CN104090729B - The method and device of mirror image synchronization is repaired by business write operation - Google Patents

Abstract

The present invention provides a method and device for repairing mirror synchronization through business write operations. The method includes: when executing the synchronization task of synchronizing the data of the main resource to the mirror resource, recording the address of the faulty resource in the faulty area to the cache resource, and marking the master resource The mirroring status of the mirroring resource is not synchronized, and the address of the faulty resource includes the address of the faulty resource in the faulty area where the mirroring resource cannot be written because the primary resource is not read successfully, or the primary resource is read successfully but the mirroring resource is not written successfully; when writing business data , if the mirror resource is successfully written, and the successfully written area belongs to the faulty area, clear the faulty resource address corresponding to the successfully written area; when the synchronization task is completed and there is no faulty resource address in the cache resource, the mirroring status is marked as synchronized. When the mirroring state is not synchronized by applying the present invention, as long as the mirroring resource writes business data successfully, the main resource can be replaced to provide mirroring protection for the business, thereby maximizing the utilization rate of the mirroring resource.

Description

Method and device for repairing mirror synchronization through business write operations

technical field

The invention relates to the field of storage technologies, in particular to a method and device for repairing mirror synchronization through business write operations.

Background technique

Mirroring technology is to write data into two independent disks at the same time, such as RAID1 (Redundant Array of Independent Disks1, Redundant Array of Independent Disks 1), or write data into volume resources established on multiple physical resources. Among them, one resource is called the main resource, and the other resource is called a mirror resource as a copy of the main resource. When the main resource fails, the business can be switched to the mirror resource to ensure that the business is not interrupted. When establishing the mirror relationship between the primary resource and the mirror resource, all the data of the primary resource is fully synchronized to the mirror resource by executing the synchronization task. When writing business data, the business data is written to the primary resource and the mirror resource simultaneously to ensure The data are consistent.

In the prior art, when performing a synchronization task, if the mirror resource cannot write data due to failure to read data from the primary resource, or fails to write data read from the primary resource to the mirror resource, the mirroring of the two resources is considered The state is unsynchronized; in the unsynchronized state, even if the mirror resource can successfully write business data, it will not be able to provide mirror protection when the primary resource fails because of the unsynchronized state, resulting in business interruption.

Contents of the invention

The present invention provides a method and device for repairing mirror synchronization through business writing operations, so as to solve the problem in the prior art that mirror resources that successfully write business data cannot provide mirror protection in an unsynchronized state.

According to the first aspect of the embodiments of the present invention, there is provided a method for repairing mirror synchronization through a business write operation, the method is applied to a storage device, and a cache resource is set on the storage device, and the method includes:

When executing the synchronization task of synchronizing the data of the primary resource to the mirror resource, record the address of the faulty resource in the faulty area to the cache resource, and mark the mirror state of the primary resource and the mirror resource as unsynchronized, wherein the address of the faulty resource Including the address of the faulty resource in the faulty area where the mirror resource cannot be written due to the unsuccessful read of the primary resource, and/or the faulty resource address of the faulty area where the read of the primary resource is successful but the writing of the mirror resource is unsuccessful;

When writing business data, it is judged whether the mirror resource is successfully written, if the mirror resource is successfully written, and the successfully written area belongs to the faulty area, then clear the faulty resource address corresponding to the successfully written area;

When the synchronization task is completed and there is no faulty resource address in the cache resource, mark the mirror state as synchronized.

According to the second aspect of the embodiments of the present invention, there is provided an apparatus for repairing mirror synchronization through a business write operation, the apparatus is applied to a storage device, and cache resources are set on the storage device, and the apparatus includes:

The synchronization task execution unit is used to record the address of the faulty resource in the fault area to the cache resource when executing the synchronization task of synchronizing the data of the primary resource to the mirror resource, and mark the mirror state of the primary resource and the mirror resource as unsynchronized, Wherein, the address of the faulty resource includes the address of the faulty resource in the faulty area where the mirror resource cannot be written because the primary resource is unsuccessfully read, and/or, the faulty resource address of the faulty area where the primary resource is successfully read but the mirror resource is unsuccessfully written;

The business writing maintenance unit is used to judge whether the mirroring resource is written successfully when writing business data, and if the mirroring resource is written successfully, and the successfully written area belongs to the faulty area, then clear the faulty resource address corresponding to the successfully written area;

The mirror synchronization repair unit is configured to mark the mirror state as synchronized when the synchronization task is completed and there is no faulty resource address in the cache resource.

