JP2000330730A - Differential copy system for disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To minimize influence on a response to a host by minimizing a data transfer quantity in a background when a copy volume is mirrored again after being separated with respect to such a disk array device on which a job volume and a copy volume which is mirrored, can be accessed by an instruction from a host while separated and can be mirrored again are mounted. SOLUTION: When a copy volume 19 is disconnected from a host, a host director 12 prepares an updated history table in which the updated histories of a job volume 18 and the volume 19 are described for each of the volumes 18 and 19 and stores the updated histories of the respective volumes 18 and 19 while being disconnected. When a re-mirroring instruction is issued, updated difference is decided by taking the OR operations of the updated history tables of the volumes 18 and 19 and only minimum data are copied in a background according to them.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a disk array device, and more particularly, to a differential copy method for a disk array device that performs efficient copy processing using an update history table.

[0002]

2. Description of the Related Art In a modern mission-critical business environment, users require 24-hour access to user data stored inside a disk array, and as a result, it is not possible to stop business and take a backup of data. It will have a significant impact on user business opportunities. Therefore, in the related art, a mirror volume of a volume used in a business is used as a host computer (hereinafter, referred to as a host) so as to back up data while minimizing the time during which the business is stopped. A copy volume that can be accessed from the server and that can be re-mirrored is created inside the disk array device. When making a backup, the copy volume is separated from the business volume, and when the separation is completed, the business volume is returned to normal business. On the other hand, the copy volume is used for backup of the transaction volume data immediately before the separation, and after the backup is completed, the transaction volume and the copy volume are re-mirrored. At this time, while the transaction volume and the copy volume are separated, there is a possibility that data is updated in the transaction volume and the copy volume, so the entire contents of the transaction volume must be copied to the copy volume. ing.

[0003] Also, JP-A-58-9296 discloses that
In a duplicated data storage device, a data storage area management storage device is provided for each, and a difference in the contents is extracted, so that the storage data of one device can be efficiently stored data of the other device. A technique for copying data to a computer has been disclosed.

[0004]

However, Japanese Patent Application Laid-Open No. 58-1983
In the prior art as described above including Japanese Patent Application Laid-Open No. 9296, since the copy volume is separated and re-mirrored after work such as backup and batch processing is completed, the entire contents of the transaction volume must be copied to the copy volume. No. For this reason, a large amount of data is transferred and copied in the background, and during this time, the response to the host is greatly affected.

The present invention has been made in view of such circumstances, and an object of the present invention is to minimize the amount of data transfer in the background and to minimize the influence on the response to the host. Another object of the present invention is to provide a differential copy method for a disk array device.

[0006]

In order to solve the above-mentioned problem, the differential copy method of the disk array device according to the present invention uses the transaction volume 18 and the copy volume in the configuration of FIG. A state transition table (FIGS. 2, 3, and 4) for determining the state of each volume while the server 19 is being operated separately from the host 19, and a state transition table generated for the transaction volume 18 and the copy volume 19 , And an update history table (FIG. 5) for each volume for determining whether or not the data update block exists in the cache. As a result, after the transaction volume 18 and the copy volume 19 are separated, the data update from the host that occurred in the transaction volume 18 is described in the data update history table of the transaction volume 18 (FIG. 5). The host director 12 writes and updates the data in the copy volume 19 in the data update history table (FIG. 5) of the copy volume 19.

When the operation such as backup is completed and the transaction volume 18 and the copy volume 19 are remirrored, the logical sum of the update flags 51 of both data update history tables (FIG. 5) is obtained. The update difference of the volume is described in the update history table (FIG. 5). Next, only the update difference between the two volumes is transferred to the operation volume 18.
Since the data is copied to the copy volume 19 in the background, the amount of data to be transferred is minimal and the effect on the response to the host is also minimal.

