WO2020029748A1

WO2020029748A1 - Methods for snapshot management and difference bitmap generation in distributed storage system, and product

Info

Publication number: WO2020029748A1
Application number: PCT/CN2019/095901
Authority: WO
Inventors: 张航; 张良翔; 尹青云
Original assignee: 华为技术有限公司
Priority date: 2018-08-08
Filing date: 2019-07-13
Publication date: 2020-02-13
Also published as: CN109165120A; CN109165120B

Abstract

Disclosed in embodiments of the present invention are methods for snapshot management and difference bitmap generation in a distributed storage system, and a product. The distributed storage system comprises N storage nodes. The N storage nodes are used for providing a storage space for a first logical unit number (LUN). The method for snapshot management comprises: a storage node Ni generates a first record; the first record is only used for recording an update address corresponding to a logical address of the first LUN distributed at the storage node Ni at a first snapshot moment; the logical address of the first LUN distributed at the storage node Ni and a logical address of a second LUN distributed at the storage node Ni belong to a same partition; the storage node Ni generates a second record; the second record is only used for recording the update address corresponding to the logical address of the first LUN distributed at the storage node Ni at the second snapshot moment; the first snapshot moment and the second snapshot moment are adjacent snapshot moments. The present invention can improve the obtaining efficiency of difference bitmaps.

Description

Method and product for generating management snapshot and difference bitmap in distributed storage system

Technical field

The present invention relates to the field of distributed storage technology, and in particular, to a method and product for generating a management snapshot and a difference bitmap in a distributed storage system.

Background technique

A distributed storage system usually adopts the architecture shown in FIG. 1A, including a storage node cluster. The user controls the storage node cluster through the client, such as triggering the storage node cluster to create a snapshot for a logical unit number (LUN), create a difference bitmap, and perform read and write operations. The backup storage system is used to back up the data of the LUNs in the storage node cluster. When the backup storage system is synchronized with the distributed storage system for the first time, the client triggers the storage node cluster to create a snapshot of the LUN, or the distributed storage system creates a snapshot of the LUN at a predetermined interval. For example, the first snapshot with version number 0 is created. The backup storage system copies all the data of the LUNs in the storage node cluster. Therefore, all the data of the snapshot created this time is copied to the backup storage system. Later, you can use snapshots to back up the incremental data of the LUNs in the storage node cluster to the backup storage system.

In a distributed storage system scenario, the structure of a storage node cluster can be as shown in Figure 1B, including N storage nodes, where N is an integer greater than or equal to 2, and each storage node contains a hard disk.

Block modification tracking (CBT), also called difference bitmap, is used to record the LUN data increment between two snapshot versions. In a distributed storage system, a LUN is distributed across multiple storage nodes. A distributed storage system contains multiple partitions, and a storage node provides storage space for multiple partitions. In a distributed storage system, when a client accesses a logical block address (LBA) in a LUN, it first determines the partition where the LBA is located, and then queries the partition mapping table to determine the storage node corresponding to the partition. The client accesses the corresponding storage. node. The storage node provides storage space corresponding to the LBA.

Because a storage node provides storage space for multiple partitions, LBAs for multiple LUNs are distributed on the same partition. In the existing distributed storage system, when a snapshot operation is performed on a certain LUN, the storage node usually records the snapshot data on the storage node with partition as the granularity. For example, if partition A of storage node A contains LBAs with multiple LUNs, then storage node A records all snapshot data on partition A. The distributed storage system needs to query all snapshot data of partition A recorded by storage node A to query the snapshot data of LBA distributed to partition A of storage node A. Generally, the LBA of a LUN is distributed on multiple storage nodes. Therefore, the snapshot incremental data collection takes a long time in a distributed storage system scenario.

Summary of the invention

Embodiments of the present invention provide a management snapshot and a differential bitmap generation method and product in a distributed storage system, which are used to improve the acquisition efficiency of the differential bitmap.

