Method and device for realizing block storage enhanced data redundancy in distributed storage scene

Abstract

The invention relates to a method and a device for realizing block storage enhanced data redundancy in a distributed storage scene, and belongs to the technical field of computers, the method comprises the following steps: receiving a first data write-in request to RBD equipment, determining a first storage area of the RBD equipment, and determining a second storage area in local auxiliary equipment, constructing a second data write-in request to the local auxiliary equipment according to the first data write-in request; and executing a first data write-in operation to the RBD equipment and a second data write-in operation to the local auxiliary equipment, and determining the condition of the dirty region mark of the second storage region according to the results of the first data write-in operation and the second data write-in operation. According to the method, the local auxiliary equipment is used for carrying out redundant backup on the service data, so that when the service cannot normally access the distributed storage cluster, the continuity of the service can still be guaranteed, and the service is prevented from being interrupted due to unavailable storage.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and in particular to a method and device for implementing block storage enhanced data redundancy in a distributed storage scenario. BACKGROUND

[0002] With the rapid development of cloud computing field, more and more applications choose to use distributed storage as the backend storage; among them, the RBD (RADOS Block Device) of Ceph becomes the first choice of core business due to its excellent performance.

[0003] However, in actual use, network fluctuations, storage cluster abnormalities and other problems will affect the access of core business to storage. At present, the redundancy mechanism of traditional distributed storage for RBD device has two kinds: redundancy mechanism based on storage pool strategy and cross-cluster redundancy mechanism based on rbd-mirror.

[0004] 1. Redundancy mechanism based on storage pool strategy This mechanism copies data into multiple copies and stores them in each storage node, and checks and repairs the data of the storage copy periodically to ensure the reliability of the data. However, it has the following main problems: (1) When the storage cluster fails, the core business will be in an unusable state, and the business can continue to provide services only after the storage cluster is completely restored.

[0005] (2) The business data and its redundant data are stored in the same cluster, and when the storage cluster fails, there is a risk of data damage and loss.

[0006] (3) When data rebalancing is performed after the storage cluster fails to recover, a large number of additional data read and write operations will be generated, which will seriously affect the business performance.

[0007] 2. Cross-cluster redundancy mechanism based on rbd-mirror This mechanism creates an additional backup storage cluster and asynchronously replicates the data of the RBD device to the backup storage cluster to achieve data redundancy of RBD. However, it has the following main problems: (1) The core business may still be in an unusable state due to access to storage exceptions, for example: The master and backup cluster failure caused by major defects of distributed storage software.

[0008] The communication link between the core business node and the master and backup cluster has a hardware failure.

[0009] There are availability problems when the core business node accesses the distributed storage cluster.

[0010] (2) The data redundancy mechanism is asynchronous operation, and there is an inevitable delay window. The redundant data is not the latest data, and the data in the window period is easy to be lost.

[0011] The traditional redundancy mechanism has a major problem: when the core business cannot normally access the distributed storage cluster, the continuity of the core business cannot be guaranteed, resulting in business interruption due to storage unavailability. SUMMARY

[0012] In view of the above analysis, the embodiments of the present application aim to provide an implementation method and device for enhancing data redundancy of block storage in a distributed storage scenario, to solve the problem that the continuity of the core business cannot be guaranteed when the core business cannot normally access the distributed storage cluster, resulting in business interruption due to storage unavailability.

[0013] On the one hand, the embodiments of the present application provide an implementation method for enhancing data redundancy of block storage in a distributed storage scenario, which comprises: receiving a first data write request to an RBD device; wherein the first data write request comprises business data to be written and index data of the business data, and the index data comprises a first data write address of the business data to the RBD device and information of the data length of the business data; determining at least one first storage area of the RBD device for storing the business data according to the index data; determining a second data write address in a local auxiliary device according to the first data write address, and determining a second storage area in the local auxiliary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area; constructing a second data write request for writing the business data to the second data write address of the local auxiliary device according to the first data write request; respectively executing a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determining the dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0014] The beneficial effects of the above technical solution are as follows: by using the local auxiliary device to make redundant backup of business data, when the business cannot normally access the distributed storage cluster, the continuity of the business can still be guaranteed, and the business interruption due to storage unavailability is avoided.

[0015] Based on the further improvement of the above method, the determination of the dirty region mark of the second storage area according to the results of the first data write operation and the second data write operation comprises: in response to the success of the first data write operation and the second data write operation, determining that the dirty region mark is not set in the second storage area, to represent the data consistency of the first storage area and the second storage area; in response to the failure of any one of the first data write operation and the second data write operation, determining that the dirty region mark is set in the second storage area for subsequent data synchronization.

[0016] The beneficial effect of the above further improvement is that the data synchronization of the first storage area and the second storage area can be quickly determined according to the dirty region mark of the second storage area.

[0017] Based on the further improvement of the above method, the method further comprises: acquiring the second storage area in the local auxiliary device with the dirty region mark by a data synchronization thread; performing data synchronization between the second storage area with the dirty region mark and the corresponding first storage area in the RBD device; and in response to the success of the data synchronization, clearing the dirty region mark of the second storage area.

[0018] The beneficial effect of the above further improvement is that the second storage area marked as a dirty region is acquired by a data synchronization thread, the data synchronization of the second storage area and the first storage area is performed, and the dirty region mark is cleared after the success of the data synchronization, which guarantees the data consistency of the first storage area and the second storage area, improves the reliability of data enhanced storage, and guarantees the strict correspondence between the dirty region mark and the data consistency.

