risk warnings but also reduces the operational costs incurred due to premature disk replacement. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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 embodiments can be obtained based on these drawings without creative effort.

[0044] Figure 1 A schematic flowchart of a storage pool management method provided for an embodiment of the present invention;

[0045] Figure 2 A risk level diagram provided for embodiments of the present invention;

[0046] Figure 3 A schematic diagram of a three-replica storage pool provided for an embodiment of the present invention;

[0047] Figure 4 A schematic diagram illustrating hard disk failure prediction provided for an embodiment of the present invention;

[0048] Figure 5 A schematic diagram of an encoder provided for an embodiment of the present invention;

[0049] Figure 6 A schematic diagram of a decoder provided for an embodiment of the present invention;

[0050] Figure 7 A schematic diagram of the structure of a storage pool management system provided for an embodiment of the present invention;

[0051] Figure 8 A schematic diagram of the structure of a computer device provided for an embodiment of the present invention;

[0052] Figure 9 A schematic diagram of the structure of a computer-readable storage medium provided for an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to specific examples and the accompanying drawings.

[0054] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities or parameters with the same name but different names. It is clear that "first" and "second" are only for the convenience of expression and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0055] In embodiments of the present invention, to improve the reliability of hard drives, hard drive manufacturers provide a specification for detecting the operating status of hard drives, namely SMART (Self-Monitoring Analysis and Reporting Technology), an automatic hard drive status detection and early warning system and specification. By detecting indicators such as the underlying data error read rate, motor start / stop count, remapped sector count, and seek error rate during hard drive operation, a threshold detection method is used to achieve non-destructive hard drive health status detection: a warning is issued if any indicator falls below a preset threshold.

[0056] There are two main redundancy protection strategies for distributed storage: multiple replicas and erasure coding. Multiple replicas, as the name suggests, involve creating multiple copies of data, achieved through data replication and distribution. Simply put, a single set of data is copied multiple times and stored on different nodes. When a node fails or the network connection drops, the system can automatically switch to other nodes to continue providing data services, such as two-replica or three-replica systems. Multiple replicas are suitable for scenarios with small data volumes and frequent data read / write operations, such as cloud storage and distributed file systems. Erasure coding is a technique that uses data encoding, fragmentation, and reassembly to divide data into multiple data blocks, encode these blocks, and distribute their storage, thereby achieving data redundancy protection. Erasure coding can provide higher data reliability and availability than multiple replicas without increasing data storage space. Erasure coding is suitable for scenarios with large data volumes and infrequent data read / write operations, such as distributed databases and large-scale data centers. In addition, distributed storage systems also have the concept of fault domains. A fault domain typically refers to a group of servers, storage, or network connection components collectively affected by a failure. In storage systems, defining fault domains helps administrators better manage storage devices and data, improving the reliability and fault tolerance of the storage system. Distributed storage fault domains have four levels: data center fault domain, rack fault domain, node fault domain, and hard disk fault domain. For example, when the fault domain is a node and the redundancy protection strategy is three replicas, it allows two nodes to fail simultaneously, with the remaining node providing normal service. As the storage scale increases and the number of nodes increases, the probability of multiple nodes failing simultaneously rises. At this point, node-level fault domains are insufficient, and the fault domain level can be escalated to the rack level. Data safety is determined by detecting whether a fault domain has been reached. Therefore, simply obtaining the health status of a hard disk is insufficient to fully determine the safety of the data in the storage system; it must also be combined with the redundancy protection strategy of the distributed storage. If the decision to replace a disk is based solely on the failure prediction results of a single hard disk, it can easily lead to a waste of operational resources and an increase in operational costs.