In a first aspect, an embodiment of the present invention provides a method for managing snapshots in a distributed storage system. The distributed storage system includes N storage nodes, and the N storage nodes are configured to provide a first logical unit number LUN. Storage space; the method includes:

The storage node Ni generates a first record; the first record is only used to record an update address corresponding to the logical address of the first LUN distributed at the storage node Ni at a first snapshot time; wherein the first LUN is distributed The logical address of the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is each value from 1 to N;

The storage node Ni generates a second record, and the second record is only used to record an updated address corresponding to a logical address where the first LUN is distributed on the storage node Ni at a second snapshot time; wherein the first The snapshot time and the second snapshot time are adjacent snapshot times.

When it needs to be explained, each record in the embodiment of the present invention separately records the logic corresponding to the data block in which the data changes when the data is written according to the received write input instruction between the current snapshot and the previous previous snapshot. Address sequence. For example, the second record records a logical address sequence corresponding to a data block in which data changes when the data is written according to the write input instruction between the first snapshot time and the second snapshot time. The write input instruction usually carries the data to be written, the identifier of the LUN to be written, and the offset of the data block corresponding to the write address in the LUN to be written. Which storage node is it? After each snapshot, the storage node Ni generates an empty record for the first LUN. Whenever a write input instruction to operate the first LUN is received and the write position is located at the storage node Ni, a record information is added to the record. , Used to record the write position, when there are multiple write input instructions to perform a write input operation on the first LUN, and the write position is located at the storage node Ni, multiple records are added to the record accordingly, and the record write input instruction is written the address of.

With the technical solution provided by the embodiment of the present invention, when obtaining the difference bitmap of the first LUN between two adjacent snapshots, the difference bitmap data is collected at the granularity of the storage node, and the storage node manages the records corresponding to the difference data according to the LUN granularity. When collecting the difference bitmap, the client only needs to establish N connections with N storage nodes. Compared to the prior art, which collects the difference bitmap with partition granularity, and the storage node manages the index corresponding to the difference data according to the partition granularity, an embodiment of the present invention The number of communication connections between the client and the storage node cluster is reduced, the efficiency of obtaining the difference bitmap is improved, and it is also beneficial for the backup storage system to perform timely backup of the data in the storage node cluster.

In some possible implementation manners of the present invention, the storage node Ni provides the client with the first record and the second record.

In some possible implementation manners of the present invention, the first record and the second record are both linked lists.

In some possible implementation manners of the present invention, the first record and the second record are both an array.

In a second aspect, an embodiment of the present invention provides a method for generating a difference bitmap in a distributed storage system, where the distributed storage system includes N storage nodes, and the N storage nodes are used for a first logical unit number. LUN provides storage space;

The client obtains the first record and the second record from the storage node Ni; wherein the first record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the first snapshot time Wherein the logical address of the first LUN distributed on the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is 1 to Each value of N; the second record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the second snapshot time; wherein the first snapshot time and The second snapshot moment is an adjacent snapshot moment;

The client generates a different position of the first LUN on the storage node Ni according to the first record and the second record.

In some possible implementation manners of the present invention, the client sends a first snapshot instruction to the storage node Ni at the first snapshot moment; the client sends the storage node Ni to the storage node Ni at the second snapshot moment Sending a second snapshot instruction; the first snapshot instruction and the second snapshot instruction are respectively used to perform a snapshot operation on the first LUN.

In a third aspect, an embodiment of the present invention provides a storage node in a distributed storage system, where the storage node is any one of N storage nodes included in the distributed storage system, and the N storage nodes are used for And providing storage space for a first logical unit number LUN, the storage node includes:

A first processing unit, configured to generate a first record; the first record is only used to record an update address corresponding to a logical address of the first LUN distributed on the storage node at a first snapshot time; A logical address of a LUN distributed on the storage node and a logical address of a second LUN distributed on the storage node belong to the same partition, and N is an integer not less than 2;

A second processing unit, configured to generate a second record, where the second record is only used to record an update address corresponding to a logical address where the first LUN is distributed on the storage node at a second snapshot moment; A snapshot time and the second snapshot time are adjacent snapshot times.

