Storage data migration method, device and system

Abstract

The invention discloses a storage data migration method, device and system, and the method comprises the steps: receiving a migration task sent by a client, the migration task comprising to-be-migrated cloud disk information; allocating a migration service for the migration task according to the to-be-migrated cloud disk information, and determining a target storage cluster; a cluster route and a data conversion protocol are constructed through the migration service, the data migration process is completed, and the data migration process comprises the steps that the migration service reads to-be-migrated data from a source storage cluster according to the cluster route, protocol conversion is conducted on the to-be-migrated data according to the data conversion protocol, and the converted data are written into a target storage cluster. By utilizing the scheme of the invention, efficient and reliable storage data migration with small influence on services can be realized.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to a method, apparatus, and system for migrating stored data. Background Technology

[0002] With the rapid development of information technology, data volume is exploding, and enterprises' demand for data storage is constantly increasing. In practical applications, it is often necessary to migrate data from one storage system to another to meet needs such as performance improvement, cost optimization, or system upgrades. Especially in public cloud environments, storage systems are usually deployed in cluster mode. When a single storage cluster fails, data needs to be migrated from the failed storage cluster to the healthy cluster. The industry's common storage data migration methods mainly include the following:

[0003] (1) Migration based on data replication tools, such as rsync, supports data synchronization from local to remote. This method relies on the host to provide system tools and requires interruption of business migration. Interruption during synchronization will lead to inconsistency of data on the target end.

[0004] (2) Migration based on cloud service provider cloud platform: Each cloud service provider provides a platform to support data migration of its own block storage products between different storage clusters. Because the storage products of each cloud service provider have differences in architecture design and IO protocol definition, the migration method based on the cloud service provider cloud platform is often only suitable for its own products and cannot be compatible with other cloud service providers' storage products, and cannot achieve cross-vendor cloud disk migration. Summary of the Invention

[0005] This invention provides a storage data migration method, apparatus, and system to achieve efficient, reliable storage data migration with minimal impact on business operations.

[0006] On one hand, embodiments of the present invention provide a method for migrating stored data, the method comprising:

[0007] Receive a migration task sent by the client, the migration task including information about the cloud disk to be migrated;

[0008] Based on the cloud disk information to be migrated, allocate migration services for the migration task and determine the target storage cluster;

[0009] The migration service constructs a cluster routing and data conversion protocol to complete the data migration process. The data migration process includes: the migration service reads the data to be migrated from the source storage cluster according to the cluster routing, performs protocol conversion on the data to be migrated according to the data conversion protocol, and writes the converted data into the target storage cluster.

[0010] Optionally, allocating migration services to the migration task based on the cloud disk information to be migrated includes:

[0011] The architecture type of the source storage cluster is determined based on the cloud disk information to be migrated. The architecture type includes: network architecture type and storage architecture type.

[0012] Assign migration services to the migration task based on the architecture type.

[0013] Optionally, the cloud disk information to be migrated includes the cloud disk ID to be migrated; determining the architecture type of the source storage cluster based on the cloud disk information to be migrated includes: looking up the cloud disk service information table based on the cloud disk ID to be migrated to determine the architecture type of the source storage cluster.

[0014] Optionally, the migration task further includes: storage requirement information; determining the target storage cluster includes: determining the target storage cluster based on the storage requirement information.

[0015] Optionally, determining the target storage cluster includes: selecting the target storage cluster based on the cloud disk information to be migrated.

[0016] Optionally, the step of constructing a cluster routing and data conversion protocol through the migration service includes: the migration service constructing a cluster routing and data conversion protocol based on the architecture type of the source storage cluster to which the cloud disk to be migrated belongs and the architecture type of the target cluster.

[0017] Optionally, the method further includes:

[0018] If the cloud disk to be migrated is currently online, the migration service obtains the client's real-time IO interface data;

[0019] For write IO interface data, the write IO interface data is merged with the read data to be migrated, the merged data is written to the source storage cluster, and the merged data is written to the target storage cluster after protocol conversion.

[0020] Optionally, the method further includes:

[0021] Before the migration service reads the data to be migrated from the source storage cluster, it obtains the current status of the cloud disk to be migrated by querying the cloud disk status record;

[0022] During the data migration process, the migration service obtains the current status of the cloud disk to be migrated based on the cloud disk online / offline notification sent by the client.