[0057] According to one aspect of the present invention, embodiments of the present invention provide a method for managing a storage pool, such as... Figure 1 As shown, it may include the following steps:

[0058] S1, predict the state of each hard disk in the storage pool, and designate the hard disks whose predicted state is risky as risky disks;

[0059] S2, obtain the redundancy protection strategy of the storage pool;

[0060] S3, based on the redundancy protection strategy of the storage pool, count the number or location of the risk disks and determine the risk level according to the number or location of the risk disks;

[0061] S4, determine the strategy for handling the risk panel based on the risk level.

[0062] The proposed solution does not rely solely on hard drive-level prediction results, but combines these results with redundancy protection strategies. Different processing logics are applied based on different redundancy protection strategies to comprehensively assess data security, thus avoiding the influence of the randomness in hard drive failure prediction. This approach utilizes both hard drive performance data and distribution data to assess risks from a storage pool perspective, thereby deciding whether to perform a hard drive replacement. This not only provides timely risk warnings but also reduces the operational costs incurred due to premature hard drive replacement.

[0063] In some embodiments, the proposed solution mainly includes four processes: data acquisition, hard disk failure prediction, data security detection, and alarm location. First, data acquisition is performed, primarily collecting two types of data: hard disk SMART indicator data and distribution data. SMART data serves as input for hard disk failure prediction, mainly consisting of monitoring indicators during hard disk operation. Distribution data serves as input for data security detection, mainly including information such as the node, rack, slot number, storage pool name, storage pool policy, and fault domain type of each hard disk. Then, hard disk failure prediction is performed based on the SMART indicator data to obtain the health status prediction results for each hard disk. Next, the prediction results and distribution data for each hard disk are aggregated and input into the data security detection module to obtain the data security status at the storage pool level. Alarm location receives the results from the hard disk failure prediction module and the data security detection module, comprehensively determines the risk level, and locates the risky disks, facilitating subsequent troubleshooting by maintenance personnel.

[0064] In some embodiments, the method of counting the number or location of the risky disks based on the redundancy protection strategy of the storage pool and determining the risk level based on the number or location of the risky disks further includes:

[0065] In response to the redundancy protection strategy being an erasure coding strategy, the number of risk disks is counted, and the risk level is determined based on the number of risk disks and the number of check blocks set in the erasure coding strategy.

[0066] In some embodiments, determining the risk level based on the number of risk disks and the number of verification blocks set in the erasure coding strategy further includes:

[0067] If the number of risk disks is greater than the number of verification blocks, the risk level is determined to be high risk;

[0068] In response to the fact that the number of risk disks equals the number of verification blocks, the risk level is determined to be medium risk;

[0069] In response to the fact that the number of risk disks is less than the number of verification blocks, the risk level is determined to be low risk.

[0070] Specifically, the redundancy protection strategy of the storage pool, the number of risky disks, and the number of parity blocks (m) can be combined to comprehensively determine whether data has reached or exceeded the fault domain. If the redundancy protection strategy of the storage pool is an erasure coding strategy (e.g., k+m:1 erasure, where m is the number of parity blocks), then the health status of the hard disks in the fault domain list of the current storage pool nodes (or racks) is statistically analyzed. If the number of risky disks in different virtual nodes (or racks) is less than the number of parity blocks (m), then the data has not reached the fault domain. If the number of risky disks in the pool is equal to the number of parity blocks (m), then the status is "reached the fault domain." If the number of risky disks in the pool is greater than or equal to m+1, then the status is "exceeded the fault domain." Figure 2 As shown, if there are no risky disks, the storage pool is considered risk-free. If there are risky disks but they haven't reached the fault domain, the risk index is I (low risk); if they reach the fault domain, the risk index is II (medium risk); and if they exceed the fault domain, the risk index is III (high risk). Alarms of varying levels can then be provided based on the risk. Simultaneously, the physical location of the risky hard drives is pinpointed, providing information such as the rack, node, and slot number to facilitate subsequent disk replacement. After replacing potentially faulty disks, data security is reassessed. Based on the risk assessment rules, alarms are downgraded or restored.