In some possible implementation manners of the present invention, the storage node further includes:

A providing unit is configured to provide the client with the first record and the second record.

In a fourth aspect, an embodiment of the present invention provides a client in a distributed storage system. The distributed storage system includes N storage nodes, and the N storage nodes are configured to provide storage for a first logical unit number LUN. Space; the client includes:

An obtaining unit, configured to obtain a first record and a second record from the storage node Ni; wherein the first record is only used to record a logical address corresponding to the first LUN distributed on the storage node Ni at a first snapshot time The updated address of the first LUN distributed on the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i For each value from 1 to N; the second record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the second snapshot time; wherein the first record The snapshot time and the second snapshot time are adjacent snapshot times;

A generating unit is configured to generate a difference position of the first LUN on the storage node Ni according to the first record and the second record.

In some possible implementation manners of the present invention, the client further includes:

A sending unit, configured to send a first snapshot instruction to the storage node Ni at the first snapshot time, and send a second snapshot instruction to the storage node Ni at the second snapshot time; the first snapshot instruction And the second snapshot instruction are respectively used to perform a snapshot operation on the first LUN.

With the technical solution provided by the embodiment of the present invention, when obtaining the difference bitmap of the first LUN between two adjacent snapshots, the difference bitmap data is collected at the granularity of the storage node, and the storage node manages the records corresponding to the difference data according to the LUN granularity. When collecting the difference bitmap, the client only needs to establish N connections with N storage nodes. Compared to the prior art, which collects the difference bitmap with partition granularity, and the storage node manages the index corresponding to the difference data according to the partition granularity, embodiments of the present invention The number of communication connections between the client and the storage node cluster is reduced, the efficiency of obtaining the difference bitmap is improved, and it is also beneficial for the backup storage system to perform timely backup of the data in the storage node cluster.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly explain the technical solutions in the embodiments of the present invention or the background art, the drawings that are needed in the embodiments of the present invention or the background art will be described below.

FIG. 1A is a schematic architecture diagram of a distributed system.

FIG. 1B is a schematic diagram of a communication connection between a client and a storage node cluster when obtaining a difference bitmap in the prior art.

2 is a schematic diagram of a communication connection between a client and a storage node cluster when obtaining a difference bitmap according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of establishing a communication connection between a client and a storage node in the prior art and an embodiment of the present invention.

FIG. 4 is an interaction schematic diagram of a distributed system in an embodiment of the present invention.

detailed description

The following describes the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention.

FIG. 1A is a schematic architecture diagram of a distributed storage system. It includes a client 101, a storage node cluster 102, and a backup storage system 103. The user controls the storage node cluster 102 through the client 101, such as triggering the storage node cluster 102 to create a snapshot of the LUN, create a difference bitmap, and perform read and write operations. The backup storage system 103 is used to back up the data of the LUN in the storage node cluster 102. When the backup storage system 103 is synchronized with the distributed storage system for the first time, the client 101 triggers the storage node cluster 102 to create a snapshot for the LUN. For example, a snapshot with version number 0 is created. The backup storage system 103 backs up all the data of the LUN in the storage node cluster 102. All the data in the snapshot created this time is copied to the backup storage system 103. The snapshot can be used later. The user continues to trigger the storage node cluster 102 to create a snapshot for the LUN through the client 101. Each time the version number of the newly created snapshot is plus 1. By comparing the differences between two adjacent snapshots, a difference bitmap can be obtained. According to the difference bitmap, the difference data corresponding to the difference bitmap of the adjacent snapshots is transmitted to the backup storage system 103, and the backup storage system 103 receives the difference. Update data is written after the data.