[0023] Optionally, each migration service can support multiple migration tasks simultaneously.

[0024] On the other hand, embodiments of the present invention also provide a storage data migration device, the device comprising: a receiving module, a task management module, and one or more migration service modules;

[0025] The receiving module is used to receive migration tasks sent by the client, and the migration tasks include information about the cloud disks to be migrated.

[0026] The task management module is used to allocate migration services to the migration task based on the cloud disk information to be migrated, and to determine the target storage cluster;

[0027] The migration service module is used to construct cluster routing and data conversion protocols according to the migration service allocated by the task management module, and complete the data migration process. The data migration process includes: reading the data to be migrated from the source storage cluster according to the cluster routing, performing protocol conversion on the data to be migrated according to the data conversion protocol, and writing the converted data into the target storage cluster.

[0028] Optionally, the migration service module is further configured to obtain the real-time IO interface data of the client when the cloud disk to be migrated is currently online; for write IO interface data, merge the write IO interface data with the read data to be migrated, write the merged data into the source storage cluster, and perform protocol conversion on the merged data before writing it into the target storage cluster.

[0029] Optionally, the migration service module is further configured to obtain the current status of the cloud disk to be migrated by querying the cloud disk status record before reading the data to be migrated from the source storage cluster; and to obtain the current status of the cloud disk to be migrated based on the cloud disk online / offline notification sent by the client during the data migration process.

[0030] On the other hand, embodiments of the present invention also provide a storage data migration system, the system comprising: a computing node, multiple storage clusters, and the aforementioned storage data migration device;

[0031] The storage cluster is used to store data corresponding to different cloud disks;

[0032] The storage data migration device is used to receive migration tasks sent by the computing node client, allocate migration services for the migration tasks, and determine the target storage cluster; and to construct cluster routing and data conversion protocols through the migration services to complete the data migration process.

[0033] The storage data migration method, apparatus, and system provided in this invention, upon receiving a migration task sent by a client, allocates a migration service for the migration task and determines the target storage cluster; the migration service constructs cluster routing and data conversion protocols to achieve data migration. During the data migration process, the migration service reads the data to be migrated from the source storage cluster according to the cluster routing, performs protocol conversion on the data to be migrated according to the data protocol, and writes the converted data to the target storage cluster. Using this invention, cloud disks can be migrated between different storage clusters. Moreover, it is compatible with different network and storage architectures, enabling mutual migration of cloud disks between different types of storage clusters and supporting different protocols. Furthermore, it can automatically adjust the number of concurrent tasks and the data migration rate according to the urgency of the migration task.

[0034] Furthermore, the migration service supports both offline and online data migration. The migration service obtains real-time IO interface data from the client. For write IO interface data, it merges the write IO interface data with the read data to be migrated, writes the merged data to the source storage cluster, and then performs protocol conversion on the merged data before writing it to the target storage cluster. This invention supports both offline and online migration methods, with the entire migration process being seamless for the business side and having zero impact on the user's actual business operations. Attached Figure Description

[0035] Figure 1 This is a flowchart of a storage data migration method provided in an embodiment of the present invention;

[0036] Figure 2 This is another flowchart of the storage data migration method provided in the embodiments of the present invention;

[0037] Figure 3 This is a schematic diagram of a storage data migration device provided in an embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of a storage data migration system provided in an embodiment of the present invention;

[0039] Figure 5 yes Figure 4 The diagram shows the data migration process in the data migration system. Detailed Implementation

[0040] To make the above-mentioned objectives, features and beneficial effects of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] To address the numerous problems with existing data migration methods, such as low migration efficiency, difficulty in ensuring data consistency, and significant impact on business systems, this invention provides a storage data migration method, apparatus, and system that enables cloud disks to migrate between different types of storage clusters, supports migration between different network architectures and different storage architectures, supports both offline and online migration methods, and ensures that the online migration process has no impact on upper-layer business.

[0042] like Figure 1 The diagram shown is a flowchart of a storage data migration method provided in an embodiment of the present invention, which includes the following steps:

[0043] Step 101: Receive the migration task sent by the client. The migration task includes information about the cloud disk to be migrated.

[0044] Step 102: Assign migration services to the migration task based on the cloud disk information to be migrated, and determine the target storage cluster.