That is, according to the differential copy method for a disk array device according to the first aspect, the data used for business is backed up to a copy volume that can be re-mirrored, so that the difference of the disk array device that manages the data can be improved. In the copy method, a host director that sends and receives commands from a host computer, a cache memory that temporarily stores data, and a plurality of disks
The disk module is composed of a disk module that constitutes an ID and a disk director that controls the disk module. The disk module is a business volume used for business and a mirror volume of the business volume, and is issued by a disconnection command from the host computer. Separately, includes a copy volume that can be accessed from the strike computer and can be recombined, and when the host computer issues a command to disconnect the transaction volume and the copy volume, the transaction volume and the copy volume An update history table that describes the update history of the above is created.

A differential copy method for a disk array device according to a second aspect of the present invention is the differential copy method for a disk array device according to the first aspect, wherein the business volume and the copy volume are operated while being separated from each other. And a state transition table for determining the status of the copy volume, and determining whether an updated block of data from the host computer generated for the transaction volume and the copy volume and the data is in the cache memory, With an update history table for each volume,
When a remirror instruction is issued from the host computer, the update difference is determined by taking the logical sum of the update history tables of the transaction volume and the copy volume.

According to a third aspect of the present invention, in the differential copy method of the disk array device according to the second aspect, the status transition table shows that the transaction volume and the copy volume are mirrored to each other. A state transition table indicating that the transaction volume and the copy volume are accessible independently of each other when the transaction volume and the copy volume are separated by a disconnection command from the host computer. It is composed of a state transition table to be displayed and a state transition table to display a state in which one becomes a master volume and one becomes a slave volume during the restoration operation of the transaction volume and the copy volume, and the re-mirror operation is performed. It is characterized by having been done.

According to a fourth aspect of the present invention, in the differential copy method of the disk array device according to the second or third aspect, the transaction volume and the copy volume are separated from each other in the update history table. While writing, a block address indicating the address of a block where data update has occurred from the host computer, and writing “1” to the block where data update has occurred from the host computer and writing “0” to the other blocks Write a flag and '1' while there is data in the cache memory, transfer the data from the cache memory to the disk director,
And a cache flag that is rewritten to '0' corresponding to the corresponding block.

According to a fifth aspect of the present invention, in the differential copy method for a disk array device according to the fourth aspect, the update history table includes a host director for each separated transaction volume and copy volume. It is characterized by being created by.

According to the differential copy method of a disk array device according to the present invention, a master volume and a slave volume are individually mounted on each of two physically separated disk array devices, and each of the disk array devices has a master volume and a slave volume. The host director establishes a link using a communication protocol with the host computer, transmits and receives commands to and from the strike computer, and issues a command to disconnect the two disk array devices from the host computer. An update history table describing the update histories of the master volume and the slave volume is created, and the update difference is determined by taking the logical sum of the update history tables of the master volume and the slave volume.

[0014]

Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the differential copy method of the disk array device according to the first embodiment of the present invention. In the figure,
The disk array device 10 according to the first embodiment includes:
Host director 1 that sends and receives commands from the host
2, a cache memory 13 for temporarily storing data, a disk director 14 for controlling read / write to a disk, and a plurality of disk drives
(Redundant Arryas of Inexpensive Disks). The disk module 17 includes a transaction volume 18 used for the host transaction and a copy volume 19 that can be separated by a mirror volume of the transaction volume, is independently accessible from the host, and can be remirrored. In.

Normally, the copy volume 19 operates as a mirror of the business volume 18 and cannot be accessed from the host. However, when a disconnection command is issued from the host, the copy volume 19 becomes independently accessible from the host, and the copy volume again becomes a mirror of the transaction volume by a remirror command from the host, and the copy volume 19 is returned from the host. Independent access is no longer possible.

