CN118035200A - Distributed file system metadata management method, device and equipment - Google Patents

Abstract

The invention discloses a method, a device and equipment for managing metadata of a distributed file system, wherein the method comprises the following steps: the distributed file system groups the metadata according to the catalogs, stores the metadata in each catalogs in the form of key value pairs and groups the catalogs one by one; wherein, metadata in the catalogue is stored according to deterministic sequence; the deterministic order does not change with changes in data attributes; the distributed file system employs an invariance-based sequential update for metadata operations. The invention solves the problems of high metadata storage maintenance cost, high metadata operation cost and complex logic.

Description

Distributed file system metadata management method, device and equipment

Technical Field

The present invention relates to the field of computer technology, and in particular to a distributed file system metadata management method, device and equipment.

Background technique

Distributed File System (DFS) connects multiple physical nodes together through a high-speed network to achieve unified management of multi-node resources, and provides users with a file system access interface and a unified file system namespace. Distributed file systems are generally designed based on the client/server model. The client provides a standard file system access interface for applications, and the server stores all data and metadata in the file system. Distributed file systems generally adopt a design architecture that separates data and metadata. File data is stored on multiple data servers, and file metadata is stored on a single or multiple metadata servers. Distributed file systems have the advantages of cost savings, easy management, good scalability, strong reliability and high availability, and have now been widely used.

In order to support metadata operations (i.e., metadata indexing and metadata scanning), the distributed file system needs to specify the storage order of metadata on persistent media (such as disks, persistent memory, etc.). In the prior art, one method is to split the metadata into two parts: index nodes and directory entries. To index the metadata, it is necessary to first access the directory entry and then access the corresponding index node, which brings the overhead of indirect indexing. Another method is to use the identifier of the parent directory corresponding to the metadata and the name of the metadata as the key, and the metadata itself as the value, and store the metadata in an ordered key-value storage. Maintaining the total order relationship of all key-value pairs in the key-value storage brings high overhead.

In order to ensure the atomicity and crash consistency of metadata operations, the distributed file system needs to perform sequential updates on the metadata operations on the disk to ensure that the metadata server can be restored to a consistent state after a crash. Multi-data operations such as metadata creation, metadata deletion, and metadata rename operations require the modification of more than one metadata, so for these operations, relying on the atomicity of unit data modification provided by the hardware cannot guarantee the atomicity of the entire operation. The main method used in the prior art is distributed transactions. An example of a distributed transaction is: before performing a metadata operation, all metadata involved are locked, and then all changes involved are written to the write-ahead log, and then the metadata changes are applied to the target metadata using algorithms such as two-phase commit. This process introduces additional overhead such as write-ahead logs, and the logic is relatively complex.

Summary of the invention

The present invention provides a distributed file system metadata management method, device and equipment, which solve the problems of high metadata storage and maintenance overhead, high metadata operation overhead and complex logic.

In order to solve the above problems, the present invention proposes a distributed file system metadata management method, comprising:

The distributed file system groups metadata by directory, stores metadata in each directory in the form of key-value pairs and groups them directory by directory; metadata in the directory is stored in a deterministic order; the deterministic order does not change with changes in data attributes;

Distributed file systems use invariant-based sequential updates for metadata operations.

In one embodiment of the present invention, the metadata in each directory is stored in the form of key-value pairs and grouped directory by directory, specifically including: storing the individual metadata in each directory in the form of key-value pairs, wherein the key includes the identifier of the parent directory and the metadata name, and the value is the metadata itself; logically grouping the metadata belonging to the same parent directory.

In one embodiment of the present invention, the invariants include: A. the parent directory metadata updated in the metadata operation is dependent on its sub-directory and sub-file metadata; B. the renaming operation will not destroy the uniqueness of the metadata identifier.

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata creation operations: first create target metadata and determine the target metadata creation time; then modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory; when an abnormality occurs in the distributed file system, modify the last data modification time and the last metadata modification time of the parent directory to the target metadata creation time; modify the number of hard links of the parent directory to 2 + the number of subdirectories for recovery.