[0019] Based on the further improvement of the above method, the execution of the data synchronization between the second storage area with the dirty region mark and the corresponding first storage area in the RBD device comprises: in response to the success of the first data write operation and the failure of the second data write operation, copying the data of the first storage area to the second storage area; and in response to the failure of the first data write operation and the success of the second data write operation, copying the data of the second storage area to the first storage area.

[0020] The beneficial effect of the above further improvement is that the reliability of the data synchronization of the first storage area and the second storage area is improved.

[0021] Based on the further improvement of the above method, the method applied to the service node further comprises: mapping the RBD device to the service node before receiving the first data write request to the RBD device; allocating the local auxiliary device for the RBD device, establishing the mapping relationship between the addresses of the RBD device and the local auxiliary device, and establishing the mapping relationship between the first storage area of the RBD device and the second storage area of the local auxiliary device; defaulting all areas of the second storage area as dirty areas, performing data synchronization between the second storage area with the dirty area mark and the corresponding first storage area in the RBD device based on the data in the RBD device, and clearing the dirty area mark after successful data synchronization.

[0022] The beneficial effect of the above further improvement scheme is that the initialization synchronization ensures the consistency of the data of the RBD device and the local auxiliary device, and provides a basis for the subsequent data redundancy function.

[0023] Based on the further improvement of the above method, before establishing the mapping relationship between the first storage area of the RBD device and the second storage area of the local auxiliary device, the method further comprises: dividing the RBD device into a plurality of first storage areas according to a preset space size, and dividing the local auxiliary device into a plurality of second storage areas according to the preset space size; wherein the preset space size is greater than the original block size of the RBD device.

[0024] The beneficial effect of the above further improvement scheme is that the convenience of management of the first storage area and the second storage area is improved.

[0025] Based on the further improvement of the above method, the method further comprises: if a failure scenario occurs in the first data write operation to the RBD device, the second data write operation to the local auxiliary device, and the data synchronization between the RBD device and the local auxiliary device, obtaining failure information and recording it into a log, and alarming according to a preset alarm strategy; the failure information comprises at least one of RBD device information, local auxiliary device information, write address information of failed write operation, and write business data length.

[0026] The beneficial effect of the above further improvement scheme is that when the RBD device or the local auxiliary device fails, the user will not be aware of the interruption of the service, and the fault cannot be solved in time, leaving hidden dangers. After the error alarm, the user or the operation and maintenance personnel can be notified in time to perform related error processing, and the fault is further expanded.

[0027] Based on the further improvement of the above method, after the respective execution of the first data write operation corresponding to the first data write request and the second data write operation corresponding to the second data write request, the method further comprises: receiving a data read request for reading the business data; in response to the distributed storage cluster corresponding to the RBD device being faulty, reading the business data by accessing the second storage area of the local auxiliary device; in response to the distributed storage cluster corresponding to the RBD device being normal, and the second storage area currently being provided with the dirty zone mark, in the case that the first data write operation is successful, reading the business data by accessing the first storage area, and in the case that the second data write operation is successful, reading the business data by accessing the second storage area; in response to the distributed storage cluster corresponding to the RBD device being normal, and the second storage area currently not being provided with the dirty zone mark, reading the business data by accessing the second storage area.

[0028] The beneficial effect of the above further improvement is to ensure the reliability of data reading and improve the efficiency of data reading.

[0029] Based on the further improvement of the above method, the method further comprises: pooling and uniformly managing one or more local redundant disks by a user mode tool to obtain one or more local storage pools; and dividing each of the local storage pools to obtain one or more local auxiliary devices.

[0030] The beneficial effect of the above further improvement is to realize the uniform management of local redundant disks and easy expansion.

[0031] In another aspect, an embodiment of the present application provides an implementation device for enhancing data redundancy of block storage in a distributed storage scenario, comprising: a receiving module configured to receive a first data write request to an RBD device; wherein the first data write request comprises service data to be written and index data of the service data, and the index data comprises a first data write address of the RBD device for writing the service data and information of a data length of the service data; a determining module configured to determine at least one first storage area of the RBD device for storing the service data according to the index data; a mapping module configured to determine a second data write address in a local secondary device according to the first data write address, and determine a second storage area in the local secondary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local secondary device, and there is a mapping relationship between the first storage area and the second storage area; a constructing module configured to construct a second data write request of writing the service data to the second data write address of the local secondary device according to the first data write request; wherein the second data write request comprises the second data write address of the local secondary device and the service data; and a writing module configured to respectively execute a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determine a dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0032] The above technical solution has the following beneficial effects: by using the local secondary device to make a redundant backup of the service data, when the service cannot normally access the distributed storage cluster, the continuity of the service can still be guaranteed, and the service is prevented from being interrupted due to unavailable storage.

[0033] The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the implementation method for enhancing data redundancy of block storage in a distributed storage scenario as described above.

[0034] The present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the implementation method for enhancing data redundancy of block storage in a distributed storage scenario as described above.

[0035] The present application also provides a computer program product comprising a computer program, wherein the computer program is executable by a processor to implement the implementation method for enhancing data redundancy of block storage in a distributed storage scenario as described above.

