A metadata management method, device, apparatus and storage medium

Abstract

The application discloses a metadata management method and device, equipment and storage medium, relates to the technical field of data management, and is applied to a file storage system and comprises the following steps: obtaining a first feature label by counting node state information of service nodes of metadata, obtaining a second feature label by counting file state information of stored files, and determining a third feature label of a current business operation and business context information; performing weighted evaluation on each service node based on a preset mapping matrix in combination with the first feature label, the second feature label and the third feature label, and determining a target service node matched with target metadata corresponding to the current business operation according to an evaluation result; and if the target service node is not a current service node corresponding to the target metadata, migrating the target metadata to the target service node to execute the current business operation. Thus, the metadata can be migrated in real time, the overhead of cross-node business operation is avoided as much as possible, and the execution effect of each business operation is ensured.

Description

Technical Field

[0001] This application relates to the field of data management technology, and in particular to a metadata management method, apparatus, device and storage medium. Background Technology

[0002] File storage systems often employ a centralized metadata service architecture, which faces single points of failure and scalability limitations. To address this issue, distributed storage systems introduce a multi-active metadata service architecture, employing technologies such as dynamic subtree partitioning and hash sharding to achieve distributed metadata management. However, these technologies still have significant shortcomings in terms of business awareness, the accuracy of dynamic routing decisions, and control over cross-node operation overhead. For example, metadata load balancing based on dynamic subtree partitioning can achieve load balancing by monitoring directory access frequency and migrating frequently accessed directory subtrees to low-load MDS (Metadata Server), but this technology suffers from processing lag, lacks business tag support, and has low load balancing accuracy.

[0003] Given the above issues, how to manage metadata to achieve effective load balancing is a problem that needs to be solved in this field. Summary of the Invention

[0004] The purpose of this application is to provide a metadata management method, apparatus, device, and storage medium that can solve the problems of low load balancing accuracy and processing lag.

[0005] To address the aforementioned technical problems, in a first aspect, embodiments of this application provide a metadata management method applied to a file storage system, comprising: The first feature label is obtained by statistically analyzing the node status information of the metadata service node, and the second feature label is obtained by statistically analyzing the file status information of the stored files. The third feature label corresponding to the current business operation and the corresponding business context information is then determined. Based on a preset mapping matrix, each service node is weighted and evaluated by combining the first feature label, the second feature label, and the third feature label, and the target service node matching the target metadata corresponding to the current business operation is determined according to the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights; If the target service node is not the current service node corresponding to the target metadata, then the target metadata is migrated to the target service node to execute the current business operation.

[0006] In some embodiments, the step of statistically analyzing the node status information of the metadata service node to obtain the first feature tag includes: The node status information of the service node is statistically analyzed to obtain the load level corresponding to the load status, the service health corresponding to the service process status, and the network quality corresponding to the network status of the service node. The load level, the service health, and the network quality are determined as the first feature label corresponding to the service node.

[0007] In some embodiments, the step of statistically analyzing the file status information of stored files to obtain the second feature label includes: The system collects the file status information of stored files, obtains the access popularity corresponding to file access operations, the file association relationship with the files involved in the file access operations, and the file storage strategy and file permission control information. The access popularity, the file association, the file storage strategy, and the file permission control information are determined as the second feature tag corresponding to the stored file.

[0008] In some embodiments, determining the third feature tag corresponding to the current business operation and the corresponding business context information includes: Based on the context information of the current business operation, a user behavior profile of relevant users within the historical period is obtained, and the context information is analyzed in a time series to obtain the time series characteristics of the business operation. The operation type of the current business operation, the user behavior profile, and the timing features of the business operation are determined as the third feature label corresponding to the current business operation.

[0009] In some embodiments, the weighted evaluation of each service node based on a preset mapping matrix, combined with the first feature label, the second feature label, and the third feature label, includes: The first feature label, the second feature label, and the third feature label are divided to obtain static feature labels in the static dimension, dynamic feature labels in the dynamic dimension, and historical feature labels in the historical dimension. Based on a preset mapping matrix, each service node is weighted and evaluated by combining the static feature labels, the dynamic feature labels, and the historical feature labels.