[0045] Specifically, the architecture type of the source storage cluster can be determined based on the information of the cloud disk to be migrated. The architecture type includes network architecture type and storage architecture type. Then, migration services are assigned to the migration task based on the determined architecture type of the source storage cluster.

[0046] In some embodiments, the cloud disk information to be migrated may be the cloud disk ID to be migrated. Accordingly, the cloud disk service information table can be looked up based on the cloud disk ID to determine the architecture type of the source storage cluster. The cloud disk service information table may include information on all online cloud disks, such as cloud disk IDs and their corresponding cloud disk-related information. The cloud disk-related information may include, but is not limited to, information such as the storage cluster corresponding to the cloud disk, the network architecture type of the storage cluster, and the storage architecture type.

[0047] The cloud disk service information table can be dynamically updated by the corresponding server based on the online / offline status of each cloud disk.

[0048] There are several ways to determine the target storage cluster, such as:

[0049] In one non-limiting embodiment, storage requirement information can be carried in the migration task. Accordingly, the target storage cluster can be determined based on this storage requirement information.

[0050] In another non-limiting embodiment, the target storage cluster can be selected based on the cloud disk information to be migrated. The cloud disk information includes the cloud disk type, which may include, for example, a regular cloud disk, an SSD (Solid-State Disk) cloud disk, etc. Different cloud disk types correspond to different underlying storage media, and different storage media correspond to different types of storage clusters.

[0051] Step 103: The migration service constructs a cluster routing and data conversion protocol to complete the data migration process. The data migration process includes: the migration service reads the data to be migrated from the source storage cluster according to the cluster routing, performs protocol conversion on the data to be migrated according to the data conversion protocol, and writes the converted data into the target storage cluster.

[0052] Specifically, the migration service can construct cluster routing and data protocols based on the architecture type of the source storage cluster to which the cloud disk to be migrated belongs and the architecture type of the target cluster. For example, if the network architecture of the source storage cluster is TCP (Transmission Control Protocol) and the network architecture of the target cluster is RDMA (Remote Direct Memory Access), then the migration service needs to construct routing and data protocols from TCP to RDMA. That is, the data to be migrated is obtained from the source storage cluster via TCP and written to the target storage cluster via RDMA.

[0053] The migration service can directly access data on the source storage cluster and directory storage cluster, parse the data format of the data to be migrated read from the source storage cluster, parse out the specific data format, and determine how to convert it.

[0054] It should be noted that if the source storage cluster and the target storage cluster have the same architecture, then no parsing or data format conversion is required.

[0055] By utilizing the constructed cluster routing and data conversion protocols, data migration can be made compatible with different network and storage architectures, enabling cloud disks to migrate between different types of storage clusters and to be compatible with different protocols.

[0056] It should be noted that in this embodiment of the invention, each migration service can support migration tasks for multiple (e.g., 4) cloud disks simultaneously.

[0057] Furthermore, the migration operation should be seamless for upper-layer services, and the data traffic generated during migration should not affect these services. Therefore, a limit can be set on the migration rate of each cloud disk to reduce the impact on upper-layer services. For example, the migration rate limit could be set to 200MB / s; another example is setting appropriate flow control policies to limit the migration bandwidth of each cloud disk based on the token bucket algorithm.

[0058] In practical implementation, the number of concurrent migration tasks and the data migration rate that each migration service can support can be set according to actual application needs, and this embodiment of the invention does not limit this.

[0059] Furthermore, the migration service not only supports offline migration of cloud disk data but also online migration of cloud disk data. The following section will combine... Figure 2 This will be explained in detail.

[0060] like Figure 2 The diagram shown is another flowchart of a storage data migration method provided in an embodiment of the present invention, including the following steps:

[0061] Step 201: Receive the migration task sent by the client. The migration task includes information about the cloud disk to be migrated.

[0062] Step 202: Assign migration services to the migration task based on the cloud disk information to be migrated, and determine the target storage cluster.

[0063] Step 203: The migration service determines whether the cloud disk to be migrated is offline or online. If it is offline, proceed to step 204; otherwise, proceed to step 205.

[0064] Specifically, before the migration service reads the data to be migrated from the source storage cluster, it can obtain the current status of the cloud disk to be migrated by querying the cloud disk status record; or during the data migration process, the migration service can obtain the current status of the cloud disk to be migrated based on the cloud disk online / offline notifications sent by the client.