FIG. 1B is a schematic diagram of a communication connection between a client and a storage node cluster when obtaining a difference bitmap in the prior art. As shown in FIG. 1B, a storage node cluster may include N storage nodes, that is, storage node 1, storage node 2, storage node 3, ..., storage node N, where N is an integer greater than or equal to 2, and each storage node may It's a hard disk. Each storage node includes 3 partitions. For example, storage node 1 includes partitions: partition 1, partition n + 1, and partition 2n + 1. Storage node 2 includes: partition 2, partition n + 2, partition 2n + 2, and storage. Node 3 includes: partition 3, partition n + 3, partition 2n + 3, and storage node N includes: partition n, partition 2n, and partition 3n. In order to obtain the difference bitmap of the snapshot, the client needs to establish a communication connection with each partition in each storage node. As shown in Figure 1B, the client needs to communicate with the three partitions in storage node 1 (partition 1, partition n + 1. The partition 2n + 1) establishes a communication connection to obtain the location of the data block where the LUN in the storage node 1 is updated. The client needs to establish a communication connection with three partitions (partition 2, partition n + 2, and partition 2n + 2) in storage node 2 in order to obtain the location of the data block where the LUN in data storage node 2 has been updated. The client needs to establish a communication connection with the three partitions (partition 3, partition n + 3, and partition 2n + 3) in the storage node 3 in order to obtain the location of the data block where the LUN in the storage node 3 is updated. The client needs to establish a communication connection with the three partitions (partition n, partition 2n, and partition 3n) in the storage node N in order to obtain the location of the data block where the LUN in the storage node N is updated.

An embodiment of the present invention discloses a method for managing snapshots in a distributed storage system. The distributed storage system includes N storage nodes, and the N storage nodes are configured to provide storage space for a first logical unit number LUN. ;

The storage node Ni generates a first record; the first record is only used to record an update address corresponding to the logical address of the first LUN distributed at the storage node Ni at a first snapshot time; wherein the first LUN distribution The logical address of the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is each value from 1 to N; that is: In one embodiment, i can be any integer value from 1 to N. The storage node Ni generates a second record, and the second record is only used to record an updated address corresponding to a logical address where the first LUN is distributed on the storage node Ni at a second snapshot time; wherein the first The snapshot time and the second snapshot time are adjacent snapshot times. For example, if a snapshot operation is performed on the first LUN every 10 minutes, for example, a snapshot operation is performed on the first LUN at 22:00, and a snapshot operation is performed again on the first LUN at 22:10, then 22:00 and 22:10 are adjacent. Snapshot moments. A snapshot moment corresponds to a version number of the snapshot. The update address corresponding to the logical address of the first LUN distributed on the storage node Ni is the position of the data block where the first LUN is distributed on the logical address of the storage node Ni to be updated.

In some possible embodiments of the present invention, the storage node Ni provides the client with the first record and the second record. The first record and the second record may be a linked list or an array.

An embodiment of the present invention also discloses a method for generating a difference bitmap in a distributed storage system. The distributed storage system includes N storage nodes, and the N storage nodes are configured to assign a first logical unit number LUN. Provide storage space; the client obtains the first record and the second record from the storage node Ni; wherein the first record is only used to record the logical address where the first LUN is distributed on the storage node Ni at the first snapshot time A corresponding update address; wherein the logical address of the first LUN distributed on the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is each value from 1 to N; the second record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the second snapshot time; A snapshot time and the second snapshot time are adjacent snapshot times; the client generates a difference position of the first LUN on the storage node Ni according to the first record and the second record.

In some possible embodiments of the present invention, the method further includes: the client sends a first snapshot instruction to the storage node Ni at the first snapshot time; the client at the second snapshot time Sending a second snapshot instruction to the storage node Ni; the first snapshot instruction and the second snapshot instruction are respectively used to perform a snapshot operation on the first LUN.

In some possible embodiments of the present invention, the first record and the second record may be a linked list or an array.