FIG. 2, FIG. 3, and FIG. 4 are state transition tables of all volumes included in the disk module 17 in the embodiment of the present invention. FIG. 2 is a state transition diagram when the transaction volume 18 and the copy volume 19 are in the mirror state, and the mirror partner is the partner volume as described in the remarks column. FIG.
FIG. 8 is a state transition diagram showing a state in which the business volume 18 and the copy volume 19 are separated by a disconnection command from the host, and both volumes can be accessed independently. In this case, when the host issues the remirror command, the master volume and the slave volume are specified. FIG. 4 is a state transition diagram of the transaction volume 18 and the copy volume 19 during the restore operation. In this case, the transaction volume 18 becomes the master volume, the copy volume 19 becomes the slave volume, and the remirror operation is performed. It indicates that.

FIG. 5 shows a volume update history table according to the embodiment of the present invention. This table contains the host director 1
2 is made up of a block address 50, an update flag 51, and a cache flag 52. While the transaction volume 18 and the copy volume 19 are separated from each other, “1” is written to the update flag 51 in a block in which data has been updated by the host. Write 1 'in advance. After data is transferred from the cache to the disk, the cache flag 52 of the corresponding block is changed to '0'.

FIG. 6 shows the state after the remirror instruction is issued.
9 is a flowchart illustrating a process when a host accesses data to the disk array device 10 during a normal operation. FIG. 7 is a flowchart showing a process that occurs inside the disk array device 10 when a normal operation is being performed and there is no access from the host after the remirror instruction is issued.

Next, the operation of the disk array device 10 according to the first embodiment shown in FIG. 1 will be described. When the transaction volume 18 and the copy volume 19 are in a mirror state, the host director 12 writes a state transition table as shown in FIG. At this time, the host cannot access the copy volume 19. However, when the host issues a disconnect command, the host director 1
2 is rewritten into a state transition table as shown in FIG. 3, the copy volume 19 becomes accessible from the host,
At the same time, the host director 12 creates and manages an update history table as shown in FIG. 5 for each volume.

When the disconnect instruction is issued, the host director 12 rewrites the state transition table as shown in FIG. 3, and updates the update flag 5 for the block updated by the host.
1 to “1”, and the data is stored in the cache memory 13.
While the cache flag 52 is set to “1”.
To While disconnected, backup can be performed from a copy volume 19 having the same contents as the business volume 18, and batch processing and the like can be performed. On the other hand, the business volume 18 can execute normal online business.

When the re-mirror command is issued, the host director 12 updates the update flag 51 of the business volume 18.
And the update flag 51 of the copy volume 19 are logically ORed. Further, the contents of the cache flag 52 of the volume designated as the master volume by the host are transferred to the update history table shown in FIG. At this time, the contents of the cache flag 52 on the side designated as the slave volume are discarded. At this point, the disk array device 10
Issues a normal end notification for the re-mirror instruction to the host.

Then, upon the normal end of the re-mirror instruction, the host can resume normal business. At this time, the flow of processing when data access from the host to the disk array device 10 occurs will be described with reference to the flowchart of FIG. When the host accesses the data (Step S60; hereinafter, the steps are omitted), the host director 12 checks whether or not access to the master volume is possible in the state transition table of FIG. 3 (S61), and the host director 12 must be the master volume. If the access is impossible (S61, No), the normal operation is performed (S62). on the other hand,
If the access results in the master volume (S
61, Yes), it is checked whether there is a Read command (S63), and in the case of a Read command (S63,
Yes), normal read for master volume
A command operation is executed (S64).

Thereafter, it is checked whether or not the update flag 51 of the corresponding block is "1" by using the update history table of FIG. 5 (S65), and if it is "1" (S65, Ye).
s) Write the read data to the corresponding block of the slave volume (S66), and update the corresponding update flag 5
1 is reset to '0' (S69), and the processing ends. on the other hand,
If it is not a Read command in step S63 (S63, No), based on the Write command (S
67), the data is written to both the master volume and the slave volume (S68), and the update flag 51 of the corresponding block is rewritten to “0” in the update history table of FIG. 5 (S69), and the process ends.