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata deletion operations: first mark the target metadata as invalid and record the deletion time of the target metadata, then modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory, and finally delete the target metadata; when an abnormality occurs in the distributed file system, modify the last data modification time and the last metadata modification time of the parent directory to the target metadata deletion time; modify the number of hard links of the parent directory to 2 + the number of subdirectories; delete the invalid target metadata for recovery.

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata renaming operations: first create new metadata with the same identifier as the old metadata, then mark the old metadata as invalid, and finally modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory of the old metadata and the new metadata, and delete the old metadata; when an abnormality occurs in the distributed file system, delete metadata with duplicate identifiers and invalid metadata; modify the last data modification time and the last metadata modification time of the old parent directory to the old metadata deletion time; modify the last data modification time and the last metadata modification time of the new parent directory to the new metadata creation time; modify the number of hard links of the two parent directories of the old metadata and the new metadata to 2+the number of subdirectories for recovery.

In one embodiment of the present invention, the metadata includes directory metadata, file metadata and hard link metadata.

A distributed file system metadata management device, comprising:

The metadata grouping storage module is used to group the metadata according to the directory, store the metadata in each directory in the form of key-value pairs and group them directory by directory; wherein the metadata in the directory is stored in a deterministic order; the deterministic order will not change with the change of data attributes;

The metadata operation update module is used to update metadata operations sequentially based on invariants.

A distributed file system metadata management device, comprising:

at least one processor; and,

A memory connected to the at least one processor via a bus; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed to implement the method described in any one of the above embodiments.

A non-volatile storage medium stores computer executable instructions, wherein the computer executable instructions are executed by a processor to implement the method described in any one of the above embodiments.

The present invention provides a distributed file system metadata management method, device and equipment, which at least include the following beneficial effects: through directory-by-directory grouping, ordered key-value storage and a sequential update strategy based on invariants, the efficiency of metadata management is effectively improved, the file system performance is optimized, the complexity of metadata management is simplified, the consistency and reliability of metadata are enhanced, and an efficient, stable and reliable metadata management method is provided for a distributed file system.

BRIEF DESCRIPTION OF THE DRAWINGS

The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

FIG1 is a schematic diagram of steps of a distributed file system metadata management method provided by an embodiment of the present invention;

FIG2 is a flowchart of a directory-by-directory ordered key-value storage metadata according to an embodiment of the present invention;

3 is a schematic diagram of performing directory metadata creation and deletion operations provided by an embodiment of the present invention;

FIG4 is a schematic diagram of performing a directory metadata renaming operation provided by an embodiment of the present invention;

5 is a schematic diagram of a distributed file system metadata management device provided by an embodiment of the present invention;

FIG6 is a schematic diagram of a distributed file system metadata management device provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be described clearly and completely in conjunction with the specific embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

It should be noted that it is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in the present invention can be combined with other embodiments without conflict. Unless otherwise defined, the technical terms or scientific terms involved in the present invention should be the usual meanings understood by people with ordinary skills in the technical field to which the present invention belongs. The words "one", "a", "a", "the" and the like involved in the present invention do not indicate a quantitative limitation and may represent the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in the present invention are intended to cover non-exclusive inclusions; the terms "first", "second", "third", etc. involved in the present invention are merely to distinguish similar objects and do not represent a specific ordering of objects.

In metadata operations, metadata scanning operations do not occur on any interval, they only occur within a directory, and they may not return the scanned metadata in lexicographical order.

The number of multivariate data operations is limited, and they all follow two metadata invariants when updating metadata: A. The parent directory metadata (last data block modification time, last metadata modification time, number of hard links) updated in the operation is dependent on its subdirectory and subfile metadata; B. The rename operation will not destroy the uniqueness of the metadata identifier (such as ID).

Based on the above two principles, the present invention proposes a distributed file system metadata management method, device and equipment, which are described in detail below.