In order to facilitate understanding of the technical solutions to be protected in the embodiments of the present invention, the technical solutions to be protected in the present invention are described below with reference to specific examples. In the following embodiments, a record is described by taking a linked list as an example.

The left side of FIG. 3 is a schematic diagram of establishing a communication connection between a client and a storage node in the prior art. As shown on the left side of the prior art, in order to obtain a difference bitmap in the prior art, it is necessary to traverse the specified logical unit (for example, the logical unit identifier is 111). For all partitions, the location of the storage block where new data is entered in each partition is found according to the index corresponding to each partition. For example, the client first establishes a communication connection with the partition with the logical unit ID of 111 and the ID of 0. The partition with the value of 0 starts to search. Through the search, two data in the partition are updated, and the corresponding storage block identifiers are 1 and 3. Further, it is necessary to continue to find the location where the data update has occurred in the partition identified by 50. To this end, the client needs to establish a communication connection with the partition identified by 50 first, and obtain the storage block corresponding to the update location in the partition by searching. Identified as 7, 9, and 10. Further, the client also needs to establish a communication connection with a partition with a logical unit ID of 111 in the storage node and a partition with an ID of 75 in the logical unit, and find the storage blocks corresponding to the update positions in the partition with the

IDs

12 and 14. Similarly, the client also needs to establish a communication connection with the partition identified as 51200 in the logical unit identified by the logical node of the storage node as 111, and find that the storage block corresponding to the update position in the partition is identified as 1023997. It can be known from the foregoing description that, in the prior art, in order to obtain the position of the changed data in each storage node, communication connections with multiple partitions need to be established separately, which takes a long time.

FIG. 2 is a schematic diagram of a communication connection between a client and a storage node cluster when obtaining a difference bitmap provided by the present invention. From this figure, in order to obtain the difference bitmap, the client only needs to establish a connection with each storage node once. Specifically, as shown in FIG. 2, the storage node cluster includes N storage nodes, that is, storage node 1, storage node 2, storage node 3, ..., storage node N, where N is an integer greater than or equal to 2, and each storage node A node can be a hard disk. Each storage node includes 3 partitions. For example, storage node 1 includes partitions: partition 1, partition n + 1, and partition 2n + 1. Storage node 2 includes: partition 2, partition n + 2, partition 2n + 2, and storage. Node 3 includes: partition 3, partition n + 3, partition 2n + 3, and storage node N includes: partition n, partition 2n, and partition 3n. In order to obtain the difference bitmap, the client only needs to establish a communication connection with each storage node, as shown in Figure 1B, the client only needs to establish a communication connection with storage node 1, and then the first LUN in storage node 1 can be obtained. The location of the data block for the data update. The client only needs to establish a communication connection with the storage node 2 to obtain the position of the data block where the first LUN in the storage node 2 is updated. The client only needs to establish a communication connection with the storage node 3 once to obtain the position of the data block where the first LUN in the storage node 3 is updated with data. The client only needs to establish a communication connection with the storage node N to obtain the position of the data block where the first LUN in the storage node N is updated with data.

The right side of FIG. 3 is a schematic diagram of establishing a communication connection between a client and a storage node in an embodiment of the present invention. As shown on the right side of FIG. 3, in order to obtain a difference bitmap, the present invention only needs to establish the same storage node with each storage node separately. The communication connection is sufficient. Specifically, by looking up the index corresponding to the logical unit identifier of 111 and two adjacent snapshots, it can be obtained that the data update position of the first LUN in the storage node is 1, 3, 7, 9 , 10, 12, 14, 1023998. It can be known from the foregoing description that in the embodiment of the present invention, the number of times that a client establishes a communication connection with each storage node is reduced, and the efficiency of obtaining a difference bitmap is improved.

