JP2006106883A - Computer system and remote replication method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the amount of data transferred from a master LU to a backup LU. <P>SOLUTION: This computer system includes a master LU 21a-1 for executing necessary processing according to requests from a master host computer; a first backup LU 21b-1 for remotely replicating master LU data; and a second backup LU 21b-2 for replicating the data of the first backup LU. When the first backup LU is recovered from closure due to a failure, all the data in the second backup LU 21b-2 are copied into the first backup UL recovered, and an amount of data corresponding to the amount that first difference information D1 not copied in the second backup LU is merged with second difference information D2 is remotely replicated from the master LU 21a-1 to the first backup LU 21b-1. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a computer system and a remote replication method for replicating data of a master logical unit of a certain storage device to a backup logical unit of another storage device installed at a remote site by a remote replication function.

  In general, in a computer system including an external storage device, data in the external storage device is appropriately backed up to another storage device. A remote replication function is used during backup processing from the external storage device to another storage device. Here, the remote replication function refers to the contents of a master logical unit (hereinafter referred to as master LU (LU: logical unit)) of a certain storage device at a certain point in time as a backup logical unit (hereinafter referred to as backup LU) of another storage device. It is a function to copy. The main use of the replicated data (snapshot data) is data backup. The reason for installing another storage device for backup at a remote site that is sufficiently distant from such a storage device is to prevent important data from being lost in the event of a disaster such as a fire, earthquake, or flood. Therefore, the availability of the computer system can be improved. The master LU refers to an LU (logical disk) from which data requested / written by the application is read / executed by execution of the application operating on the host computer. The backup LU is an LU (logical disk) that holds a copy of the master LU data. In order to realize the remote replication function and avoid the loss of master logical LU data in the event of a disaster, a backup LU having the same size as the master LU is prepared at a remote site different from the master LU.

Conventionally, the following two types of methods are used for collecting snapshot data.
The first data collection method synchronizes the contents of the master LU and the backup LU until the snapshot data collection is instructed. Normally, the computer system is always operating. Therefore, for example, when the host computer gives an instruction to collect snapshot data after the end of business, the backup LU is disconnected from the master LU and synchronized. Stop. Thereafter, the host computer reads / writes data only to the master LU, while snapshot data is collected from the master LU to the backup LU in response to a snapshot data collection instruction. After the synchronization is stopped, the difference information between the data written in the master LU and the snapshot data collected in the backup LU is managed. For example, when resynchronization is performed due to the start of a job, the master LU based on the difference information is managed. By writing the difference data to the backup LU, it is not necessary to copy the entire contents of the master LU, and resynchronization can be performed in a short time.

  In the second data collection method, after receiving the snapshot data collection instruction, the master LU data and the backup LU data are synchronized. Prior to this synchronization, the master LU differential data is copied to the backup LU based on the difference information. If a new write occurs to the master LU at this time, the old data read from the master LU → the old data to the backup LU. Need to be processed in the order of writing → writing new data to the master LU.

  The technical content of the present invention can be said to be a technique similar to the first data collection method among the two types of data collection methods. In this first data collection method, after a backup LU from which snapshot data has been collected is backed up to an external magnetic tape, for example, a resynchronization process is required to make the contents of the backup LU coincide with the master LU again. . In this resynchronization, only the difference data of the master LU is copied to the backup LU based on the difference information, and the data transfer amount is suppressed.

  By the way, in the first data collection method, if the HDD of the backup LU fails due to various factors, a so-called LU blockage state occurs in which the data of the backup LU cannot be used. In this case, all data must be copied from the master LU to the backup LU for resynchronization after the backup LU is closed. However, in remote replication, since the line capacity between the master LU and the backup is not large, there is a problem that the burden is too large to transfer a large amount of data from the master LU.

  Therefore, it is conceivable to add another backup LU with the backup LU as a master to the storage device installed at the remote site for backup. In other words, the configuration is such that a plurality of backup LUs are provided in the storage device at the remote site.