FIG1 is a schematic diagram of the steps of a distributed file system metadata management method provided by an embodiment of the present invention. The distributed file system metadata management method shown in FIG1 can be executed by a distributed file system metadata management device, and the distributed file system metadata management device can be set on a server. For example, the server can be a physical server including an independent host, a virtual server carried by a host cluster, a cloud server, etc., which is not limited by the embodiment of the present invention. The distributed file system metadata management method can include the following steps:

S110: The distributed file system groups the metadata according to directories, stores the metadata in each directory in the form of key-value pairs, and groups them directory by directory.

The metadata in the directory is stored in a deterministic order; the deterministic order does not change with changes in data attributes.

Specifically, the distributed file system uses a key-value storage system that groups and orders each directory to store metadata. The metadata between directories does not need to be ordered, and the metadata within a directory can be stored in an order other than the lexicographic order.

In one embodiment of the present invention, the metadata includes directory metadata, file metadata, and hard link metadata.

Specifically, metadata mainly includes three categories, namely directory metadata, file metadata and hard link metadata. The "ID" mentioned below refers to the identifier of metadata, which only needs to ensure global uniqueness and does not necessarily appear in the form of ID in the traditional sense, that is, number.

For directory metadata and file metadata, the information includes but is not limited to: metadata ID, whether it is valid, metadata type, metadata permission information (including user ID, group ID and access rights), metadata timestamp information (including creation time, deletion time, last data modification time, last access time and last metadata modification time), metadata corresponding directory or file size, metadata corresponding hard link number, metadata corresponding file block mapping information, etc. The above metadata information may add or delete certain items according to different implementations.

The hard link metadata is used to point to the metadata of the file being hard linked. Its information includes the metadata ID, whether it is valid, the metadata type, and the metadata ID of the file it points to.

In one embodiment of the present invention, the metadata in each directory is stored in the form of key-value pairs and grouped directory by directory, specifically including: storing the individual metadata in each directory in the form of key-value pairs, wherein the key includes the identifier of the parent directory and the metadata name, and the value is the metadata itself; and logically grouping the metadata belonging to the same parent directory.

Specifically, a single metadata is stored in the key-value storage system in the form of a key-value pair, where the key includes the ID of the parent directory and the metadata name, and the value is the metadata itself.

Metadata belonging to the same parent directory are logically grouped together, which guarantees the relative order of metadata belonging to the same parent directory, i.e., grouped together. Metadata belonging to different parent directories may or may not be further grouped. They may be ordered or unordered.

The metadata belonging to the same parent directory only has a logical organizational relationship, and there is no requirement for its organizational relationship at the physical storage level. The logical organizational relationship means that it supports scanning the metadata in the same group in a certain order. The present invention does not have constraints on the storage location level of metadata grouped together. For example, it can be distributed to different metadata servers. In order to symmetrically perform hash calculations to obtain hash values (integers), they are sorted according to the hash values. The "deterministic order" here is defined as: if one metadata has key A and the other metadata has key B, then in this order, either A<B or A>B, and this order will not change with the scale of the data structure itself.

FIG. 2 is a flow chart of a distributed file system metadata organization method according to an embodiment of the present invention.

Specifically, in Figure 2, a hash table is used to index between directories, the key of which is a directory identifier (which can be a directory ID), and the value is an ordered tree index structure corresponding to the directory. The hash table does not guarantee the orderliness between directories, that is, it does not guarantee that the key-value pairs between directories maintain a specific order such as a lexicographic order.

Each directory has an ordered tree index structure that stores key-value pairs corresponding to the metadata. The key of each metadata key-value pair is the metadata name, and the value is the corresponding metadata. The ordered tree index structure in each directory ensures the hash order between key-value pairs, that is, after the key is input into the existing hash function to calculate the hash value, the key-value pairs are sorted according to the size of the hash value obtained.

It should be noted that Figure 2 is only an example of a directory-by-directory ordered key-value storage implementation, wherein a hash table index may not be used between directories, and a tree index in hash order may not be used within a directory, and the index structure must meet the implementation requirements of the present invention.