FIG. 4 is an interaction schematic diagram of a distributed storage system in an embodiment of the present invention. The distributed system includes: a client, a storage node cluster, and a backup storage system; wherein the storage node cluster includes N storage nodes, and M LUNs that can be identified by the client are provided on the N storage nodes. N and M are integers, N≥2, and M≥1. The interaction diagram shown in FIG. 4 includes the following steps:

S401. The user sends a first trigger instruction to the client. The first trigger instruction instructs the storage node cluster to create snapshots with version numbers V-1 for the M LUNs, respectively.

S402. The client obtains a first trigger instruction.

S403. The client sends a first trigger instruction to the storage node cluster.

S404. The storage node cluster obtains a first trigger instruction, and creates a snapshot of version number V-1 for each of the M LUNs according to the first trigger instruction; the first trigger instruction instructs the storage node cluster to the M Create snapshots of version V-1 for each LUN.

S405. Any one of the N storage nodes in the storage node cluster creates an index corresponding to the snapshot version number V for each of the M LUNs.

. S406 user sends at least zero to the client user inputs a write command, the user writes a zero input at least any one of an instruction to write the user instruction input K comprising: a write data D _k and D _k of the data to be written An input position P _k , the write position P _k includes: a logical unit identifier LUN_idk ′ of the logical unit to be written to the data D _k and a storage block corresponding to the write position P _k in the to-be-written The offset Y _{k in the} logical unit; the idk ′ is an integer, and 1 ≦ idk ′ ≦ M.

S407. The client obtains the at least zero user write input instructions.

S408. The client sends the at least zero user write input instructions to the storage node cluster.

S409. Acquires the storage node clusters at least zero user inputs a write command, the storage node of the cluster write any user input command K, the writing position determining P _k P _k according to the write position located in the storage Node j, adding an entry to the index created by the storage node j for a LUN with a LUN identified as LUN_idk 'and a snapshot version number of the LUN of V, and recording the position _Pk in the added entry, so Let j be an integer, 1≤j≤N.

When it should be noted that the client obtains zero user write input instructions means that the client does not obtain user write input instructions. In this case, steps S408 and S409 are not performed, and no entry is added to the index.

S410. The user sends a second trigger instruction to the client. The second trigger instruction instructs the storage node cluster to create snapshots with version numbers V for the M LUNs, respectively.

S411. The client obtains a second trigger instruction.

S412. The client sends a second trigger instruction to the storage node cluster.

S413. The storage node cluster obtains the second trigger instruction, and creates snapshots with version numbers V for the M LUNs according to the second trigger instruction.

S414. The user sends a first instruction to the client.

The first instruction carries a LUN identifier LUN_id of a specified LUN and a snapshot version number V-1 and V of a snapshot created on the specified LUN; the first instruction instructs obtaining the snapshot version number of the specified LUN to be V-1 And a difference bitmap between two adjacent snapshots whose snapshot version number is V, the id and V are integers, V≥1, 1≤id≤M.

S415. The client obtains the first instruction.

S416. The client establishes a communication connection with each of the N storage nodes.

S417. The client searches for an index Liv corresponding to the specified LUN and the snapshot whose snapshot version number is V established by any storage node i of the N storage nodes; the index Liv records all The position information of each data updated between two adjacent snapshots is described, where i is an integer and 1 ≦ i ≦ N.

S418. The client determines the difference bitmap according to an index set {L ₁ v, L ₂ v, ... L _NV }.

S419. The storage node cluster sends the differential data and the position of each differential data in the snapshot of the designated LUN with a snapshot version number of V to the backup storage system.

In some possible implementation manners of the present invention, the position of the differential data and each differential data in the snapshot of the snapshot version number V of the specified LUN may be obtained by the storage node cluster from the client.

S420. The backup storage system uses the received differential data and the position of each differential data in the snapshot of the snapshot version number V of the specified LUN to update.