Applying the embodiment of the present invention, when the synchronization task is executed through the buffer resource record, the mirror resource cannot be written due to the unsuccessful read of the master resource, or the failure resource address of the fault area where the master resource is successfully read but the mirror resource is not written successfully. When the mirroring status of the resource is not synchronized, as long as the business is successfully written to the mirroring resource, the faulty resource address in the corresponding faulty area will be cleared, and the mirroring status can be updated to synchronized when the synchronization task is completed and there is no faulty resource address in the cache resource. This ensures that as long as the mirror resource writes business data successfully, it can replace the primary resource and provide mirror protection for the business, thereby maximizing the utilization of mirror resources.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description serve to explain the principles of the invention.

FIG. 1 is a schematic diagram of an application scenario of an embodiment of the present invention;

FIG. 2 is a flow chart of mirror resource execution synchronization tasks in an embodiment of the present invention;

FIG. 3 is a flow chart of a method for restoring the synchronization status of mirrored resources in an embodiment of the present invention;

Fig. 4 is a hardware environment architecture diagram of the operation of the device for restoring the mirror resource synchronization state in an embodiment of the present invention;

Fig. 5 is a block diagram of an apparatus for restoring the synchronization status of mirrored resources in an embodiment of the present invention.

detailed description

The mirror technology in the present invention is to write data into two independent disks at the same time, such as RAID1, or write data into volume resources established on multiple physical resources. Referring to Figure 1, it is a schematic diagram of an application scenario of an embodiment of the present invention:

In Figure 1, the client device communicates with the storage device through the network, and the storage device establishes a mirror relationship for two independent resources, one of which is the main resource and the other is the mirror resource, and the storage space on the main resource and the mirror resource According to the stripe (block) division of the array where the resources are located, each stripe has a corresponding resource address. When service data is generated on the client device, the storage device writes the service data into the primary resource and the mirror resource. When writing business data, you can use the synchronous mirroring writing method, that is, write the business data into the main resource and the mirroring resource at the same time; you can also use the asynchronous mirroring writing method, that is, write the business data into the dedicated cache (cache) first, and then write it to the Business data is written to the main resource and the mirror resource respectively from the dedicated cache. In the present invention, there is no limitation on the way of writing service data.

In the present invention, a cache resource is provided on the storage device to record the address of the faulty resource in the faulty area where the mirror resource cannot be written due to the failure to read data from the primary resource when the primary resource and the mirror resource perform a synchronization task. Preferably, UPS (Uninterruptible Power Supply, uninterruptible power supply) support may be provided for the cache resource, so as to ensure that the address of the faulty resource recorded in the cache resource will not be lost due to power failure of the storage device. Based on the address of the faulty resource recorded in the cache resource, when writing business data, as long as the business is successfully written to the mirror resource, the address of the faulty resource in the corresponding faulty area will be cleared, and when the synchronization task is completed and there is no faulty resource address in the cache resource, the The mirroring status is updated to synchronized, so as to ensure that as long as the mirroring resource writes business data successfully, mirroring protection can be provided for the business. The solution of the present invention will be described in detail below through specific examples.

Referring to Fig. 2, it is a flow chart of mirror resource execution synchronization task in one embodiment of the present invention:

Step 200: Start to execute the synchronization task of synchronizing the main resource to the mirror resource.

Step 201: sequentially synchronizing the data of a stripe on the primary resource array to the corresponding stripe of the array where the mirror resource is located.

When establishing a mirror relationship between the primary resource and the mirror resource, the data of the primary resource is synchronized to the mirror resource by executing a synchronization task. When executing the synchronization task, the data of each stripe of the array corresponding to the online primary resource is synchronized one by one to the corresponding stripe of the array where the mirror resource is located. The synchronization process includes reading data from the primary resource and writing data to the mirror resource .