[0010] In some embodiments, determining the target service node that matches the target metadata corresponding to the current business operation based on the evaluation result includes: Determine the response time of each service node in response to the current business operation, as represented by the evaluation results; The service node corresponding to the shortest response time is determined as the target service node that matches the target metadata corresponding to the current business operation.

[0011] In some embodiments, migrating the target metadata to the target service node includes: If the target metadata is in a occupied state, then wait until the target metadata is in an unoccupied state and then lock the target metadata; The locked target metadata is migrated to the target service node; Accordingly, after migrating the target metadata to the target service node, the process further includes: If the verification result corresponding to the target metadata indicates that the migration was successful, then the relevant data in the current service node is deleted, and the lock on the target metadata is released. If the verification result corresponding to the target metadata indicates that the migration was unsuccessful, then the relevant data in the target service node is deleted, and the lock on the target metadata is released.

[0012] Secondly, this application provides a metadata management device for use in a file storage system, comprising: The feature tag generation module is used to statistically analyze the node status information of the metadata service node to obtain the first feature tag, and statistically analyze the file status information of the stored file to obtain the second feature tag, and determine the third feature tag corresponding to the current business operation and the corresponding business context information. The service node evaluation module is used to perform weighted evaluation of each service node based on a preset mapping matrix, combined with the first feature label, the second feature label, and the third feature label, and to determine the target service node that matches the target metadata corresponding to the current business operation based on the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights; The metadata migration module is used to migrate the target metadata to the target service node to perform the current business operation when the target service node is not the current service node corresponding to the target metadata.

[0013] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the metadata management method described above.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the metadata management method described above.

[0015] As can be seen from the above technical solution, the technical solution of this application is applied to a file storage system. First, the node status information of the metadata service node is statistically analyzed to obtain a first feature label, and the file status information of the stored files is statistically analyzed to obtain a second feature label. Then, a third feature label corresponding to the current business operation and the corresponding business context information is determined. Then, based on a preset mapping matrix, the first feature label, the second feature label, and the third feature label are combined to perform a weighted evaluation of each service node, and the target service node matching the target metadata corresponding to the current business operation is determined according to the evaluation result. The preset mapping matrix is ​​a matrix representing the mapping relationship between feature labels and weights. If the target service node is not the current service node corresponding to the target metadata, the target metadata is migrated to the target service node to execute the current business operation. The advantages of this technical solution are as follows: By statistically analyzing the information of service nodes and stored files, relevant feature tags can be generated. These feature tags are then combined with the feature tags of business operations to perform weighted evaluations of each service node, ensuring the accuracy of the evaluation. Furthermore, by migrating the relevant target metadata to the appropriate target service node before executing the business operation, routing accuracy can be improved. Real-time dynamic migration of metadata in the file storage system ensures the effectiveness of each business operation and minimizes the overhead of cross-node business operations. Attached Figure Description

[0016] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of a metadata management method provided in this application embodiment; Figure 2 A flowchart illustrating a specific metadata management method provided in this application embodiment; Figure 3 This application provides a specific code diagram illustrating the determination of a target service node. Figure 4 A schematic diagram of a metadata management device provided in an embodiment of this application; Figure 5 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0019] The terms "comprising" and "having," and any variations thereof, in the specification and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may include steps or units not listed.

[0020] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] like Figure 1 As shown, this embodiment discloses a metadata management method applied to a file storage system, including: Step S11: Statistically analyze the node status information of the metadata service node to obtain the first feature label, and statistically analyze the file status information of the stored files to obtain the second feature label, and determine the third feature label corresponding to the current business operation and the corresponding business context information.