[0071] In some embodiments, the method of counting the number or location of the risky disks based on the redundancy protection strategy of the storage pool and determining the risk level based on the number or location of the risky disks further includes:

[0072] In response to the redundancy protection strategy being a multi-replica strategy, the location and sum of each of the risk disks are obtained;

[0073] The risk level is determined based on the location of each risk disk and the number of replicas set in the multi-replica strategy.

[0074] In some embodiments, determining the risk level based on the number of risk disks, the location of each risk disk, and the number of replicas set in the multi-replica strategy further includes:

[0075] Obtain the fault domain type;

[0076] In response to the fault domain type being a node or a cabinet, the nodes or cabinets containing the risk disk are counted based on the location of each risk disk;

[0077] If the number of nodes or cabinets with the risk disk is greater than the number of replicas, the risk level is determined to be high risk.

[0078] In response to the number of nodes or cabinets containing the risk disk being equal to the number of replicas, the risk level is determined to be high risk;

[0079] If the number of nodes or cabinets with the risk disk is less than the number of replicas, the risk level is determined to be high risk.

[0080] Specifically, if the storage pool's redundancy protection strategy is a multi-replica strategy, the physical distribution of hard drives needs to be considered. Based on whether the pool's fault domain type is a node or a rack, and the physical distribution of each hard drive in the current pool's hard drive list, the health status of the hard drives distributed across each node (or rack) is statistically analyzed, and these hard drives are ranked according to their health status. If the number of risk states for different nodes (or racks) in the pool's hard drive physical distribution is less than m, the status is "not yet in the fault domain"; if the number of risk states for different nodes (or racks) in the pool's hard drive physical distribution is equal to the number of parity blocks m, the status is "in the fault domain"; if the number of risk states for different nodes (or racks) in the pool's hard drive physical distribution is greater than or equal to m+1, the status is "exceeded the fault domain". Figure 2 As shown, if there are no risky disks, the storage pool is considered risk-free. If there are risky disks but they haven't reached the fault domain, the risk index is I (low risk); if they reach the fault domain, the risk index is II (medium risk); and if they exceed the fault domain, the risk index is III (high risk). Alarms of varying levels can then be provided based on the risk. Simultaneously, the physical location of the risky hard drives is pinpointed, providing information such as the rack, node, and slot number to facilitate subsequent disk replacement. After replacing potentially faulty disks, data security is reassessed. Based on the risk assessment rules, alarms are downgraded or restored.

[0081] For example, Figure 3 The 3-replica storage pool shown has a node-level fault domain, tolerating a maximum of two disks failing simultaneously on different nodes. When the number of failed disks is less than two, the data is safe as it has not yet reached the fault domain. When two disks on different nodes fail simultaneously, pool 1 will reach the fault domain. At this point, the data has not yet been lost, but the risk of data loss is already high, triggering an alarm and prompting timely disk replacement. When three disks fail simultaneously, pool 1 will exceed the fault domain, indicating data loss. This situation should be avoided; disk replacement should be performed before exceeding the fault domain.

[0082] In some embodiments, predicting the state of each hard drive in the storage pool further includes:

[0083] The system polls to obtain multiple SMART features for each of the hard drives, wherein the correlation coefficient between any two of the multiple SMART features is less than a threshold, and the correlation coefficient between each of the SMART features and the hard drive failure is greater than the threshold.

[0084] The state is predicted using multiple SMART features of each of the hard drives.

[0085] Specifically, the hard drive SMART information is collected, mainly including indicators such as the underlying data read error rate, motor start-stop count, number of remapped sectors, seek error rate, cumulative power-on time, and motor start-up retry, and these values ​​are persisted to the database for subsequent fault prediction.