[0065] Step 204: The migration service reads the data to be migrated from the source storage cluster according to the cluster route, performs protocol conversion on the data to be migrated according to the data conversion protocol, and writes the converted data into the target storage cluster.

[0066] Step 205: The migration service obtains real-time IO interface data from the client. For write IO interface data, it merges the write IO interface data with the read data to be migrated, writes the merged data to the source storage cluster, and performs protocol conversion on the merged data before writing it to the target storage cluster.

[0067] By merging the write I / O interface data with the read data to be migrated before writing it to the source storage cluster, the integrity of the data on the source storage cluster can be guaranteed. If an exception occurs during the migration process causing migration failure, the data on the source storage cluster can be rolled back, effectively ensuring the integrity of the migrated data.

[0068] The storage data migration method provided in this invention supports both offline and online migration methods. The entire migration process is seamless for the business side and has zero impact on the user's actual business.

[0069] It should be noted that during the migration process, cloud disks may be brought online or taken offline. The migration service can detect this automatically and perform different migration operations based on the current status of the cloud disks. This will not terminate the migration process or affect the customer's business.

[0070] In this embodiment of the invention, each migration task is a full data migration at the cloud disk granularity. During migration, data needs to be migrated sequentially from the starting position on the cloud disk. Data migration can be performed in units of data shards (also known as data partitions). For example, the minimum data shard for data migration is 256KB, but other sizes are also possible and can be modified according to actual needs. After each migration, the current migrated position is recorded so that data can be read from that position in the next migration.

[0071] In some embodiments, to ensure the consistency of the migrated data, the migration service can also perform CRC checks on the corresponding data shards of the source storage cluster and the target storage cluster after each data shard is migrated. If the CRC check fails, the migration can be terminated and rolled back immediately.

[0072] The storage data migration method provided in this invention, upon receiving a migration task sent by a client, allocates a migration service to the migration task and determines the target storage cluster; it then constructs a cluster routing and data conversion protocol through the migration service to achieve data migration. This invention is compatible with different network and storage architectures, enabling mutual migration of cloud disks between different types of storage clusters and supporting different protocols. Furthermore, it supports not only offline migration of cloud disk data but also online migration.

[0073] Accordingly, embodiments of the present invention also provide a storage data migration apparatus, such as... Figure 3 The diagram shown is a structural schematic of the device.

[0074] The storage data migration device 300 of this embodiment includes: a receiving module 301, a task management module 302, and one or more migration service modules 303. Wherein:

[0075] The receiving module 301 is used to receive migration tasks sent by the client, and the migration tasks include information about the cloud disks to be migrated;

[0076] The task management module 302 is used to allocate migration services to migration tasks based on the cloud disk information to be migrated, and to determine the target storage cluster;

[0077] The migration service module 303 is used to construct cluster routes and data protocols according to the migration service allocated by the task management module, and complete the data migration process. The data migration process includes: reading the data to be migrated from the source storage cluster according to the cluster route, performing protocol conversion on the data to be migrated according to the data conversion protocol, and writing the converted data into the target storage cluster.

[0078] The task management module 302 can determine the architecture type of the source storage cluster based on the information of the cloud disk to be migrated. The architecture type includes network architecture type and storage architecture type. Then, it allocates migration services to the migration task based on the determined architecture type of the source storage cluster. In addition, the task management module 302 can determine the target storage cluster based on the information of the cloud disk to be migrated or storage requirement information. This embodiment of the invention does not limit this.

[0079] The storage data migration device provided in this embodiment of the invention can support both offline and online migration of cloud disk data. Specifically, before the migration service module 303 reads the data to be migrated from the storage cluster, it can obtain the current status of the cloud disk to be migrated by querying the cloud disk status record; during the data migration process, it obtains the current status of the cloud disk to be migrated based on the cloud disk online / offline notification sent by the client. When the cloud disk to be migrated is currently online, it obtains the real-time IO interface data from the client; for write IO interface data, it merges the write IO interface data with the read data to be migrated, writes the merged data to the source storage cluster, and performs protocol conversion on the merged data before writing it to the target storage cluster.

[0080] It should be noted that each migration service module 303 can support multiple (e.g., 4) cloud disk migration tasks simultaneously. The number of concurrent migration tasks and data migration rate that each migration service module 303 can support can be set according to actual application needs, and this embodiment of the invention does not limit this.