Step 202: Determine whether the main resource is online, if yes, execute step 203; otherwise, execute step 208.

Step 203: Judging whether the master resource data is successfully read, if yes, go to step 204; otherwise, go to step 207.

Step 204: Determine whether the mirror resource is online, if yes, execute step 205; otherwise, execute step 208.

Step 205: Judging whether the data read from the primary resource is successfully written to the mirror resource, if yes, go to step 206; otherwise, go to step 207.

Step 206: Determine whether the synchronization task has been executed, if yes, end the current process; otherwise, return to step 201.

Step 207: Record the address of the faulty resource in the fault area where the mirror resource cannot be written or the mirror resource is written unsuccessfully due to the failure to read the primary resource to the cache resource, and mark the mirror status of the primary resource and the mirror resource as unsynchronized, and return to step 206 .

In the present invention, a cache space is set on the storage device, and when the reading of the main resource data or the writing of the mirror resource data fails, the address of the failed fault resource is recorded in the cache resource. When recording the address of the faulty resource, the LUN (Logical Unit Number, Logical Unit Number) identifier corresponding to the faulty resource obtained through combination calculation and the sector location of the actual physical layout may be used as the faulty resource address.

Step 208: Suspend the synchronization task and end the current process.

Referring to FIG. 3 , it is a flow chart of a method for repairing mirror synchronization through business write operations in an embodiment of the present invention. This embodiment is based on the synchronization execution process shown in FIG. 2 , assuming that the mirror status is not synchronized and the cache resource records the fault resource address:

Step 301: Write business data.

In the present invention, the business data can be written in a synchronous writing manner or an asynchronous writing manner. No matter which writing method is used, the corresponding operation is performed according to the result of whether the main resource and the mirror resource are finally successfully written. As an example, the business writing logic process described in this embodiment starts to judge whether the mirror resource is successfully written, that is, first judges the writing result of the mirror resource, and then judges the writing result of the main resource; similarly, the business writing logic process also It can be judged from whether the writing of the main resource is successful, that is, the writing result of the main resource is judged first, and then the writing result of the mirror resource is judged, which will not be repeated in this embodiment.

Step 302: Determine whether the image resource is written successfully, if yes, execute step 303; otherwise, execute step 309.

Step 303: When the successfully written area belongs to the faulty area, clear the faulty resource address corresponding to the successfully written area in the cache.

During the execution of the synchronization task shown in Figure 2, the cache resource records the address of the faulty resource that cannot be written to the mirror resource because the primary resource fails to read data, or the address of the faulty resource that the primary resource reads successfully but the mirror resource fails to write . During the business writing process, when the business data is written to the fault area corresponding to a fault resource address, if the mirror resource is successfully written, the fault resource address of the fault area recorded in the cache resource can be cleared.

Step 304: the synchronization task is completed and there is no faulty resource address in the cache resource, and the mirror status is marked as synchronized.

In the present invention, service writing can occur during the execution of the synchronization task, or after the execution of the synchronization task is completed. Among them, when the synchronization task is completed and there is no faulty resource address in the cache resource, it means that after multiple operations of writing business data, all areas on the mirror resource have successfully written the correct data. At this time, the mirror status is changed from "unsynchronized" to "Synced".

Step 305: Judging whether the main resource is written successfully, if yes, go to step 308; otherwise, go to step 306.

Step 306: Replace the mirror resource with the main resource, and return the business success.

In this embodiment, when the writing of the mirror resource is successful but the writing of the main resource is unsuccessful, since the mirror status has been marked as "synchronized", the mirror resource can immediately replace the master resource to provide mirror protection for the business.

Step 307: Record the address of the faulty resource in the faulty area where the primary resource fails to be written to the cache resource, and mark the mirror status as unsynchronized, and end the current process.

After the mirror resource replaces the primary resource in step 306, the original mirror resource becomes the new primary resource, and the original primary resource becomes the new mirror resource. At this time, the new mirror resource needs to record the address of the faulty resource to the cache resource because the business write fails. Since the new mirror resource cannot provide mirror protection, the mirror status is changed from "synchronized" to "not synchronized" again, and the mirror synchronization task can be restarted according to the existing synchronization mechanism.