[0022] To achieve accurate metadata load balancing, this embodiment first statistically analyzes the node status information of the metadata service nodes and the file status information of the files stored in the file storage system. This yields a first feature tag corresponding to the node status information and a second feature tag corresponding to the file status information. Correspondingly, for file-related business operations, a third feature tag corresponding to the business operation can be generated by combining the current business operation with related business operation context information. It is understood that the node status information of the metadata service nodes may include, but is not limited to, node hardware information (processor, memory, network, storage media, etc.), operating load, and metadata service health. File status information may include, but is not limited to, information recorded when a file or its corresponding directory is created, modified, accessed, or deleted, specifically such as file type, path, size, creation time, modification time, and storage / access permissions. Business operations and related context information may include, but are not limited to, operation type, read / write size, user information, client information, access time, and network latency.

[0023] In one specific embodiment, the process of statistically analyzing the node status information of metadata service nodes to obtain the first feature label involves first statistically analyzing the node status information of metadata service nodes to obtain the load level corresponding to the load status, the service health corresponding to the service process status, and the network quality corresponding to the network status of the service node; then, the load level, service health, and network quality are determined as the first feature label corresponding to the service node. It is understood that after statistical analysis, the node status information of metadata service nodes can generate feature labels such as load level related to load, service health corresponding to the service process status, and network quality corresponding to the network status, which are denoted as the first feature label. Further, the load level can consider factors such as processor utilization, memory utilization, and input / output latency; the service health can consider factors such as the liveness status of the service process, RPC (Remote Procedure Call) response latency, and disk array status; and the network quality can consider factors such as ping latency, packet loss rate, and bandwidth utilization between service nodes.

[0024] In another specific embodiment, during the process of statistically analyzing the file status information of stored files to obtain the second feature label, the file status information of the stored files is first statistically analyzed to obtain the access popularity corresponding to the file access operation, the file association relationship corresponding to the files involved in the file access operation, and the file storage policy and file permission control information. Then, the access popularity, file association relationship, file storage policy, and file permission control information are determined as the second feature label corresponding to the stored file. It is understood that files stored in the file storage system correspond to directories, and file-related feature labels also include directory-related feature labels; that is, the statistically obtained file status information of the stored files also includes information corresponding to the relevant directories; the feature labels of files / directories characterize the attributes of the data itself and the characteristics of the operation. Specifically, based on the statistically obtained file status information of the stored files, access popularity corresponding to the file access operation, file association relationship corresponding to the files involved in the file access operation, and file storage policy and file permission control information can be generated. These feature labels are denoted as the second feature label. Furthermore, access popularity tags can consider factors such as file / directory creation time, modification time, access time, and operation frequency per unit time; file association tags can consider factors such as whether files in a directory are frequently renamed in batches, read continuously, or deleted synchronously; file storage strategies and file permission control information can consider factors such as storage quotas, redundancy strategies (replication or erasure coding), user access control lists, and storage pools in the file storage system. Understandably, by identifying highly coupled file sets and assigning them the same "association" feature tag, it can be ensured that related metadata is stored on the same node as much as possible during the subsequent load balancing phase, reducing cross-node lock contention.

[0025] In another specific embodiment, in determining the third feature label corresponding to the current business operation and its corresponding business context information, firstly, user behavior profiles of relevant users within a historical period are statistically obtained based on the context information of the current business operation, and time-series analysis is performed on the context information to obtain the time-series features of the business operation; then, the operation type, user behavior profile, and time-series features of the current business operation are determined as the third feature label corresponding to the current business operation. It is understandable that after the information related to the business operation is statistically analyzed, feature labels for the operation type, user behavior profile, and time-series features of the business operation can be obtained, denoted as the third feature label. Further, the operation type label can specifically consider whether the current business operation is metadata-intensive (the key to system functionality lies in metadata) or data-intensive (such as file read / write); the user behavior profile can consider business operation habits within a historical period (e.g., a user's habit of performing large-scale data backups early Monday morning); the time-series features of the business operation specifically refer to considering the periodic patterns of metadata access over time, such as identifying the peak and trough characteristics of business operations, and generating predictive time feature labels.

[0026] Step S12: Based on the preset mapping matrix, perform weighted evaluation on each service node by combining the first feature label, the second feature label and the third feature label, and determine the target service node that matches the target metadata corresponding to the current business operation according to the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights.