[0081] Accordingly, embodiments of the present invention also provide a storage data migration system, such as... Figure 4 The diagram shown is a structural schematic of this data migration storage system.

[0082] The storage data migration system includes: compute node 401 and multiple storage clusters 402, such as Figure 4 The storage clusters A and B shown also include a storage data migration device 300. Wherein:

[0083] The compute node 401 includes a cloud host, on which a cloud disk is mounted. The cloud disk is a virtual hard disk that provides the cloud host with scalable and persistent disk space.

[0084] Storage cluster 402 is used to store data corresponding to different cloud disks;

[0085] The storage data migration device 300 is used to receive migration tasks sent by the client of the computing node 401 through the trigger, assign migration service Trans to the migration task, and determine the target storage cluster; construct cluster routing and data transformation protocol through migration service Trans to complete the data migration process, for example, reading data from storage cluster A (source storage cluster), parsing and transforming the read data and writing it to storage cluster B.

[0086] In this embodiment, the storage data migration device 300 can support not only offline migration of cloud disk data, but also online migration of cloud disk data. The specific processes of offline migration and online migration can be found in the previous description of the storage data migration device 300 embodiment of the present invention, and will not be repeated here.

[0087] It should be noted that in actual deployment, storage clusters 402 can be divided by region, into different storage clusters, such as storage clusters in Beijing, Shanghai, Guangzhou, etc., or referred to as the first storage cluster, the second storage cluster, the third storage cluster, etc. Storage clusters in the same region use the same network architecture and storage architecture, while storage clusters in different regions can use the same or different network architectures and storage architectures. The network architecture can be, for example, RDMA (Remote Direct Memory Access) or TCP (Transmission Control Protocol), and the storage architecture can be, for example, a unified storage architecture or UDisk architecture (a backend architecture for high-performance, highly available block storage services provided by Alibaba Cloud in its cloud computing platform), etc.

[0088] To ensure that the migration system does not excessively consume network and I / O resources and avoid impacting existing network services, while also considering migration efficiency, it is possible to limit the number of cloud disks that can be migrated simultaneously in each availability zone (the region where the cloud disk storage cluster is located) (e.g., the default is to support the simultaneous migration of 48 cloud disks), the maximum number of cloud disks that each migration service can support (e.g., 4), and the upper limit of the migration rate for each cloud disk (e.g., set to 200MB / s).

[0089] In the storage data migration system provided in this embodiment of the invention, the storage data migration device 300 can also provide an API (Application Programming Interface) to support dynamic adjustment of the above parameters, and can automatically adjust the parameters according to the urgency of the migration task.

[0090] The following is combined Figure 5 right Figure 4 The storage data migration process in the storage data migration system is described in detail.

[0091] Reference Figure 5 and Figure 4 When the Trigger receives a cloud disk migration task initiated by the XClient client, it obtains information about the source and target cloud disk storage clusters from the access server, selects the migration service Trans-n to execute the migration task, and sends the migration task to the Trans-n service.

[0092] After receiving the migration task, the Trans-n migration service reads the base data (i.e., the data in the source storage cluster corresponding to the cloud disk to be migrated) from the source storage cluster and writes the read base data to the target storage cluster. If the cloud disk to be migrated is online, for example, if the client XClient is interacting with other clients via IO interface data, for write IO interface data, the Trans-n migration service merges the write IO interface data with the read base data, writes the merged data to the source storage cluster, and performs protocol conversion on the merged data before writing it to the target storage cluster.

[0093] Simultaneously, the trigger maintains heartbeat communication with the migration service Trans-n to determine the liveness status of Trans-n. Additionally, the migration service Trans-n also sends migration progress and status information to the access server.

[0094] It should be noted that the storage data migration device 300 can be set on the access server or independent of the access server, and this embodiment of the invention does not limit this.

[0095] In specific implementation, the modules / units included in the various devices and equipment in the data migration system described in the above embodiments can be software modules / units, hardware modules / units, or a combination of both.

[0096] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article indicates that the preceding and following related objects have an "or" relationship.

[0097] In the embodiments of this invention, "multiple" refers to two or more.

[0098] The descriptions of "first," "second," etc., appearing in the embodiments of this invention are for illustrative purposes and to distinguish the objects being described. They do not indicate any particular order and do not imply any special limitation on the number of devices in the embodiments of this invention. They do not constitute any limitation on the embodiments of this invention.