[0086] After data collection, preprocessing is performed, primarily feature selection and data completion. Feature selection mainly utilizes the Pearson correlation coefficient. The Pearson correlation coefficient measures the correlation between variables, ranging from -1 to 1. A larger absolute value indicates a higher correlation and stronger redundancy between the two variables. First, the Pearson correlation coefficient between each pair of SMART features is calculated. If the absolute value of the coefficient between two features is large, one of them is removed. Then, the correlation between SMART features and faults is calculated, and features with weak correlation are removed. The remaining features are used as the final model input features. After determining the features, missing features need to be filled. If the acquired SMART values ​​are null, forward imputation is used. In some embodiments, several relevant features can be selected during the model training phase, and only these selected relevant features need to be collected during data acquisition.

[0087] In some embodiments, predicting the state using a plurality of SMART features of each of the hard disks further includes:

[0088] The multiple SMART features are input into the encoder to obtain the first sample features;

[0089] The features of the first sample are input into the decoder to obtain the reconstructed sample;

[0090] The reconstructed sample is then input into the encoder again to obtain the second sample feature;

[0091] Calculate the norm between the first sample feature and the second sample feature and compare it with a preset threshold;

[0092] In response to the norm being greater than the preset threshold, the corresponding hard drive is marked as being in a risky state.

[0093] Specifically, after obtaining the preprocessed data, hard drive failure prediction is performed. The specific prediction process is as follows: Figure 4As shown in the diagram, preprocessed hard drive SMART data samples x are used as input. Encoder 1 extracts features to obtain the feature representation z of the sample. The decoder reconstructs the features z to obtain the reconstructed sample x′. Finally, encoder 2 extracts features to obtain the feature representation z′ of the reconstructed sample x′. Then, the L2 norm A(X) = z′² of z and z′ is calculated and compared with a model threshold. If it is greater than the threshold, the sample is marked as a risky hard drive; otherwise, it is marked as a healthy hard drive, thus obtaining the fault prediction results for each hard drive. The structures of the two encoders are shown in the diagram. Figure 5 As shown, the code first passes through a 4x4 convolutional layer and a LeakyReLU layer, then two convolutional block structures, each block consisting of a 4x4 convolutional layer, a batch normalization layer, and a LeakyReLU layer, and finally a 4x4 convolutional layer. The decoder structure is as follows. Figure 6 As shown, the process first goes through three transposed convolutional block structures, each consisting of a 4x4 transposed convolutional layer, a batch normalization layer, and a ReLU layer. Finally, it goes through a 4x4 transposed convolutional layer and a Tanh layer.

[0094] Although hard drive failure prediction technology can predict failed drives in advance, it still has a high false alarm rate. Furthermore, since distributed storage systems provide data services to upper-layer applications through virtual storage pools with redundancy protection strategies, the risk of a single hard drive failure does not necessarily mean that data security is threatened. Determining whether to replace a drive solely based on the failure prediction results of a single hard drive can easily lead to a waste of operational resources and increased maintenance costs. Therefore, this invention proposes a distributed storage data security detection method based on hard drive failure prediction and redundancy protection strategies, which has the following advantages:

[0095] 1. It does not rely solely on hard drive-level prediction results, but combines hard drive prediction results with redundancy protection strategies. Based on different redundancy protection strategies, different processing logics are adopted to comprehensively judge whether the data is safe, thus avoiding the impact of the randomness of hard drive failure prediction.

[0096] 2. This patent utilizes both hard drive performance data and hard drive distribution data to assess risks from the perspective of storage pools and in conjunction with fault domain judgment, thereby deciding whether to perform a disk replacement operation. This not only provides timely risk warnings but also reduces the maintenance costs increased due to premature disk replacement.

[0097] 3. This patent combines hard drive failure prediction results and data security risk detection results to classify risk levels, which is more reasonable and more in line with business scenarios than a single risk, and helps maintenance personnel to judge risks.

[0098] 4. After replacing a risky disk, the storage system will automatically recognize the hard drive and re-perform a storage pool-level data security check. It can automatically downgrade or restore the alarm, making the process more scientific.