[0027] In this embodiment, the above steps yield a first feature label corresponding to node status information, a second feature label corresponding to file status information, and a third feature label corresponding to the current business operation and its corresponding business context information. Then, a weighted evaluation of each service node is performed by combining pre-set mapping weights representing the mapping relationship between feature labels and weights. It is understood that this weighted evaluation can assess whether a service node is suitable for handling the current business operation; this evaluation process involves the matching relationship between the service node and the target metadata corresponding to the current business operation, such as the matching degree between feature labels, service health, load level, and other factors. Then, based on the evaluation results, the target service node matching the target metadata corresponding to the current business operation is determined. It is understood that the target service node corresponding to the same metadata may not be the same at different time periods, meaning that metadata needs to be managed periodically to optimize the execution effect of business operations.

[0028] In one specific embodiment, the process of weighted evaluation of each service node based on a preset mapping matrix and combining a first feature label, a second feature label, and a third feature label involves first dividing the first, second, and third feature labels into static feature labels (static dimension), dynamic feature labels (dynamic dimension), and historical feature labels (historical dimension). Then, based on the preset mapping matrix, each service node is weighted and evaluated using these static, dynamic, and historical feature labels. It is understandable that feature labels can be dimensionally divided according to the changes in the factors considered for feature labels, combining static, dynamic, and historical dimensions; for example, load level corresponds to static and historical dimensions, access popularity corresponds to dynamic dimensions, and user behavior profiles correspond to historical dimensions. This allows for the setting of different calculation weights for feature labels by combining static, dynamic, and historical dimensions and the actual effects of each influencing factor. Furthermore, by maintaining and dividing the first feature label, second feature label, and third feature label by combining static, dynamic, and historical dimensions, we can obtain the corresponding static dimension label, dynamic dimension label, and historical dimension label (some feature labels correspond to two dimensions, such as historical + static). Then, by combining the pre-set mapping matrix, we can perform weighted evaluation on each service node, quantify the matching relationship between the service node and the target metadata of the current business operation, and obtain the evaluation result corresponding to each service node.

[0029] In another specific embodiment, the process of determining the target service node matching the target metadata corresponding to the current business operation based on the evaluation results first involves determining the response time of each service node in response to the current business operation, as represented by the evaluation results. Then, the service node with the shortest response time is determined as the target service node matching the target metadata corresponding to the current business operation. It is understood that the evaluation results can also reflect the response time of service nodes in response to the current business operation. Combining the response times, the service node with the fastest response can be determined, i.e., the target service node matching the target metadata corresponding to the current business operation. It should be noted that the evaluation results can reflect the relationship between the service node and the current business operation at multiple levels, such as response time and system overhead. Then, a suitable service node can be selected as the target service node from the evaluation results based on the actual situation.

[0030] Step S13: If the target service node is not the current service node corresponding to the target metadata, then migrate the target metadata to the target service node to execute the current business operation.

[0031] In this embodiment, the target service node matching the target metadata corresponding to the current business operation can be determined through the above steps. It is understood that, in a continuous time period, the target service node corresponding to a single metadata may be the same (when the system state remains basically unchanged) or different (when the system state changes significantly). When the target service node and the service node currently corresponding to the metadata are the same node, no operation is required, and the current business operation can be processed directly through the target service node. When the target service node and the current service node corresponding to the metadata are different, in order to improve the execution effect of the current business operation, it is necessary to migrate the current target metadata to the target service node so that the current business operation can be executed through the target service node, thereby achieving metadata load balancing and improving the processing effect of the business operation.

[0032] In one specific embodiment, during the migration of target metadata to the target service node, it is necessary to consider whether the metadata is currently occupied, such as being locked by other processes. If the target metadata is occupied, it is locked only after the target metadata becomes unoccupied. Then, the locked target metadata is migrated to the target service node. Furthermore, after migrating the metadata to the target service node, the migrated metadata can be verified. If the verification result indicates successful migration, the relevant data in the current service node is deleted, and the lock on the target metadata is released. If the verification result indicates unsuccessful migration, the relevant data in the target service node is deleted, and the lock on the target metadata is released. This locking and verification method ensures the reliability of the metadata migration.