[0036] The technical solutions in the present application can be combined with each other to realize more preferred combination solutions. Other features and advantages of the present application will be described in the subsequent description, and some advantages will become apparent from the description, or will be understood by implementing the present application. The purposes and other advantages of the present application can be realized and obtained from the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0037] The accompanying drawings are included to provide a further understanding of the present application, and are incorporated in and constitute a part of this application, illustrate embodiments of the present application, and together with the description serve to explain the principles of the present application. Figure 1 is a flow diagram of an implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 2 is an implementation schematic diagram of a local auxiliary device in the implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 3 is a system architecture diagram of the implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 4 is a data writing flow diagram in the implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 5 is a data reading flow diagram in the implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 6 is a structure schematic diagram of an implementation device of block storage enhanced data redundancy in a distributed storage scenario provided by the present application; Figure 7 An example of an electronic device is shown in the physical structure schematic diagram. DETAILED DESCRIPTION

[0038] The preferred embodiments of the present application will be specifically described below in conjunction with the drawings, wherein the drawings constitute a part of the present application, and are used to illustrate the principles of the embodiments of the present application, and are not used to limit the scope of the present application.

[0039] Term explanation: Ceph: a unified distributed storage system with high performance, high reliability and high scalability. It is widely used in cloud platforms, big data and other scenarios.

[0040] RBD: a remote block storage device provided by Ceph, which can be mounted on a client node through a network and perform data read / write operations.

[0041] Figure 1It is the flow diagram of the implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application. As shown in Figure 1 The method comprises the following steps. Step S1, receiving a first data write request to an RBD device; wherein the first data write request comprises service data to be written and index data of the service data, and the index data comprises a first data write address of the service data written to the RBD device and information of the data length of the service data.

[0042] The implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application can be applied to a server, and the module implementing the method of the application can be abstracted as a functional module on the server. The RBD device is a provided remote block storage device, which is mounted on the server and performs data read and write operations. The first data write request to the RBD device can be sent by an external device or by an internal module of the server, such as a functional module of a user layer of the server.

[0043] The first data write request comprises service data to be written and index data of the service data, and the index data comprises a first data write address of the service data written to the RBD device and information of the data length of the service data.

[0044] Step S2, determining at least one first storage area of the RBD device for storing the service data according to the index data.

[0045] The first storage area is determined according to the first data write address and the data length. The first storage area is used for storing the service data to be written.

[0046] The first data write address and the data length are referred to as index data, and the index data is used to determine the first storage area.

[0047] Step S3, determining a second data write address in a local auxiliary device according to the first data write address, and determining a second storage area in the local auxiliary device according to the first storage area; wherein the local auxiliary device is a local block storage device, there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area.

[0048] The local auxiliary device is a block storage device based on a local redundant device. There is a mapping relationship between the addresses of the RBD device and the local auxiliary device. Based on the mapping relationship between the addresses of the RBD device and the local auxiliary device, the second data write address in the local auxiliary device is determined according to the first data write address, and the second storage area in the local auxiliary device is determined according to the first storage area. That is, the first storage area of the RBD device and the second storage area in the local auxiliary device also have a mapping relationship.

[0049] Step S4, constructing a second data write request for writing the service data to the second data write address of the local auxiliary device according to the first data write request; wherein the second data write request comprises the second data write address of the local auxiliary device and the service data.

[0050] According to the first data write request, a second data write request for the local auxiliary device is constructed, and the second data write request comprises the second data write address of the local auxiliary device and the service data, and is used to request to write the service data to the second data write address of the local auxiliary device.

[0051] Step S5, respectively performing a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determining the dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation; wherein the first data write operation is used to write the service data to the RBD device according to the first data write address, and the second data write operation is used to write the service data to the local auxiliary device according to the second data write address.

[0052] Respectively performing a first data write operation and a second data write operation, wherein the first data write operation is used to write the service data to the RBD device according to the first data write address, and the second data write operation is used to write the service data to the local auxiliary device according to the second data write address. The dirty area mark of the second storage area is determined according to the results of the first data write operation and the second data write operation. According to the dirty area mark of the second storage area, it can be known whether the first data write operation and the second data write operation are both successful. If both are successful, the data of the first storage area and the second storage area is consistent, and subsequent data synchronization is not needed. Otherwise, subsequent data synchronization is needed. Thus, the data of the local auxiliary device and the RBD device is finally consistent, so as to reliably realize redundant backup.

[0053] In the case of determining the dirty region marker of the second storage region according to the results of the first data write operation and the second data write operation, whether to set the dirty region marker can be determined according to the results of the first data write operation and the second data write operation. In addition, if the second storage region is previously provided with the dirty region marker, whether to clear or maintain the dirty region marker can be determined according to the results of the first data write operation and the second data write operation.

[0054] The distributed storage scenario block storage enhanced data redundancy implementation method provided by the application can guarantee the continuity of the service and avoid service interruption due to storage unavailability by using the local auxiliary device to make a redundant backup of the service data when the service cannot normally access the distributed storage cluster.

[0055] According to the distributed storage scenario block storage enhanced data redundancy implementation method provided by the application, the case of determining the dirty region marker of the second storage region according to the results of the first data write operation and the second data write operation includes: in response to the success of the first data write operation and the second data write operation, determining that the dirty region marker is not set in the second storage region, to represent that the data of the first storage region and the second storage region is consistent; and in response to the failure of any one of the first data write operation and the second data write operation, determining that the dirty region marker is set in the second storage region, to be used for subsequent data synchronization.

[0056] In the case of the success of the first data write operation to the RBD device and the second data write operation to the local auxiliary device, it is determined that the dirty region marker is not set in the second storage region, that is, the second storage region is not marked as a dirty region, to represent that the data of the first storage region and the second storage region is consistent.