Step 308: return business success, and end the current process.

Step 309: Record the address of the faulty resource in the faulty area where the mirroring resource fails to be written to the cache resource, and wait for the execution of the next synchronization task, ending the current process.

Since the mirror resource cannot be written successfully, the mirror resource cannot provide mirror protection for the business, and the mirror resource needs to be repaired when the next synchronization task is executed.

Step 310: Judging whether the main resource is written successfully, if yes, go to step 308; otherwise, go to step 311.

Step 311: return business failure, end the current process.

Combining the embodiments shown in Fig. 2 and Fig. 3, it can be known that the present invention records the fault area where the mirror resource cannot be written because the primary resource fails to be read when the synchronization task is executed through the cache resource record, or the primary resource is successfully read but the mirror resource is not successfully written. Faulty resource address. When the mirror status of the primary resource and the mirror resource is not synchronized, as long as the business is successfully written to the mirror resource, the faulty resource address corresponding to the faulty area will be cleared, and when the synchronization task is completed and there is no faulty resource address in the cache resource, you can Update the mirror status to synchronized to ensure that as long as the mirror resource writes business data successfully, the primary resource can be replaced to provide mirror protection for the business, thereby maximizing the utilization of mirror resources.

Referring to Fig. 4, it is a hardware environment architecture diagram of the operation of the device for restoring the mirror resource synchronization state in an embodiment of the present invention:

The embodiment of the device for repairing mirror synchronization through service write operation of the present invention can be applied to storage devices. The device embodiments can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory for operation by the CPU of the device where it is located. From the hardware level, as shown in Figure 4, it is a hardware structure diagram of the device where the device for repairing mirror synchronization through business write operations in the present invention is located, except for the CPU, memory, network interface, and non-volatile memory shown in Figure 4 In addition to the non-volatile memory, the device where the device in the embodiment is located may generally include other hardware.

Referring to FIG. 5 , it is a block diagram of an apparatus for repairing mirror synchronization through business write operations in an embodiment of the present invention.

The device includes: a synchronization task execution unit 510 , a business write maintenance unit 520 and a mirror synchronization repair unit 530 .

Wherein, the synchronization task execution unit 510 is configured to record the address of the faulty resource in the faulty area into the cache resource when executing the synchronization task of synchronizing the data of the primary resource to the mirror resource, and mark the mirror status of the primary resource and the mirror resource as Not synchronized, wherein the address of the faulty resource includes the address of the faulty resource in the faulty area where the mirror resource cannot be written because the primary resource is unsuccessfully read, or the faulty resource address of the faulty area where the primary resource is read successfully but the mirror resource is unsuccessfully written;

The business writing maintenance unit 520 is used to judge whether the mirroring resource is written successfully when writing business data, if the mirroring resource is written successfully, and the successfully written area belongs to the faulty area, then clear the faulty resource address corresponding to the successfully written area ;

The mirror synchronization repair unit 530 is configured to mark the mirror state as synchronized when the synchronization task is completed and there is no faulty resource address in the cache resource.

Optionally, the business writing maintenance unit 520 can also be used to determine whether the writing of the main resource is successful if the writing of the mirror resource is successful and the mirror status is synchronized when writing business data, and if the writing of the main resource is not If successful, replace the mirror resource with the primary resource, and return the business success, and record the address of the faulty resource in the fault area where the primary resource failed to be written to the cache resource, and mark the mirror status as unsynchronized ; If the main resource is successfully written, it returns business success.

Optionally, the business write maintenance unit 520 may also be configured to record the failed resource address of the failed write in the cache resource when writing business data, if the write of the mirror resource is unsuccessful, and wait for the The synchronization task execution unit executes the next synchronization task.

Optionally, the business writing maintenance unit 520 can also be used to judge whether the writing of the primary resource is successful if the writing of the mirror resource is unsuccessful when writing the business data, and if the writing is successful, return the business success; Success, return business failure.

For the implementation process of the functions and effects of each unit in the above device, please refer to the implementation process of the corresponding steps in the above method for details, and will not be repeated here.