By logically organizing metadata of the same parent directory by directory grouping, metadata under the directory can be easily searched, accessed, and operated. This grouping method reduces the search scope of the system when processing metadata, thereby improving the efficiency of metadata management. Since the directory-by-directory grouping method reduces the search scope of metadata, the distributed file system can locate relevant metadata more quickly when processing file read, write, delete, rename, and other operations, thereby improving the overall performance of the distributed file system. Compared with maintaining a unified total order relationship for all metadata, this directory-by-directory grouping method only needs to maintain a total order relationship for the metadata under each directory. This relaxes the requirements for metadata storage order and reduces the system's additional overhead in maintaining metadata order relationships.

S120: The distributed file system adopts an invariant-based sequential update for metadata operations.

Specifically, the distributed file system uses invariant-based sequential updates for metadata operations to ensure their atomicity and crash consistency.

In one embodiment of the present invention, the invariants include: A. the parent directory metadata updated in the metadata operation is dependent on its sub-directory and sub-file metadata; B. the rename operation will not destroy the uniqueness of the metadata identifier.

Based on storing metadata in an ordered key-value storage system, metadata operations are managed by using an invariant-based sequential update method, making metadata management simpler and more direct. This method simplifies the logic of ensuring the atomicity of file system operations, reduces system complexity and performance overhead, and improves the performance and stability of the file system, enabling it to better cope with high-concurrency and large-scale data processing scenarios.

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata creation operations: first create target metadata and determine the target metadata creation time; then modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory; when an abnormality occurs in the distributed file system, modify the last data modification time and the last metadata modification time of the parent directory to the target metadata creation time; modify the number of hard links of the parent directory to 2 + the number of subdirectories for recovery.

Specifically, for metadata creation operations, it first creates the target metadata (including the creation time of the target metadata), and then modifies the last data modification time, the last metadata modification time, and the number of hard links of the parent directory metadata. When a file system abnormality occurs, it is restored by modifying the last data modification time and the last metadata modification time of the parent directory to the target metadata creation time; and modifying the number of hard links of the parent directory to 2 + the number of subdirectories.

For example, FIG. 3 is a schematic diagram of performing directory metadata creation/deletion operations according to an embodiment of the present invention.

In (a) of FIG3 , it creates directory metadata /A/B at time t=5. First, it inserts the /A/B directory metadata key-value pair and attaches its creation time btime=5. Then, it updates the last data block modification time ctime and the last metadata modification time mtime of the parent directory metadata to the creation time btime=5 of /A/B, and updates the number of hard links nlink of /A to the number of its subdirectories + 2 (nlink=3 in this example).

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata deletion operations: first mark the target metadata as invalid and record the deletion time of the target metadata, then modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory, and finally delete the target metadata; when an abnormality occurs in the distributed file system, modify the last data modification time and the last metadata modification time of the parent directory to the target metadata deletion time; modify the number of hard links of the parent directory to 2 + the number of subdirectories; delete the invalid target metadata for recovery.

Specifically, for metadata deletion operations, it first marks the target metadata as invalid and records the target metadata deletion time, modifies the last data modification time, last metadata modification time and number of hard links of the parent directory metadata, and finally deletes the target metadata. When a file system abnormality occurs, it modifies the last data modification time and last metadata modification time of the parent directory to the target metadata deletion time; modifies the number of hard links of the parent directory to 2 + the number of subdirectories; and deletes the invalid target metadata for recovery.

For example, in (b) of FIG3 , it deletes the directory metadata /A/B at time t=6. First, it marks the /A/B directory metadata key-value pair as invalid, and modifies its deletion time dtime to the current time 6. Then, it updates the last data block modification time ctime and the last metadata modification time mtime of the parent directory metadata to the deletion time dtime=6 of /A/B, and updates the hard link number nlink of /A to the number of its subdirectories + 2 (in this example, nlink=2). Finally, it deletes the metadata key-value pair /A/B.