[0057] In the case of the failure of any one of the first data write operation to the RBD device and the second data write operation to the local auxiliary device, it is determined that the dirty region marker is set in the second storage region, that is, the second storage region is marked as a dirty region, to represent that the data of the first storage region and the second storage region needs to be synchronized subsequently.

[0058] The distributed storage scenario block storage enhanced data redundancy implementation method provided by the application can represent that the data of the first storage region and the second storage region needs to be synchronized by marking the second storage region as a dirty region, and represent that the data of the first storage region and the second storage region is consistent by not marking the second storage region as a dirty region, so that whether the data of the first storage region and the second storage region needs to be synchronized can be quickly obtained according to the dirty region marker of the second storage region.

[0059] According to the implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application, the method further comprises: obtaining, by a data synchronization thread, the second storage area in the local auxiliary device provided with the dirty area mark; performing data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device; and in response to successful data synchronization, clearing the dirty area mark of the second storage area.

[0060] The data synchronization thread is configured to achieve consistent data storage of the RBD device and the local auxiliary device. Periodic checking can be performed by the data synchronization thread to obtain the second storage area in the local auxiliary device provided with the dirty area mark, that is, the second storage area whose data is inconsistent with the corresponding first storage area, and data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device is performed. After successful data synchronization, the dirty area mark of the second storage area is cleared, indicating that the data of the first storage area and the second storage area becomes consistent.

[0061] The implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application uses a data synchronization thread to obtain the second storage area marked as a dirty area, performs data synchronization between the second storage area and the first storage area, and clears the dirty area mark after successful data synchronization, thereby ensuring the data consistency of the first storage area and the second storage area, improving the reliability of data enhanced storage, and ensuring strict correspondence between the dirty area mark and data consistency.

[0062] According to the implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application, the data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device comprises: in response to successful first data write operation and failed second data write operation, copying the data of the first storage area to the second storage area; and in response to failed first data write operation and successful second data write operation, copying the data of the second storage area to the first storage area.

[0063] When performing data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device, the data needs to be copied from the party with successful write to the other party. Therefore, if the first data write operation is successful and the second data write operation fails, the data of the first storage area is copied to the second storage area; and if the first data write operation fails and the second data write operation is successful, the data of the second storage area is copied to the first storage area.

[0064] The distributed storage scene block storage enhanced data redundancy implementation method provided by the application improves the reliability of data synchronization of the first storage area and the second storage area.

[0065] The distributed storage scene block storage enhanced data redundancy implementation method provided by the application is applied to a business node, and before receiving a first data write request to an RBD device, the method further comprises: mapping the RBD device to the business node; allocating the local auxiliary device for the RBD device, establishing a mapping relationship between the addresses of the RBD device and the local auxiliary device, and establishing a mapping relationship between the first storage area of the RBD device and the second storage area of the local auxiliary device; defaulting all areas of the second storage area as dirty areas, based on the data in the RBD device, performing data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device, and clearing the dirty area mark after successful data synchronization.

[0066] The RBD device can be mapped to the business node by using a tool, the local auxiliary device is allocated for the RBD device, and the mapping of the RBD device and the local auxiliary device is established. All areas of the second storage area are marked as dirty areas by default, and during initialization, the data in the RBD device is used as a reference to perform data synchronization between the second storage area provided with the dirty area mark and the corresponding first storage area in the RBD device, and the dirty area mark is cleared after successful data synchronization. This function can be skipped through parameters to prevent unnecessary data synchronization operations.

[0067] The distributed storage scene block storage enhanced data redundancy implementation method provided by the application ensures that the data of the RBD device and the local auxiliary device is consistent through initialization synchronization, and provides an implementation basis for subsequent data redundancy functions.

[0068] According to the distributed storage scene block storage enhanced data redundancy implementation method provided by the application, before establishing the mapping relationship between the first storage area of the RBD device and the second storage area of the local auxiliary device, the method further comprises: dividing the RBD device into a plurality of first storage areas according to a preset space size, and dividing the local auxiliary device into a plurality of second storage areas according to the preset space size; wherein the preset space size is greater than the original block size of the RBD device.

[0069] For the convenience of management, the RBD device can be divided into a plurality of first storage areas according to a preset space size, and the local auxiliary device can be divided into a plurality of second storage areas, and a mapping relationship between the first storage areas and the second storage areas is established. Since the business data is usually stored in multiple blocks in succession, setting the preset space size to be greater than the original block size of the RBD device can facilitate efficient management of the first storage areas and the second storage areas.

[0070] The implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application improves the convenience of management of the first storage areas and the second storage areas.

[0071] According to the implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application, if a failure scenario occurs when the first data write operation to the RBD device, the second data write operation to the local auxiliary device, and data synchronization between the RBD device and the local auxiliary device are performed, failure information is obtained and recorded in a log, and an alarm is given according to a preset alarm strategy; the failure information includes at least one of RBD device information, local auxiliary device information, write address information of the failed write operation, and write business data length.

[0072] If a failure scenario occurs when the first data write operation to the RBD device and the second data write operation to the local auxiliary device are performed, and data synchronization between the RBD device and the local auxiliary device is performed according to the dirty zone mark, the failure information is recorded in the log, and the recorded information includes RBD device information, local auxiliary device information, write address information of the failed write operation, write business data length, etc. The error information is output according to an alarm strategy with gradually increasing alarm intervals, and the failure alarm time of the same storage area can be 1 min, 10 min after the failure, and then every hour. The setting of the alarm interval can prevent the same error information from constantly occupying system resources.