Next, a process in the disk array device 10 in the case where there is no access from the host after the normal end of the restore command will be described with reference to the flowchart of FIG.
That is, from the block N = 0 of the update history table of FIG. 5 created by the logical sum of the master volume and the slave volume (S70), it is checked whether the update flag 51 is "1" (S71), and the update flag is set to "1". If it is 0 '(S71, No), there is no update, so block N
Is incremented by one (S76).

On the other hand, in step S71, the update flag 51
Is "1" (S71, yes), it is checked whether the cache flag 52 is "1" (S72).
If it is '1' (S72, Yes), the data in the cache is written to both the master and slave volumes (S73), and the update flag 5 in the update history table of FIG.
1 is reset to '0' (S75). If the cache flag is “0” in step S72 (S7
2, No), the data of the corresponding block is read from the master volume and written to the corresponding block of the slave volume (S74). Then, the update flag 51 is reset to '0' (S75).

In any case, the process proceeds to step S
When the update flag 51 is reset to "0" at 75, one block N is added (S76), and it is determined whether or not the result does not exceed the maximum block number (S77). If the number of blocks has been reached (S77, No), the process ends (S77, Yes).
s), returning to step S71, and thereafter, repeating the same steps. During this time, if there is access from the host, the process is suspended once, the flowchart of FIG. 6 is executed, and then the flowchart of FIG. 7 is restarted. In the flowchart of FIG. 7, when the block address exceeds the maximum block address in step S77, all the updated contents of the master volume are copied to the copy volume 19, and the master volume and the copy volume 19 are in a mirrored state. The access management using the update history table ends.

Next, a second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 8 is a block diagram showing a configuration of a differential copy method of the disk array device according to the second embodiment of the present invention. In FIG. 8, a master volume 90 and a slave volume 84 are mounted on physically separated disk array devices 86 and 80, respectively. Also, the disk array device 8
A link 85 using a communication protocol with the host is established between the host director 81 of the disk array device 86 and the host director 87 of the disk array device 86.

In this configuration, for example, it is assumed that an old disk array device 86 is replaced by a new disk array device 80. After suspending the operation and disconnecting the disk array device 86 from the host, the disk array device 80 is connected to the host, and the disk array device 80 and the disk array device 86 are connected by the link 85. 8
1 creates the update history table shown in FIG. 5 in which the update flags 51 of all the blocks are “1” and the cache flags 52 are “0”, and starts normal business with the host.

When there is no access from the host, the data of the master volume 90 is copied to the slave volume 84 via the link 85 according to the flowchart of FIG. At this time, if there is access from the host, it follows the flowchart of FIG. However, in FIG.
In the step of writing the write command in step S67, data is written to both the master volume and the slave volume as in step S68. However, in the case of the configuration of the second embodiment shown in FIG. Step S68 is read so that data is written only to the slave volume.

When there is no access from the host, the process follows the flowchart of FIG. In the flowchart of FIG. 7, when the maximum block address is reached (S77,
No), all of the contents of the master volume 90 of the old product have been copied to the slave volume 84 of the new product, and the link 85 connecting the disk array device 80 and the disk array device 86 is removed, and the disk array is removed. Device 86 can be removed. Thus, the replacement of the old product disk array device 86 and the new product disk array device 80 is completed only by interrupting the business only during the connection with the host and the table creation time.

The embodiment described above is an example for explaining the present invention, and the present invention is not limited to the above embodiment, and various modifications are possible within the scope of the invention. Of course there is.

[0032]

As described above, according to the differential copy method of the disk array device of the present invention, the update history table for managing the update of the transaction volume and the copy volume being detached is provided, and the update history table is updated when the remirror is executed. By taking the logical sum of the update flags in the history table, the update difference is determined. Then, since only the data of the update difference is copied from the master volume to the slave volume in the background, only necessary minimum data is transferred in the background, and the influence on the host is minimized. Therefore, after operating the transaction volume and copy volume separately, the update difference between the two volumes is calculated for re-mirroring, and the minimum necessary data is copied, thereby minimizing the effect on the response to the host. Can be suppressed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a differential copy method of a disk array device according to a first embodiment of the present invention.