In one embodiment of the present invention, metadata operations are updated sequentially based on invariants, specifically including: for metadata renaming operations: first create new metadata with the same identifier as the old metadata, then mark the old metadata as invalid, and finally modify the last data modification time, the last metadata modification time and the number of hard links of the parent directory of the old metadata and the new metadata, and delete the old metadata; when an abnormality occurs in the distributed file system, delete metadata with duplicate identifiers and invalid metadata; modify the last data modification time and the last metadata modification time of the old parent directory to the old metadata deletion time; modify the last data modification time and the last metadata modification time of the new parent directory to the new metadata creation time; modify the number of hard links of the two parent directories of the old metadata and the new metadata to 2+the number of subdirectories for recovery.

Specifically, for metadata renaming operations, it first creates new metadata with the same ID as the old metadata, then marks the old metadata as invalid, and finally modifies the last data modification time, last metadata modification time, and number of hard links of the parent directory metadata of the new and old metadata, and deletes the old metadata. When a file system abnormality occurs, it deletes metadata with duplicate IDs and invalid metadata; modifies the last data modification time and last metadata modification time of the old parent directory to the old metadata deletion time; modifies the last data modification time and last metadata modification time of the new parent directory to the new metadata creation time; and modifies the number of hard links of the two parent directories to 2 + the number of subdirectories for recovery.

For example, FIG4 is a schematic diagram of an embodiment of the present invention performing a directory metadata renaming operation, which renames the directory metadata /A/C to /B/C at time t=5. The ID of the C directory is 4. First, it creates the /B/C directory metadata, which has the same ID as /A/C, 4, and its creation time btime=5. Then, it marks the /A/C directory metadata as invalid.

Then, it updates the last data block modification time, the last metadata modification time and the number of hard links of the two involved parent directories, namely /A and /B. It updates the last data block modification time and the last metadata modification time of /A to the deletion time dtime of /A/C (in this example, dtime=5), updates the number of hard links nlink of /A to the number of its subdirectories + 2 (in this example, nlink=2), updates the last data block modification time ctime and the last metadata modification time mtime of /B to the creation time btime of /B/C (in this example, btime=5), and updates the number of hard links nlink of /B to the number of its subdirectories + 2 (in this example, nlink=3). Finally, delete the metadata key-value pair /A/C.

It should be noted that the examples of directory metadata creation/deletion and renaming operations listed in Figures 3 and 4 are because they will modify the number of hard links, not because the method is not applicable to other types of metadata. For other types of metadata such as file metadata, metadata can also be correctly updated according to the method of the present invention.

In the case of anomalies in the distributed file system, metadata consistency can be restored by modifying the last data modification time, the last metadata modification time, and the number of hard links of the parent directory, and deleting invalid or duplicate metadata. This recovery mechanism ensures that metadata remains correct and complete even when a system failure occurs.

The above is a distributed file system metadata management method provided by an embodiment of the present invention. Based on the same inventive idea, an embodiment of the present invention also provides a corresponding distributed file system metadata management device, as shown in FIG5 , including:

The metadata grouping storage module 510 is used to group the metadata according to the directory, store the metadata in each directory in the form of key-value pairs and group them directory by directory; wherein the metadata in the directory is stored in a deterministic order; the deterministic order will not change with the change of data attributes;

The metadata operation update module 520 is used to update metadata operations in a sequential manner based on invariants.

The embodiment of the present invention further provides a corresponding distributed file system metadata management device, as shown in FIG6 , including:

At least one processor 602 (processor), a communication interface 604 (Communications Interface), a memory 606 (memory) and a communication bus 608; wherein the processor 602, the communication interface 604, and the memory 606 communicate with each other via the communication bus 608; the processor 602 can call the logic instructions stored in the memory 606 to enable at least one processor 602 to execute the steps of the above-mentioned embodiments.

Based on the same idea, some embodiments of the present invention also provide a medium corresponding to the above method.

Some embodiments of the present invention provide a storage medium storing computer executable instructions, and the computer executable instructions are executed by a processor to implement the steps of the above embodiments.

On the other hand, an embodiment of the present application also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the steps of a distributed file system metadata management method provided in the above embodiments.