[0099] Based on the same inventive concept, according to another aspect of the present invention, embodiments of the present invention also provide a storage pool management system 400, such as... Figure 7 As shown, it includes:

[0100] The prediction module 401 is configured to predict the state of each hard disk in the storage pool and to designate the hard disks with the predicted state as risk disks.

[0101] Module 402 is configured to obtain the redundancy protection strategy of the storage pool;

[0102] The risk determination module 403 is configured to count the number or location of the risk disks based on the redundancy protection strategy of the storage pool and determine the risk level based on the number or location of the risk disks.

[0103] Strategy module 404 is configured to determine a strategy for handling the risk panel based on the risk level.

[0104] In some embodiments, the risk determination module 403 is further configured to:

[0105] In response to the redundancy protection strategy being an erasure coding strategy, the number of risk disks is counted, and the risk level is determined based on the number of risk disks and the number of check blocks set in the erasure coding strategy.

[0106] In some embodiments, the risk determination module 403 is further configured to:

[0107] If the number of risk disks is greater than the number of verification blocks, the risk level is determined to be high risk;

[0108] In response to the fact that the number of risk disks equals the number of verification blocks, the risk level is determined to be medium risk;

[0109] In response to the fact that the number of risk disks is less than the number of verification blocks, the risk level is determined to be low risk.

[0110] In some embodiments, the risk determination module 403 is further configured to:

[0111] In response to the redundancy protection strategy being a multi-replica strategy, the location and sum of each of the risk disks are obtained;

[0112] The risk level is determined based on the location of each risk disk and the number of replicas set in the multi-replica strategy.

[0113] In some embodiments, the risk determination module 403 is further configured to:

[0114] Obtain the fault domain type;

[0115] In response to the fault domain type being a node or a cabinet, the nodes or cabinets containing the risk disk are counted based on the location of each risk disk;

[0116] If the number of nodes or cabinets with the risk disk is greater than the number of replicas, the risk level is determined to be high risk.

[0117] In response to the number of nodes or cabinets containing the risk disk being equal to the number of replicas, the risk level is determined to be high risk;

[0118] If the number of nodes or cabinets with the risk disk is less than the number of replicas, the risk level is determined to be high risk.

[0119] In some embodiments, the prediction module 401 is further configured to:

[0120] The system polls to obtain multiple SMART features for each of the hard drives, wherein the correlation coefficient between any two of the multiple SMART features is less than a threshold, and the correlation coefficient between each of the SMART features and the hard drive failure is greater than the threshold.

[0121] The state is predicted using multiple SMART features of each of the hard drives.

[0122] In some embodiments, the prediction module 401 is further configured to:

[0123] The multiple SMART features are input into the encoder to obtain the first sample features;

[0124] The features of the first sample are input into the decoder to obtain the reconstructed sample;

[0125] The reconstructed sample is then input into the encoder again to obtain the second sample feature;

[0126] Calculate the norm between the first sample feature and the second sample feature and compare it with a preset threshold;

[0127] In response to the norm being greater than the preset threshold, the corresponding hard drive is marked as being in a risky state.

[0128] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 8 As shown, embodiments of the present invention also provide a computer device 501, comprising:

[0129] At least one processor 520; and

[0130] The memory 510 stores a computer program 511 that can run on the processor. When the processor 520 executes the program, it performs the steps of any of the memory pool management methods described above.

[0131] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 9 As shown, embodiments of the present invention also provide a computer-readable storage medium 601, which stores a computer program 610. When the computer program 610 is executed by a processor, it performs the steps of any of the storage pool management methods described above.

[0132] Finally, it should be noted that 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 program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods.

[0133] Furthermore, it should be understood that the computer-readable storage medium (e.g., memory) described herein may be volatile memory or non-volatile memory, or may include both volatile memory and non-volatile memory.