[0033] Therefore, this application can generate relevant feature tags by statistically analyzing the information of service nodes and stored files. By combining the feature tags of business operations, each service node is weighted and evaluated. The business operation is then executed after the relevant target metadata is migrated to the appropriate target service node. In this way, metadata routing driven by feature tags can improve routing accuracy, dynamically migrate metadata in the file storage system in real time, ensure the response speed of each business operation, reduce system resource load, and reduce performance problems caused by insufficient resources and asynchronous locks in cross-node processing.

[0034] like Figure 2 As shown, this embodiment discloses a specific metadata management method, involving feature tag generation, dynamic routing decision-making, metadata load balancing, and dynamic adjustment of tag weights in the feedback optimization process; specifically including: In this embodiment, feature tags corresponding to service nodes, files / directories, and business contexts are first generated. For service node feature tags, the node status information of the service node is scanned according to a configured timeout period (e.g., default 10 seconds), including but not limited to node hardware information, operating load, and metadata service health. For file / directory feature tags, when directories and files stored in the file storage system are created, modified, accessed, or deleted, basic file information is recorded, including but not limited to file type, path, size, creation time, modification time, and permissions. For business operation context feature tags, business operations are recorded, including but not limited to operation type, read / write size, user information, client information, access time, and network latency. Then, combining the above information, node-related feature tags (e.g., load level, service health, network quality), file / directory feature tags (e.g., access popularity, associated operations, storage strategy, permission aggregation), and business context feature tags (e.g., operation type, user behavior profile, time-series pattern, and adapted clients) can be generated.

[0035] Then, during the dynamic routing decision-making process, by combining the generated feature labels and integrating features from three dimensions—static, dynamic, and historical—as well as the actual effects of the influencing factors corresponding to the feature labels, a mapping matrix between labels and weights can be constructed. The weight factors can be dynamically adjusted based on the feedback optimization process. A specific mapping matrix is ​​shown in Table 1 below.

[0036] Table 1. A specific mapping matrix

[0037] Furthermore, the formula for evaluating service nodes is as follows: ; Where ω refers to the normalized weight coefficient of each feature label (the influence factor in the mapping matrix), f d ( ) refers to the activation function of the feature labels in the dynamic dimension (such as Sigmoid normalization), η refers to the dynamic weight ratio (e.g., an initial value of 0.6, which is automatically adjusted according to load fluctuations), LSTM refers to Long Short-Term Memory network, and T s T represents the static dimension of the feature labels. d H represents the feature labels for dynamic dimensions. h The historical dimension features are defined by the feature labels; the number of feature labels used for calculation, i.e., the number of s, d, and h, can be freely combined according to the actual situation; the final result is W. i This characterizes the matching status between service node i and the target metadata corresponding to the current business operation. It should be noted that abnormal service nodes, such as those that are unresponsive for extended periods, can be directly ignored during the service node evaluation process. In a specific implementation, such as... Figure 3 The code shown is for determining the target service node. The weights are optimized by the matching degree of service health and feature tags, and the appropriate service node is finally selected dynamically.

[0038] The above steps yield evaluation results for each service node and identify the target service node suitable for the current business operation's target metadata. If the identified target service node is inconsistent with the current service node of the target metadata, metadata migration is required, i.e., entering the metadata load balancing phase. Specifically, metadata migration can be executed according to the configured time period (hourly), triggered immediately when the current service node is determined not to be the best service node for the target metadata, or triggered manually. Furthermore, relevant metadata can be compressed during the migration process to reduce bandwidth consumption. In a specific embodiment, metadata migration is achieved through atomic operations. First, it is determined whether the target service node represented by the evaluation result is the current service node corresponding to the target metadata. If the service nodes are inconsistent, the migration object, i.e., the target metadata, is selected. Before migration, it is necessary to determine whether the target metadata is occupied or locked. The target metadata is migrated to the target service node only when it is no longer occupied or locked. Further, after migration, the accuracy and completeness of the data can be verified. If the verification is successful, the relevant data in the current service node is deleted; if the verification fails, the relevant data in the target service node is deleted. Finally, the locking status of the target metadata is released. It should be noted that migration failure may be due to metadata being in use and unable to be migrated; it may also be due to insufficient memory to copy; or it may be due to hardware reasons such as network jitter or abnormal transmission. When migration fails, retrying is the simplest solution. After several retries, an alarm can be triggered or the process can be carried over to the next time period to continue retrying.