Each embodiment of the present invention is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, medium and program product embodiments, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

The devices, media, and program products provided in the embodiments of the present invention correspond one-to-one to the methods. Therefore, the devices, media, and program products also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices, media, and program products will not be repeated here.

It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or method. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the presence of other identical elements in the process, method, commodity or method including the elements.

The above are only embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail above with general descriptions and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made on the basis of the present invention. Therefore, these modifications or improvements made on the basis of not departing from the spirit of the present invention all belong to the scope of protection claimed in the present invention.

Claims (10)

1.A method for managing metadata of a distributed file system, comprising:

The distributed file system groups the metadata according to the catalogs, stores the metadata in each catalogs in the form of key value pairs and groups the catalogs one by one; wherein, metadata in the catalogue is stored according to deterministic sequence; the deterministic order does not change with changes in data attributes;

the distributed file system employs an invariance-based sequential update for metadata operations.

2. The method for managing metadata of a distributed file system according to claim 1, wherein the storing of metadata in each directory in the form of key-value pairs and the grouping of directories comprises:

storing the individual metadata in each directory in the form of key-value pairs, wherein the keys comprise identifiers of parent directories and metadata names, and the values are metadata per se;

Metadata belonging to the same parent directory is logically grouped together.

3. The distributed file system metadata management method of claim 1, wherein the invariance comprises:

A. The parent directory metadata updated in the metadata operation is dependent on its child directory and child file metadata;

B. the renaming operation does not destroy the uniqueness of the metadata identifier.

4. A method for metadata management for a distributed file system according to claim 3, wherein the metadata operations are updated in an order based on invariance, and specifically comprising:

For metadata creation operations: firstly, creating target metadata, and determining target metadata creation time; the last data modification time, the last metadata modification time and the hard link number of the father catalog are modified again;

When the distributed file system is abnormal, the last data modification time and the last metadata modification time of the parent directory are modified into target metadata creation time; the hard link number of the parent directory is modified to be 2+ subdirectory number for recovery.

5. A method for metadata management for a distributed file system according to claim 3, wherein the metadata operations are updated in an order based on invariance, and specifically comprising:

For metadata delete operations: marking target metadata as invalid, recording the deleting time of the target metadata, modifying the last data modification time, the last metadata modification time and the hard link number of the parent directory, and finally deleting the target metadata;

When the distributed file system is abnormal, the last data modification time and the last metadata modification time of the parent directory are modified into target metadata deletion time; modifying the hard link number of the parent directory to be 2+ subdirectory number; invalid target metadata is deleted for recovery.

6. A method for metadata management for a distributed file system according to claim 3, wherein the metadata operations are updated in an order based on invariance, and specifically comprising:

for metadata renaming operations: creating new metadata identical to the old metadata identifier, marking the old metadata as invalid, and finally modifying the last data modification time, the last metadata modification time and the hard link number of the parent directory of the old metadata and the new metadata, and deleting the old metadata;

Deleting the metadata repeated by the identifier and the invalid metadata when the distributed file system is abnormal; modifying the last data modification time of the old father catalog into the old metadata deletion time; modifying the last data modification time of the new parent directory and the last metadata modification time into new metadata creation time; the hard link numbers of the two parent directories of the old metadata and the new metadata are modified to be 2+ child directory numbers for recovery.

7. The distributed file system metadata management method of claim 1 wherein the metadata comprises directory metadata, file metadata, and hard link metadata.

8. A distributed file system metadata management apparatus, comprising:

The metadata grouping storage module is used for grouping metadata according to catalogs, storing the metadata in each catalogue in a key value pair mode and grouping the catalogues one by one; wherein, metadata in the catalogue is stored according to deterministic sequence; the deterministic order does not change with changes in data attributes;

And the metadata operation updating module is used for updating the metadata operation in an order based on invariance.

9. A distributed file system metadata management apparatus, comprising:

At least one processor; and

A memory communicatively coupled to the at least one processor via a bus; wherein,

The memory stores instructions executable by the at least one processor to implement the method of any one of claims 1-7.

10. A non-volatile storage medium storing computer executable instructions for execution by a processor to implement the method of any one of claims 1-7.