[0134] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the system as a whole. Those skilled in the art can implement the functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0135] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0136] It should be understood that, as used herein, the singular form “a” is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, “and / or” refers to any and all possible combinations of one or more of the associated listed items.

[0137] The embodiment numbers disclosed in the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0138] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0139] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A method for managing a storage pool, characterized in that, Includes the following steps: Predict the state of each hard drive in the storage pool, and designate the hard drives whose predicted state is risky as risky drives. Obtain the redundancy protection strategy for the storage pool; Based on the redundancy protection strategy of the storage pool, the number and location of the risk disks are counted, and the risk level is determined according to the number and location of the risk disks; The strategy for handling the risk panel is determined based on the aforementioned risk level; Based on the redundancy protection strategy of the storage pool, the number and location of the risky disks are counted, and the risk level is determined according to the number and location of the risky disks, further including: In response to the redundancy protection strategy being a multi-replica strategy, the location of each of the risk disks is obtained; The risk level is determined based on the location of each risk disk and the number of replicas set in the multi-replica strategy. Determining the risk level based on the number of risk disks, the location of each risk disk, and the number of replicas set in the multi-replica strategy further includes: Obtain the fault domain type; In response to the fault domain type being a node or a cabinet, the nodes or cabinets containing the risk disk are counted based on the location of each risk disk; If the number of nodes or cabinets with the risk disk is greater than the number of replicas, the risk level is determined to be high risk. In response to the number of nodes or cabinets containing the risk disk being equal to the number of replicas, the risk level is determined to be high risk; If the number of nodes or cabinets with the risk disk is less than the number of replicas, the risk level is determined to be high risk.

2. The method as described in claim 1, characterized in that, Predicting the state of each hard drive in the storage pool further includes: The system polls to obtain multiple SMART features for each of the hard drives, wherein the correlation coefficient between any two of the multiple SMART features is less than a threshold, and the correlation coefficient between each of the SMART features and the hard drive failure is greater than the threshold. The state is predicted using multiple SMART features of each of the hard drives.

3. The method as described in claim 2, characterized in that, Predicting the state using several SMART features of each of the hard disks further includes: The multiple SMART features are input into the encoder to obtain the first sample features; The features of the first sample are input into the decoder to obtain the reconstructed sample; The reconstructed sample is then input into the encoder again to obtain the second sample feature; Calculate the norm between the first sample feature and the second sample feature and compare it with a preset threshold; In response to the norm being greater than the preset threshold, the corresponding hard drive is marked as being in a risky state.

4. A storage pool management system, characterized in that, include: The prediction module is configured to predict the state of each hard drive in the storage pool and identify hard drives with a predicted state of risk as risk drives. The acquisition module is configured to acquire the redundancy protection strategy of the storage pool; The risk determination module is configured to count the number and location of the risk disks based on the redundancy protection strategy of the storage pool and determine the risk level based on the number and location of the risk disks. The strategy module is configured to determine a strategy for handling the risk panel based on the risk level. The risk assessment module is also configured as follows: In response to the redundancy protection strategy being a multi-replica strategy, the location of each of the risk disks is obtained; The risk level is determined based on the location of each risk disk and the number of replicas set in the multi-replica strategy. The risk assessment module is also configured as follows: Obtain the fault domain type; In response to the fault domain type being a node or a cabinet, the nodes or cabinets containing the risk disk are counted based on the location of each risk disk; If the number of nodes or cabinets with the risk disk is greater than the number of replicas, the risk level is determined to be high risk. In response to the number of nodes or cabinets containing the risk disk being equal to the number of replicas, the risk level is determined to be high risk; If the number of nodes or cabinets with the risk disk is less than the number of replicas, the risk level is determined to be high risk.

5. A computer device, comprising: At least one processor; as well as A memory storing a computer program executable on the processor, characterized in that the processor executes the program by performing the steps of the method as described in any one of claims 1-3.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it performs the steps of the method as described in any one of claims 1-3.

Images