Other embodiments of the invention will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present invention, these modifications, uses or adaptations follow the general principles of the present invention and include common knowledge or conventional technical means in the technical field not disclosed in the present invention . The specification and examples are to be considered exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

It should be understood that the present invention is not limited to the precise constructions which have been described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims (10)

1. A method for repairing mirror synchronization by business write operations, characterized in that, the method is applied to a storage device, and cache resources are set on the storage device, and the method includes: When executing the synchronization task of synchronizing the data of the primary resource to the mirror resource, record the address of the faulty resource in the faulty area to the cache resource, and mark the mirror state of the primary resource and the mirror resource as unsynchronized, wherein the address of the faulty resource Including the address of the faulty resource in the faulty area where the mirror resource cannot be written due to the unsuccessful read of the primary resource, or the faulty resource address of the faulty area where the read of the primary resource is successful but the writing of the mirror resource is unsuccessful; When writing business data, it is judged whether the mirroring resource is written successfully, if the mirroring resource is written successfully, and the successfully written area belongs to the faulty area, then clear the faulty resource address corresponding to the successfully written area; wherein, when writing business data Use synchronous writing method or asynchronous writing method; When the synchronization task is completed and there is no faulty resource address in the cache resource, mark the mirror state as synchronized. 2. The method according to claim 1, characterized in that the method further comprises: When writing business data, if the mirror resource is successfully written and the mirror status is synchronized, judge whether the master resource is successfully written; if the master resource is not successfully written, replace the master resource with the mirror resource and return to the business success; Recording the address of the faulty resource in the faulty area where the primary resource fails to be written to the cache resource, and marking the mirror state as unsynchronized. 3. The method according to claim 2, wherein the method further comprises: When writing business data, if the main resource is successfully written, it will return business success. 4. The method according to claim 1, wherein the method further comprises: When writing business data, if the writing of the mirror resource is unsuccessful, record the address of the failed resource to the cache resource, and wait for the execution of the next synchronization task. 5. method according to claim 4, is characterized in that, described method also comprises: When writing business data, if the mirror resource is not written successfully, it is judged whether the main resource is written successfully. If the writing is successful, the business is returned successfully. If the writing is not successful, the business is returned as failed. 6. A device for repairing mirror synchronization through a business write operation, characterized in that the device is applied to a storage device, and cache resources are set on the storage device, and the device includes: The synchronization task execution unit is used to record the address of the faulty resource in the fault area to the cache resource when executing the synchronization task of synchronizing the data of the primary resource to the mirror resource, and mark the mirror state of the primary resource and the mirror resource as unsynchronized, Wherein, the address of the faulty resource includes the faulty resource address of the faulty area where the mirror resource cannot be written due to the unsuccessful reading of the primary resource, or the faulty resource address of the faulty area where the primary resource is successfully read but the mirroring resource is unsuccessfully written; The business writing maintenance unit is used to judge whether the mirroring resource is written successfully when writing business data, and if the mirroring resource is written successfully, and the successfully written area belongs to the faulty area, then clear the faulty resource address corresponding to the successfully written area; Among them, when writing business data, adopt synchronous writing method or asynchronous writing method; The mirror synchronization repair unit is configured to mark the mirror state as synchronized when the synchronization task is completed and there is no faulty resource address in the cache resource. 7. The device of claim 6, wherein: The business write maintenance unit is also used to judge whether the writing of the main resource is successful if the writing of the mirror resource is successful and the mirror state is synchronized when writing business data, and if the writing of the main resource is unsuccessful, write the mirror image The resource replaces the primary resource, and returns the business success, and records the address of the faulty resource in the faulty area where the primary resource fails to be written to the cache resource, and marks the mirror status as unsynchronized. 8. The device of claim 7, wherein: The business write maintenance unit is further configured to return business success if the main resource is successfully written when writing business data. 9. The apparatus of claim 6, wherein: The business write maintenance unit is also used to record the failed resource address of the failed write in the buffer resource when writing the business data, and wait for the synchronization task execution unit to execute the next step. synchronization tasks. 10. The apparatus of claim 9, wherein: The business write maintenance unit is also used for judging whether the master resource is successfully written when writing business data, if the writing of the mirror resource is unsuccessful, and if the writing is successful, return the business success, and if the writing is unsuccessful, return the business failure.