[0073] The implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application can timely notify the user or the operation and maintenance personnel to perform related error processing after the error alarm, so as to avoid further expansion of the fault.

[0074] According to the implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application, after the first data write operation corresponding to the first data write request and the second data write operation corresponding to the second data write request are respectively performed, the method further comprises: receiving a data read request for reading the business data; in response to a failure of the distributed storage cluster corresponding to the RBD device, reading the business data by accessing the second storage area of the local auxiliary device; in response to normality of the distributed storage cluster corresponding to the RBD device and the second storage area currently being provided with the dirty area mark, reading the business data by accessing the first storage area in the case that the first data write operation is successful, and reading the business data by accessing the second storage area in the case that the second data write operation is successful; and in response to normality of the distributed storage cluster corresponding to the RBD device and the second storage area currently not being provided with the dirty area mark, reading the business data by accessing the second storage area.

[0075] After receiving the data read request for reading the business data, if the distributed storage cluster corresponding to the RBD device fails, the business data is read by accessing the second storage area of the local auxiliary device because the probability of failure of the local auxiliary device is extremely low; if the distributed storage cluster corresponding to the RBD device is normal, it is judged whether the second storage area is a dirty area, if the second storage area is a dirty area, that is, provided with the dirty area mark, it indicates that the data of the first storage area and the second storage area is inconsistent; if the first data write operation is successful, it indicates that the RBD device successfully writes data, and the first storage area of the RBD device is accessed to read the business data; if the second data write operation is successful, it indicates that the local auxiliary device successfully writes data, and the second storage area of the local auxiliary device is accessed to read the business data; if the distributed storage cluster corresponding to the RBD device is normal and the second storage area is currently not provided with the dirty area mark, it indicates that the data of the RBD device and the local auxiliary device is consistent, and in order to improve the data read efficiency, the business data is read by accessing the second storage area.

[0076] The implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application guarantees the reliability of data reading and improves the data read efficiency.

[0077] According to the implementation method for enhancing data redundancy of block storage in a distributed storage scenario provided by the application, the method further comprises: pooling and uniformly managing one or more local redundant disks by a user state tool to obtain one or more local storage pools; and dividing each local storage pool to obtain one or more local auxiliary devices.

[0078] Figure 2It is the implementation principle diagram of the local auxiliary device in the implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application. Figure 2 As shown in the figure, one or more local redundant disks in the system can be pooled and uniformly managed by a user mode tool, that is, the local redundant disks are abstracted and aggregated into a logically distributed storage pool.

[0079] The implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application realizes the unified management and easy expansion of the local redundant disks by pooling and uniformly managing one or more local redundant disks, obtaining one or more local storage pools, and dividing each local storage pool into one or more local auxiliary devices.

[0080] According to the implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application, the method is implemented in the kernel layer.

[0081] Figure 3 It is the system architecture diagram of the implementation method of the block storage enhanced data redundancy in the distributed storage scene provided by the application. Figure 3 As shown in the figure, the application adopts structured design and decouples each function. User layer: the user layer tool realizes function control, including enabling function, disabling function, mapping / cancelling mapping RBD device with enhanced redundancy function, viewing data synchronization state between RBD device and local auxiliary device and local disk management and other functions.

[0082] Kernel layer: the kernel layer is the core part of the application, which completes the IO mirroring function of RBD device and auxiliary device in the RBD enhanced redundancy module, and realizes the strong consistency function of local data and RBD data.

[0083] Hardware layer: the data bearing entity composed of RBD backend storage and auxiliary device.

[0084] The application implements the implementation method of the block storage enhanced data redundancy in the distributed storage scene in the kernel layer, and the user layer is not aware and has high execution efficiency.

[0085] The application provides a block storage enhanced data redundancy implementation method in a distributed storage scenario, which improves execution efficiency and user experience by implementing enhanced data redundancy in the kernel layer and providing a user state management tool.

[0086] The application provides a block storage enhanced data redundancy implementation method in a distributed storage scenario, which has the following advantages: 1. Reducing business storage path dependency: When the core business cannot normally access the distributed storage cluster, the back-end storage of the business will be downgraded to local storage, avoiding the interruption of the core business due to unavailable storage.

[0087] 2. Guaranteeing the availability of redundant data: The carrier of redundant data is a local physical device, independent of the storage cluster. When the storage cluster experiences catastrophic failure and causes business data damage or loss, the local redundant data is still available.

[0088] 3. Guaranteeing the validity of redundant data: Real-time data synchronization is achieved based on local IO replication, there is no asynchronous delay window, and the redundant data is guaranteed to be the latest data.

[0089] 4. Providing operation and maintenance buffer: When the distributed storage cluster recovers, the local storage is used as the back-end storage of the business. This saves valuable time for the recovery of the storage cluster and avoids affecting business performance due to balancing data.

[0090] To further illustrate the block storage enhanced data redundancy implementation method in a distributed storage scenario provided by the application, two examples are given below to introduce.

[0091] Figure 4 is a data write flowchart in the block storage enhanced data redundancy implementation method in a distributed storage scenario provided by the application. As Figure 4 shown, the method comprises: I. Normal data write flow 1. As shown in Figure 4 ①, when the core business writes data, the write request is listened to and copied. Among them, copying the write request means constructing a write request to the local auxiliary device according to the write request to the RBD device, the data in the write request to the RBD device is the original data, and the data in the write request to the local auxiliary device is the mirror data.