[0099] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A method of migrating data storage, the method comprising: The method comprises: receiving a migration task sent by a client, the migration task comprising cloud disk information to be migrated; allocating a migration service for the migration task according to the cloud disk information to be migrated, and determining a target storage cluster; constructing a cluster routing and data conversion protocol through the migration service, and completing a data migration process, the data migration process comprising: reading, by the migration service, to-be-migrated data from a source storage cluster according to the cluster routing, performing protocol conversion on the to-be-migrated data according to the data conversion protocol, and writing the converted data to the target storage cluster.

2. The method of claim 1, wherein, The allocation of the migration service for the migration task according to the cloud disk information to be migrated comprises: determining an architecture type of the source storage cluster according to the cloud disk information to be migrated, the architecture type comprising: a network architecture type and a storage architecture type; allocating a migration service for the migration task according to the architecture type.

3. The method of claim 2, wherein, The cloud disk information to be migrated comprises a cloud disk ID to be migrated; and the determination of the architecture type of the source storage cluster according to the cloud disk information to be migrated comprises: looking up a cloud disk service information table according to the cloud disk ID to be migrated, and determining the architecture type of the source storage cluster.

4. The method of claim 1, wherein, The migration task further comprises storage requirement information; and the determination of the target storage cluster comprises: determining the target storage cluster according to the storage requirement information.

5. The method of claim 1, wherein, The determination of the target storage cluster comprises: selecting the target storage cluster according to the cloud disk information to be migrated.

6. The method of claim 1, wherein, The construction of the cluster routing and data conversion protocol through the migration service comprises: constructing, by the migration service, a cluster routing and data conversion protocol according to an architecture type of a source storage cluster to which the cloud disk to be migrated belongs and an architecture type of the target cluster.

7. The method of claim 1, wherein, The method further comprises: if the cloud disk to be migrated is currently in an online state, acquiring, by the migration service, real-time IO interface data of the client; for write IO interface data, merging the write IO interface data with read to-be-migrated data, writing the merged data to the source storage cluster, and writing the merged data to the target storage cluster after protocol conversion.

8. The method of claim 7, wherein, The method further comprises: acquiring, by the migration service, a current state of the cloud disk to be migrated by querying cloud disk state records before reading the to-be-migrated data from the source storage cluster; during the data migration process, acquiring, by the migration service, the current state of the cloud disk to be migrated according to cloud disk online-offline notifications sent by the client.

9. The method of claim 1 to 8, wherein, Each migration service supports multiple migration tasks simultaneously.

10. A data migration storage apparatus, characterized by comprising: The apparatus comprises a receiving module, a task management module, and one or more migration service modules; the receiving module is configured to receive a migration task sent by a client, the migration task comprising cloud disk information to be migrated; the task management module is configured to allocate a migration service for the migration task according to the cloud disk information to be migrated, and determine a target storage cluster; the task management module is configured to allocate a migration service for the migration task according to the architecture type. The migration service module is configured to construct a cluster routing and a data conversion protocol according to the migration service allocated by the task management module, complete a data migration process, and the data migration process includes: reading to-be-migrated data from a source storage cluster according to the cluster routing, performing protocol conversion on the to-be-migrated data according to the data conversion protocol, and writing the converted data to the target storage cluster.

11. The storage data migration apparatus according to claim 10, wherein, The migration service module is further configured to, when the to-be-migrated cloud disk is currently in an online state, acquire real-time IO interface data of the client; for write IO interface data, merge the write IO interface data with the read to-be-migrated data, write the merged data to the source storage cluster, and write the merged data to the target storage cluster after performing protocol conversion on the merged data.

12. The storage data migration apparatus according to claim 11, wherein, The migration service module is further configured to, before reading the to-be-migrated data from the source storage cluster, acquire a current state of the to-be-migrated cloud disk by querying cloud disk state records; and during the data migration process, acquire the current state of the to-be-migrated cloud disk according to cloud disk online / offline notifications sent by the client.

13. A data migration system storing, The system comprises: a computing node, a plurality of storage clusters, and the storage data migration apparatus according to any one of claims 10 to 12. The storage cluster is configured to store data corresponding to different cloud disks. The storage data migration apparatus is configured to receive a migration task sent by a client of the computing node, allocate a migration service for the migration task, and determine a target storage cluster; construct a cluster routing and a data conversion protocol through the migration service, and complete a data migration process.

Images