[0039] Furthermore, the feedback optimization phase, following metadata load balancing, involves statistically analyzing the proportion of cross-metadata service processing operations across each metadata service node. The closer this proportion is to 0, the better the metadata load balancing effect. The weights corresponding to feature labels can be dynamically adjusted based on the processing status of these operations. For example, through an LSTM online learning mechanism, relevant weights are automatically adjusted based on the real-time load balancing effect sequence. Specifically, the proportion of cross-metadata service processing operations refers to the proportion where the "current metadata service node" is not equal to the "best metadata service node." Further, online learning primarily involves analyzing the entire process, collecting information from the entire process, including collected information, generated labels, decision weights, migration results, and the final load balancing effect. This information is then fed to the LSTM to predict and adjust weights, thus dynamically updating feature label weights and optimizing subsequent evaluation and load balancing decisions.

[0040] Therefore, this embodiment generates feature tags for each directory and file by statistically analyzing basic file information, metadata service node information, and business operation context, and then using feature engineering. The generated feature tags are then categorized into static, dynamic, and historical dimensions. Weighting factors are assigned, and a weight calculation model is used to evaluate each service node to determine the optimal node. Furthermore, load balancing is initiated periodically or manually, and metadata is migrated based on the dynamic routing decision results, thereby improving the performance of the metadata service. After metadata load balancing, the balancing effect is analyzed, and an LSTM model can be used to automatically adjust the feature tag weights, achieving closed-loop optimization of the load balancing effect. This enables adaptive load balancing of the metadata service, reducing the probability of business operations being processed across service nodes and improving the overall performance of the metadata service.

[0041] like Figure 4 As shown, this embodiment discloses a metadata management device applied to a file storage system, comprising: The feature tag generation module 11 is used to statistically analyze the node status information of the metadata service node to obtain the first feature tag, and statistically analyze the file status information of the stored file to obtain the second feature tag, and determine the third feature tag corresponding to the current business operation and the corresponding business context information. The service node evaluation module 12 is used to perform weighted evaluation of each service node based on a preset mapping matrix, combined with the first feature label, the second feature label, and the third feature label, and to determine the target service node that matches the target metadata corresponding to the current business operation based on the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights; Metadata migration module 13 is used to migrate the target metadata to the target service node to perform the current business operation when the target service node is not the current service node corresponding to the target metadata.

[0042] Therefore, this application can generate relevant feature tags by statistically analyzing the information of service nodes and stored files, and then perform weighted evaluation of each service node by combining the feature tags of business operations. After migrating the relevant target metadata to the appropriate target service node, the business operation is executed. This can improve routing accuracy, dynamically migrate metadata in the file storage system in real time, ensure the execution effect of each business operation, and minimize the overhead of cross-node business operations.

[0043] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 This is a structural diagram of an electronic device according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. Specifically, the electronic device may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the metadata management method disclosed in any of the foregoing embodiments. Furthermore, the electronic device in this embodiment may specifically be an electronic computer.

[0044] In this embodiment, the power supply 23 is used to provide operating voltage for various hardware devices on the electronic device; the communication interface 24 can create a data transmission channel between the electronic device and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0045] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0046] The operating system 221 is used to manage and control the various hardware devices on the electronic device and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the metadata management method executed by the electronic device as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0047] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed metadata management method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0048] Furthermore, this application also discloses a computer program product, including a computer program / instruction that, when executed by a processor, implements the aforementioned disclosed metadata management method.