[0092] 2. As shown in Figure 4 ②, according to the address information of the write request and the data length of the business data to be written, the index data is obtained, and the first storage area corresponding to the index data is calculated, as shown in Figure 4 ③, according to the first storage area, the corresponding second storage area in the local auxiliary device is determined, and the area is marked as a dirty area.

[0093] 3. As shown in "④" of FIG. 3, the original data is written to the first storage area of the RBD device through the network, thereby realizing writing the original data to the distributed storage cluster. Figure 4 Figure 4 4. As shown in "⑤" of FIG. 3, the copied data is written to the second storage area of the local auxiliary device.

[0094] 4. As shown in "⑥" of FIG. 3, after the data writing to the RBD and the local auxiliary device is successful, the dirty area mark of the corresponding second storage area in the index data is cleared. Figure 4

[0095] II. Abnormal data writing process 1. As shown in "①" of FIG. 4, when the core business writes data, the writing request is listened to and copied. The copying of the writing request refers to constructing the writing request to the local auxiliary device according to the writing request to the RBD device. The data in the writing request to the RBD device is the original data, and the data in the writing request to the local auxiliary device is the mirror data. Figure 4 2. As shown in "②" of FIG. 4, according to the address information of the writing request and the data length of the business data to be written, the index data is obtained, and the first storage area corresponding to the index data is calculated, as shown in "③" of FIG. 4.

[0096] Figure 4 3. As shown in "④" of FIG. 4, the original data is written to the first storage area of the RBD device through the network, thereby realizing writing the original data to the distributed storage cluster. Figure 4 4. As shown in "⑤" of FIG. 4, the copied data is written to the second storage area of the local auxiliary device.

[0097] Figure 4 4. When the writing to any back-end storage including the RBD device and the local auxiliary device fails, the second storage area corresponding to the index data will remain as a dirty area, and will be rewritten by an independent synchronization thread in the background asynchronously. Figure 4 5. As shown in "⑦" of FIG. 4, the data synchronization thread periodically acquires the second storage area marked as "dirty area".

[0098] 6. As shown in "⑧" of FIG. 4, the data in the second storage area is written to the first storage area of the RBD device through the network, thereby realizing writing the data to the distributed storage cluster.

[0099] Figure 4 6. As shown in "⑧" of FIG. 4, the data in the second storage area is written to the first storage area of the RBD device through the network, thereby realizing writing the data to the distributed storage cluster.

[0100] 6. As shown in "⑧" of FIG. 4, the data in the second storage area is written to the first storage area of the RBD device through the network, thereby realizing writing the data to the distributed storage cluster. Figure 4 ​​​​​As shown in "⑧", the dirty area of the dirty area is constructed IO request, and the data of the first storage area is copied from the RBD device to the second storage area of the local auxiliary device when the storage cluster is normal; when the storage cluster recovers from the abnormal state, the data of the second storage area is copied from the local auxiliary device to the first storage area of the RBD device.

[0101] 7. As Figure 5 As shown in "⑨", after completing the background data synchronization, the dirty area mark of the corresponding second storage area in the index data is cleared.

[0102] Figure 5 The figure is a data reading process schematic diagram of the implementation method of the block storage enhanced data redundancy in the distributed storage scenario provided by the application. As Figure 6 The method comprises the following steps: When the core business reads the business data, if the distributed storage cluster fails, the local auxiliary device can be directly accessed for data reading. Otherwise, the RBD enhanced redundancy kernel module will select different reading paths according to whether the second storage area corresponding to the index data is dirty or not. The index data is constructed according to the reading address and data length of the business data, and the reading address of the business data is consistent with the writing address of the business data: When it is detected that the second storage area corresponding to the index data is a non-dirty area, the local auxiliary device is directly read through the fast path.

[0103] When it is detected that the second storage area corresponding to the index data is a dirty area, the writing state of the business data in the first storage area and the second storage area is obtained, if the writing state of the first storage area is successful, the slow path is degraded to directly read the distributed storage RBD; if the writing state of the second storage area is successful, the local auxiliary device is directly read through the fast path.

[0104] The implementation method of the block storage enhanced data redundancy in the distributed storage scenario provided by the application has the following technical advantages: 1. Hardware reliability of local storage: based on the local redundancy device, multiple disks are combined to form a redundant disk, and the fault tolerance ability of the redundant disk is guaranteed at the hardware level.

[0105] 2. Local storage pooling: multiple redundant disks are combined to realize unified management and address linear arrangement of local disks, and the mapping of RBD and local storage pool area is realized by marking the area in the local storage pool corresponding to RBD. Based on the above implementation, the dynamic expansion support in the running process of the local storage pool is convenient for users to carry out seamless hardware addition and project upgrading in the future.

[0106] 3. Write IO mirror, index data and independent synchronization thread: writing IO mirror ensures the timeliness of data writing, and based on index data, the region of failed write IO is marked, and finally through independent synchronization thread, background asynchronous rewriting is carried out to ensure stable data writing.

[0107] 4. Fusion of local storage and distributed storage based on index data: based on index data, data region division and synchronization of divided data are realized, and local dirty marking and small range synchronization of data are realized.