FIG. 2 is a state transition diagram when a transaction volume and a copy volume are in a mirror state in the embodiment of the present invention.

FIG. 3 is a state transition diagram showing a state in which a transaction volume and a copy volume are separated by a disconnection command from a host, and both volumes can be accessed independently.

FIG. 4 is a state transition diagram during a restore operation of a transaction volume and a copy volume.

FIG. 5 is a volume update history table according to the embodiment of the present invention.

FIG. 6 is a flowchart illustrating a process when a host accesses data to a disk array device during a normal operation after a remirror instruction is issued.

FIG. 7 is a flowchart illustrating a process that occurs inside the disk array device when a normal operation is being performed and there is no access from the host after a remirror instruction is issued.

FIG. 8 is a block diagram showing a configuration of a differential copy method of a disk array device according to a second embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 Disk array apparatus 11 Communication path with host 12 Host director 13 Cache memory 14 Disk director 15 Internal bus 16 Communication path between disk director and disk module 17 Disk module 18 Business volume 19 Copy volume 50 Address 51 Update flag 52 Cache flag 80 New disk array device 81 Host director 82 Cache memory 83 Disk director 84 Slave volume 85 Link between new disk array device and old disk array device 86 Old disk array device 87 Host director 88 Cache memory 89 Disk director 90 Master volume

Claims (6)

[Claims] The present invention relates to a differential copy method for a disk array device that manages data by backing up data used in business to a copy volume that can be re-mirrored, and a host director that transmits and receives commands from a host computer. A cache memory for temporarily storing data; a disk module in which a plurality of disks constitute a RAID; and a disk director for controlling the disk module. The disk module comprises: a business volume used in business; A copy volume that is a mirror volume of the transaction volume and that can be separated from the host computer by a disconnection command and that can be accessed from the host computer and that can be recombined. Generating an update history table describing an update history between the business volume and the copy volume when a command to disconnect the business volume and the copy volume is issued from the disk array device. method. 2. The differential copy method for a disk array device according to claim 1, wherein the status of the transaction volume and the copy volume is determined while the transaction volume and the copy volume are operated separately. A state transition table for determining whether or not the updated block of data from the host computer and the data generated for the transaction volume and the copy volume are present in the cache memory; A disk array comprising: an update history table; and when a remirror instruction is issued from the host computer, a logical sum of the update history tables of the transaction volume and the copy volume is determined to determine an update difference. Device differential copy method. 3. The differential copy method for a disk array device according to claim 2, wherein the state transition table indicates that the transaction volume and the copy volume are in a mirror state to each other. A table indicating that the transaction volume and the copy volume are independently accessible when the transaction volume and the copy volume are separated by a disconnection command from the host computer. During the restore operation of the transition table and the transaction volume and the copy volume, one of the transition volumes becomes the master volume,
A state transition table for displaying a state in which one of the slave volumes becomes a slave volume and performs a remirror operation. 4. The differential copy method for a disk array device according to claim 2, wherein the update history table includes information from the host computer while the transaction volume and the copy volume are separated. A block address indicating the address of the block in which data has been updated, an update flag for writing "1" from the host computer to the block in which data has been updated, and "0" for other blocks, and data in the cache memory. And a cache flag for writing data "1" while the data is in the cache memory, transferring data from the cache memory to the disk director, and then rewriting the data to "0" in accordance with the corresponding block. Device differential copy method. 5. The differential copy method for a disk array device according to claim 4, wherein said update history table is created by said host director for each of said separated transaction volume and said copy volume. Differential copy method of the disk array device to be used. 6. A master volume and a slave volume are individually mounted on each of two physically separated disk array devices, and the host director of each of the disk array devices is:
A link using a communication protocol with the host computer is established to transmit and receive commands to and from the host computer. When the host computer issues a command to disconnect two disk array devices, the master An update history table that describes an update history between a volume and the slave volume is created, and an OR of the update history tables of the master volume and the slave volume is determined to determine an update difference. Differential copy method.