[0049] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0050] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0051] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0052] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0053] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only intended to help understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A metadata management method, characterized in that, Applied to file storage systems, including: The first feature label is obtained by statistically analyzing the node status information of the metadata service node, and the second feature label is obtained by statistically analyzing the file status information of the stored files. The third feature label corresponding to the current business operation and the corresponding business context information is then determined. Based on a preset mapping matrix, each service node is weighted and evaluated by combining the first feature label, the second feature label, and the third feature label, and the target service node matching the target metadata corresponding to the current business operation is determined according to the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights; If the target service node is not the current service node corresponding to the target metadata, then the target metadata is migrated to the target service node to execute the current business operation.

2. The metadata management method according to claim 1, characterized in that, The first feature tag is obtained by statistically analyzing the node status information of the metadata service node, including: The node status information of the service node is statistically analyzed to obtain the load level corresponding to the load status, the service health corresponding to the service process status, and the network quality corresponding to the network status of the service node. The load level, the service health, and the network quality are determined as the first feature label corresponding to the service node.

3. The metadata management method according to claim 1, characterized in that, The step of statistically analyzing the file status information of stored files to obtain the second feature label includes: The system collects the file status information of stored files, obtains the access popularity corresponding to file access operations, the file association relationship with the files involved in the file access operations, and the file storage strategy and file permission control information. The access popularity, the file association, the file storage strategy, and the file permission control information are determined as the second feature tag corresponding to the stored file.

4. The metadata management method according to claim 1, characterized in that, The process of determining the third feature tag corresponding to the current business operation and the corresponding business context information includes: Based on the context information of the current business operation, a user behavior profile of relevant users within the historical period is obtained, and the context information is analyzed in a time series to obtain the time series characteristics of the business operation. The operation type of the current business operation, the user behavior profile, and the timing features of the business operation are determined as the third feature label corresponding to the current business operation.

5. The metadata management method according to claim 1, characterized in that, The weighted evaluation of each service node based on a preset mapping matrix, combined with the first feature label, the second feature label, and the third feature label, includes: The first feature label, the second feature label, and the third feature label are divided to obtain static feature labels in the static dimension, dynamic feature labels in the dynamic dimension, and historical feature labels in the historical dimension. Based on a preset mapping matrix, each service node is weighted and evaluated by combining the static feature labels, the dynamic feature labels, and the historical feature labels.

6. The metadata management method according to claim 1, characterized in that, The step of determining the target service node that matches the target metadata corresponding to the current business operation based on the evaluation results includes: Determine the response time of each service node in response to the current business operation, as represented by the evaluation results; The service node corresponding to the shortest response time is determined as the target service node that matches the target metadata corresponding to the current business operation.

7. The metadata management method according to any one of claims 1 to 6, characterized in that, The migration of the target metadata to the target service node includes: If the target metadata is in a occupied state, then wait until the target metadata is in an unoccupied state and then lock the target metadata; The locked target metadata is migrated to the target service node; Accordingly, after migrating the target metadata to the target service node, the process further includes: If the verification result corresponding to the target metadata indicates that the migration was successful, then the relevant data in the current service node is deleted, and the lock on the target metadata is released. If the verification result corresponding to the target metadata indicates that the migration was unsuccessful, then the relevant data in the target service node is deleted, and the lock on the target metadata is released.

8. A metadata management device, characterized in that, Applied to file storage systems, including: The feature tag generation module is used to statistically analyze the node status information of the metadata service node to obtain the first feature tag, and statistically analyze the file status information of the stored file to obtain the second feature tag, and determine the third feature tag corresponding to the current business operation and the corresponding business context information. The service node evaluation module is used to perform weighted evaluation of each service node based on a preset mapping matrix, combined with the first feature label, the second feature label, and the third feature label, and to determine the target service node that matches the target metadata corresponding to the current business operation based on the evaluation results; the preset mapping matrix is ​​a matrix that represents the mapping relationship between feature labels and weights; The metadata migration module is used to migrate the target metadata to the target service node to perform the current business operation when the target service node is not the current service node corresponding to the target metadata.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the metadata management method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the metadata management method as described in any one of claims 1 to 7.

Images