[0108] 1) In the abnormal scene of distributed storage short-time failure, shutdown maintenance, storage network temporary disconnection, etc., the abnormal data region is recorded, small range data synchronization after storage cluster recovery is realized, the recovery speed is accelerated, and the business pressure of distributed storage cluster in the recovery process is reduced.

[0109] 2) In the scene of long-time failure or even replacement and reconstruction of distributed storage, at this time, all regions of index data are dirty regions, full disk synchronization to distributed storage is carried out, and the key technical problem of business interruption after failure of the previous distributed storage cluster is solved.

[0110] 5. Data read path optimization: 1) When the distributed storage cluster is abnormal, the local disk is directly read to ensure that the business will not be interrupted due to the failure to read data.

[0111] 2) When the distributed storage cluster is normal, if the index data of the read data region is not dirty, the local disk is directly read through fast path to realize high-performance optimized data reading; if the index data of the region is dirty, the data read path is determined according to the data writing state before RBD and local auxiliary equipment to ensure the reliability of reading data.

[0112] The distributed storage scenario block storage data redundancy enhancement implementation device provided by the application is described below. The distributed storage scenario block storage data redundancy enhancement implementation device described below can be correspondingly referred to the distributed storage scenario block storage data redundancy enhancement implementation method described above.

[0113] Figure 6 is a structural schematic diagram of the distributed storage scenario block storage data redundancy enhancement implementation device provided by the application. As Figure 7As shown, the device comprises a receiving module 10, a determining module 20, a mapping module 30, a constructing module 40 and a writing module 50, wherein: the receiving module 10 is configured to receive a first data write request to an RBD device; wherein the first data write request comprises service data to be written and index data of the service data, and the index data comprises a first data write address of the service data to the RBD device and information of a data length of the service data; the determining module 20 is configured to determine at least one first storage area of the RBD device for storing the service data according to the index data; the mapping module 30 is configured to determine a second data write address in a local secondary device according to the first data write address, and determine a second storage area in the local secondary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local secondary device, and there is a mapping relationship between the first storage area and the second storage area; the constructing module 40 is configured to construct a second data write request of writing the service data to the second data write address of the local secondary device according to the first data write request; and the writing module 50 is configured to respectively execute a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determine a dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0114] The distributed storage scenario block storage enhances the implementation device of data redundancy, which backs up the service data by using the local secondary device, and guarantees the continuity of the service when the service cannot normally access the distributed storage cluster, and avoids the interruption of the service due to the unavailability of the storage.

[0115] Figure 7 An example of an electronic device is shown in the entity structure diagram, such as ​As shown, the electronic device can include a processor 710, a communication interface 720, a memory 730, and a communication bus 740, wherein the processor 710, the communication interface 720, and the memory 730 complete mutual communication through the communication bus 740. The processor 710 can invoke the logical instructions in the memory 730 to execute the implementation method of block storage enhancement data redundancy in a distributed storage scenario, which includes: receiving a first data write request to an RBD device; wherein the first data write request includes service data to be written and index data of the service data, and the index data includes a first data write address of the service data to the RBD device and information of a data length of the service data; determining at least one first storage area of the RBD device for storing the service data according to the index data; determining a second data write address in a local auxiliary device according to the first data write address, and determining a second storage area in the local auxiliary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area; constructing a second data write request of writing the service data to the second data write address of the local auxiliary device according to the first data write request; respectively executing a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determining the dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0116] In addition, the logical instructions in the memory 730 described above can be implemented in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0117] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the implementation method of block storage enhanced data redundancy in a distributed storage scenario provided by the above-mentioned methods, the method comprising: receiving a first data write request to an RBD device; wherein the first data write request comprises service data to be written and index data of the service data, the index data comprising a first data write address of the RBD device for writing the service data and information of a data length of the service data; determining at least one first storage area of the RBD device for storing the service data according to the index data; determining a second data write address in a local auxiliary device according to the first data write address, and determining a second storage area in the local auxiliary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area; constructing a second data write request of writing the service data to the second data write address of the local auxiliary device according to the first data write request; respectively executing a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determining a dirty area mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0118] In yet another aspect, the present application also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method for implementing block storage enhanced data redundancy in a distributed storage scenario as described above, the method comprising: receiving a first data write request to an RBD device; wherein the first data write request is to write service data and index data of the service data to the RBD device, the index data comprising a first data write address to write the service data to the RBD device and information of a data length of the service data; determining at least one first storage area of the RBD device for storing the service data according to the index data; determining a second data write address in a local secondary device according to the first data write address, and determining a second storage area in the local secondary device according to the first storage area; wherein there is a mapping relationship between the addresses of the RBD device and the local secondary device, and there is a mapping relationship between the first storage area and the second storage area; constructing a second data write request to write the service data to the second data write address in the local secondary device according to the first data write request; respectively executing a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determining a dirty region mark of the second storage area according to the results of the first data write operation and the second data write operation.

[0119] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the program can be stored in a computer readable storage medium. The computer readable storage medium includes a disk, an optical disk, a read-only memory, a random access memory, etc.

[0120] The device embodiments described above are merely illustrative, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0121] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software plus necessary universal hardware platforms, and of course can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a plurality of instructions to cause a computer device to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

[0123] The above is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited to this, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. An implementation method of block storage enhanced data redundancy in a distributed storage scenario, characterized in that, include: Receive a first data write request to the RBD device; wherein the first data write request includes service data to be written and index data of the service data, and the index data includes information on the first data write address to write the service data to the RBD device and the data length of the service data; Based on the index data, at least one first storage area of ​​the RBD device is determined for storing the business data; The second data write address in the local auxiliary device is determined based on the first data write address, and the second storage area in the local auxiliary device is determined based on the first storage area; wherein, there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area; Based on the first data write request, a second data write request is constructed to write the business data to the second data write address of the local auxiliary device; Execute a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determine the dirty area marking status of the second storage area based on the results of the first data write operation and the second data write operation.