In the embodiment of the present invention, when obtaining the difference bitmap of a specified LUN between two adjacent snapshots, the difference bitmap data is collected at the granularity of the storage node, the storage node manages the index corresponding to the difference data according to the LUN granularity, and the client collects the difference bitmap Only need to establish N connections with N storage nodes. Compared with the prior art, which collects the difference bitmap with partition granularity, and the storage node manages the index corresponding to the difference data according to the partition granularity, the client and the storage node cluster in the embodiment of the present invention The number of communication connections is reduced, the efficiency of obtaining the difference bitmap is improved, and it is also beneficial for the backup storage system to perform timely backup of the data in the storage node cluster.

An embodiment of the present invention further provides a storage node in a distributed storage system, where the storage node is any one of N storage nodes included in the distributed storage system, and the N storage nodes are used for A logical unit number LUN provides storage space, and the storage node includes:

An embodiment of the present invention further provides a client in a distributed storage system, where the distributed storage system includes N storage nodes, and the N storage nodes are used to provide storage space for a first logical unit number LUN; The clients include:

An embodiment of the present invention also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, the method includes a method for managing snapshots in any of the distributed storage systems described in the foregoing method embodiments. Some or all steps.

An embodiment of the present invention also provides a computer storage medium, where the computer storage medium may store a program, and when the program is executed, the method includes generating a difference bitmap in any one of the distributed storage systems described in the foregoing method embodiments. Part or all of the steps of a method.

An embodiment of the present invention further provides an application program, where the application program is used to execute a method for managing snapshots in a distributed storage system according to an embodiment of the present invention at runtime.

An embodiment of the present invention further provides an application program, where the application program is used to execute a method for generating a difference bitmap in a distributed storage system according to an embodiment of the present invention at runtime.

The steps in the method of the embodiment of the present invention may be adjusted, combined, and deleted according to actual needs.

The units in the apparatus of the embodiment of the present invention may be combined, divided, and deleted according to actual needs.

A person of ordinary skill in the art may understand that all or part of the processes in the method of the foregoing embodiments are implemented. The processes may be completed by a computer program instructing related hardware. The program may be stored in a computer-readable storage medium. When the program is executed, Can include the processes of the method embodiments described above.

In the embodiment of the present invention, in an implementation, the storage node and the client may have a general computer structure, including a memory and a processor, the memory communicates with the processor, the memory includes computer instructions, and the processor executes the computer instructions in the memory, respectively. It is used to implement the functions of the storage node and the client in the embodiments of the present invention.

Each unit structure of the storage node and the client in the embodiment of the present invention may be implemented in hardware, such as a processor and a memory, or a software module, such as a computer instruction in the memory. The embodiment of the present invention does not limit this. Those of ordinary skill in the art may understand that all or part of the processes in the method of the foregoing embodiments are implemented. The processes may be completed by a computer program instructing related hardware. The program may be stored in a computer-readable storage medium. Can include the processes of the method embodiments described above. The foregoing storage medium includes: a read-only memory (ROM) or a random access memory (RAM), a magnetic disk, or an optical disk, which can store various program codes.

Claims

A method for managing snapshots in a distributed storage system, wherein the distributed storage system includes N storage nodes, and the N storage nodes are used to provide storage space for a first logical unit number LUN, the method include:

The storage node Ni generates a first record; the first record is only used to record an update address corresponding to the logical address of the first LUN distributed at the storage node Ni at a first snapshot time; wherein the first LUN distribution The logical address of the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is each value from 1 to N;

The storage node Ni generates a second record, and the second record is only used to record an updated address corresponding to a logical address where the first LUN is distributed on the storage node Ni at a second snapshot time; wherein the first The snapshot time and the second snapshot time are adjacent snapshot times.
The method according to claim 1, further comprising:

The storage node Ni provides the client with the first record and the second record.
The method according to claim 1 or 2, wherein the first record and the second record are both linked lists.
The method according to claim 1 or 2, wherein the first record and the second record are both arrays.
A method for generating a difference bitmap in a distributed storage system, wherein the distributed storage system includes N storage nodes, and the N storage nodes are used to provide storage space for a first logical unit number LUN;