2. The method for enhancing data redundancy in block storage under a distributed storage scenario according to claim 1, characterized in that, The process of determining the dirty zone marker of the second storage area based on the results of the first data write operation and the second data write operation includes: In response to the success of both the first data write operation and the second data write operation, it is determined that the dirty area marker is not set in the second storage area to indicate that the data in the first storage area and the second storage area are consistent; In response to the failure of either the first data write operation or the second data write operation, it is determined that the dirty area marker is set in the second storage area for subsequent data synchronization.

3. The method for enhancing data redundancy in block storage under a distributed storage scenario as described in claim 2, characterized in that, The method further includes: The second storage area marked with the dirty area is obtained from the local auxiliary device through the data synchronization thread; Perform data synchronization between the second storage area marked with the dirty area and the corresponding first storage area in the RBD device; In response to the successful data synchronization, the dirty area marker in the second storage area is cleared.

4. The method for enhancing data redundancy in block storage under a distributed storage scenario as described in claim 3, characterized in that, The process of performing data synchronization between the second storage area marked with the dirty area and the corresponding first storage area in the RBD device includes: In response to the success of the first data write operation and the failure of the second data write operation, the data in the first storage area is copied to the second storage area; In response to the failure of the first data write operation and the success of the second data write operation, the data in the second storage area is copied to the first storage area.

5. The method for enhancing data redundancy in block storage under a distributed storage scenario according to claim 3, characterized in that, The method is applied to a service node, and before receiving the first data write request to the RBD device, the method further includes: Map the RBD device to the service node; The local auxiliary device is allocated to the RBD device, a mapping relationship is established between the addresses of the RBD device and the local auxiliary device, and a mapping relationship is established between the first storage area of ​​the RBD device and the second storage area of ​​the local auxiliary device; By default, the entire area of ​​the second storage region is marked as a dirty region. Based on the data in the RBD device, data synchronization is performed between the second storage region marked with the dirty region and the corresponding first storage region in the RBD device, and the dirty region mark is cleared after the data synchronization is successful.

6. The method for enhancing data redundancy in block storage under a distributed storage scenario according to claim 5, characterized in that, Before establishing the mapping relationship between the first storage area of ​​the RBD device and the second storage area of ​​the local auxiliary device, the method further includes: The RBD device is divided into multiple first storage areas according to a preset space size, and the local auxiliary device is divided into multiple second storage areas according to the preset space size; wherein the preset space size is larger than the original block size of the RBD device.

7. The method for enhancing data redundancy in block storage under a distributed storage scenario according to claim 3, characterized in that, The method further includes: If a failure occurs during the first data write operation to the RBD device, the second data write operation to the local auxiliary device, or the data synchronization between the RBD device and the local auxiliary device, the failure information is obtained and recorded in the log, and an alarm is triggered according to a preset alarm policy. The failure information includes at least one of the following: RBD device information, local auxiliary device information, write address information of the failed operation, and the length of the written business data.

8. The method for enhancing data redundancy in block storage under a distributed storage scenario according to claim 1, characterized in that, After executing the first data write operation corresponding to the first data write request and the second data write operation corresponding to the second data write request, the method further includes: Receive a data read request to read the business data; In response to a failure of the distributed storage cluster corresponding to the RBD device, the service data is read by accessing the second storage area of ​​the local auxiliary device; If the distributed storage cluster corresponding to the RBD device is normal and the second storage area is currently set with the dirty area marker, then if the first data write operation is successful, the business data can be read by accessing the first storage area; if the second data write operation is successful, the business data can be read by accessing the second storage area. If the distributed storage cluster corresponding to the RBD device is normal and the second storage area is not currently marked with the dirty area flag, then the business data can be read by accessing the second storage area.

9. The method for enhancing data redundancy in block storage under a distributed storage scenario according to any one of claims 1 to 6, characterized in that, The method further includes: One or more local redundant disks can be pooled and managed in a unified manner using user-space tools to obtain one or more local storage pools; Each of the local storage pools is divided into one or more local auxiliary devices.

10. A device for implementing enhanced data redundancy in block storage under a distributed storage scenario, characterized in that, include: The receiving module is configured to: receive a first data write request to an RBD device; wherein the first data write request includes service data to be written and index data of the service data, and the index data includes information on a first data write address for writing the service data to the RBD device and the data length of the service data; The determining module is configured to: determine, based on the index data, at least one first storage area of ​​the RBD device for storing the service data; The mapping module is configured to: determine a second data write address in the local auxiliary device based on the first data write address, and determine a second storage area in the local auxiliary device based on the first storage area; wherein, there is a mapping relationship between the addresses of the RBD device and the local auxiliary device, and there is a mapping relationship between the first storage area and the second storage area; The construction module is configured to: construct a second data write request based on the first data write request to write the business data to the second data write address of the local auxiliary device; The write module is configured to: execute a first data write operation corresponding to the first data write request and a second data write operation corresponding to the second data write request, and determine the dirty area marking status of the second storage area based on the results of the first data write operation and the second data write operation.

Images