The client obtains the first record and the second record from the storage node Ni; wherein the first record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the first snapshot time Wherein the logical address of the first LUN distributed on the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i is 1 to Each value of N; the second record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the second snapshot time; wherein the first snapshot time and The second snapshot moment is an adjacent snapshot moment;

The client generates a different position of the first LUN on the storage node Ni according to the first record and the second record.
The method according to claim 5, further comprising:

Sending, by the client, a first snapshot instruction to the storage node Ni at the first snapshot moment;

The client sends a second snapshot instruction to the storage node Ni at the second snapshot moment; the first snapshot instruction and the second snapshot instruction are respectively used to perform a snapshot operation on the first LUN.
The method according to claim 5 or 6, wherein the first record and the second record are both linked lists.
The method according to claim 5 or 6, wherein the first record and the second record are both arrays.
A storage node in a distributed storage system, wherein the storage node is any one of N storage nodes included in the distributed storage system, and the N storage nodes are used for a first logical unit. No. LUN provides storage space, and the storage node includes:

A first processing unit, configured to generate a first record; the first record is only used to record an update address corresponding to a logical address of the first LUN distributed on the storage node at a first snapshot time; A logical address of a LUN distributed on the storage node and a logical address of a second LUN distributed on the storage node belong to the same partition, and N is an integer not less than 2;

A second processing unit, configured to generate a second record, where the second record is only used to record an update address corresponding to a logical address where the first LUN is distributed on the storage node at a second snapshot moment; A snapshot time and the second snapshot time are adjacent snapshot times.
The storage node according to claim 9, wherein the storage node further comprises:

A providing unit is configured to provide the client with the first record and the second record.
The storage node according to claim 9 or 10, wherein the first record and the second record are both linked lists.
The storage node according to claim 9 or 10, wherein the first record and the second record are both an array.
A client in a distributed storage system, wherein the distributed storage system includes N storage nodes, and the N storage nodes are used to provide storage space for a first logical unit number LUN; the client include:

An obtaining unit, configured to obtain a first record and a second record from the storage node Ni; wherein the first record is only used to record a logical address corresponding to the first LUN distributed on the storage node Ni at a first snapshot time The updated address of the first LUN distributed on the storage node Ni and the logical address of the second LUN distributed on the storage node Ni belong to the same partition, N is an integer not less than 2, and i For each value from 1 to N; the second record is only used to record the update address corresponding to the logical address of the first LUN distributed on the storage node Ni at the second snapshot time; wherein the first record The snapshot time and the second snapshot time are adjacent snapshot times;

A generating unit is configured to generate a difference position of the first LUN on the storage node Ni according to the first record and the second record.
The client according to claim 13, wherein the client further comprises:

A sending unit, configured to send a first snapshot instruction to the storage node Ni at the first snapshot time, and send a second snapshot instruction to the storage node Ni at the second snapshot time; the first snapshot instruction And the second snapshot instruction are respectively used to perform a snapshot operation on the first LUN.
The client according to claim 13 or 14, wherein the first record and the second record are both linked lists.
The client according to claim 13 or 14, wherein the first record and the second record are both an array.
A storage node in a distributed storage system, wherein the storage node is any one of N storage nodes included in the distributed storage system, and the N storage nodes are used for a first logical unit. LUN provides storage space, the storage node includes a memory and a processor, the memory is in communication with the processor, the memory contains computer instructions, and the processor executes the computer instructions in the memory for The method according to any one of claims 1-4 is performed.
A storage node in a distributed storage system, characterized in that the distributed storage system includes N storage nodes, and the N storage nodes are used to provide storage space for a first logical unit number LUN; the client The inclusion point includes a memory and a processor, the memory is in communication with the processor, the memory includes computer instructions, and the processor executes the computer instructions in the memory for performing any of claims 5-8 The method described.