  However, as described above, a system in which a backup storage device is installed at a remote site sufficiently away from the master storage device and a plurality of backup LUs are provided in the backup storage device has already been proposed ( Patent Document 1).

In this system, the master LU in the storage device that is the master and the first backup LU in the backup storage device are always synchronized, and the contents of the first backup LU are stored in the backup storage device. Copy processing to the second backup LU. In this state, if a file in the main LU is lost due to some reason, the contents of the second backup LU are put on the secondary host, the file to be recovered on the secondary host is copied to the first backup LU, and the first The backup LU is used as a master, resynchronized with the master LU in the storage apparatus, and necessary files are recovered.
JP 2003-233518 A (see FIG. 2)

  However, the technique of Patent Document 1 described above is a process of copying only a file to be recovered from the backup LU by resynchronization when a file in the master LU is lost due to some factor. Therefore, this is a process for recovering data in units of file levels in the logical disk, and the amount of data transfer does not matter so much.

  However, if all data in the master LU is lost and all data is required, copy all data from the second backup LU to the first backup LU, and then copy all data from the first backup LU to the master LU. Need to copy. In general, in remote replication, the line capacity between the master storage device and the backup storage device is not large. Therefore, when all data is copied by resynchronization, transfer of large-capacity data becomes a heavy burden as described above.

  The present invention has been made in view of the above circumstances, and a computer system and a remote replication that reduce the amount of data transferred from the master LU to the backup LU and reduce the burden on the transmission line when recovering after the backup LU is closed. It aims to provide a method.

(1) In order to solve the above-described problem, a computer system according to the present invention includes a first storage device having a master LU that executes necessary processing based on a request from a master host computer, and data of the master LU. A first backup LU for remote replication and a second storage device at a remote site having the first backup LU as a master and a second backup LU for replicating the data of the first backup LU. A computer system,
The second storage device performs remote replication of data of the master LU to the first backup LU and replicates data of the first backup LU to the second backup LU. When one backup LU is blocked due to a failure and then recovered, all data copy means for copying all data of the second backup LU to the recovered first backup LU, and the first backup LU The second difference information obtained from the data update position of the first backup LU and the data copy position of the second backup LU at the time of the block is transmitted to the first storage device, and as a result, the second The replenishment data of the master LU returned from the storage device is the first And a data supplement recovery means for copying replenished backup LU provided for,
Based on the recovery of the first backup LU, the first storage device starts from the data update position of the master LU and the copy position remote-replicated from the master LU to the first backup LU before the blocking. The first backup information is calculated from the master LU based on information obtained by merging the first difference information with the second difference information sent from the second storage device. Both difference information merge processing means for remotely replicating the supplementary data to the LU is provided.

  Therefore, according to the present invention, after the recovery by exchanging the first backup LU, etc., after copying all the data of the second backup LU to the first backup LU, the above configuration is used. Further, remote replication is performed from the master LU to the first backup LU based on the merged information of the second difference information obtained on the second storage device side and the first difference information obtained on the first storage device. Since the data is replenished, the amount of data transferred from the master LU to the first backup LU can be reduced by the total amount of data in the second backup LU, thereby reducing the load on the transmission line by reducing the amount of data transferred. It is possible to enter the resynchronization promptly.

  The first storage device includes a synchronization processing unit that executes synchronization processing between the master LU and the first and second backup LUs, and the master processing is performed in accordance with the abolition of synchronization processing by the synchronization processing unit. Remote replication processing means for remotely replicating LU data to the first backup LU, and reading of data to the master LU by execution of an application on the master host computer after the failure due to failure of the first backup LU Data read / write execution means for executing / write, and the data update position of the master LU at the time of the remote replication processing including the time of execution of the read / write of the data and the first from the master LU until the blockage Remotely replicated copy to other backup LU If newly added r and history data recording means for recording the location, it can be possible to calculate the first difference information.

  The second storage device includes a replication processing means for replicating the data of the first backup LU to the second backup LU during the remote replication processing, and the first backup LU during the replication processing. If new history data recording means for recording the data update position and the data copy position of the second backup LU is added, the second difference information can be calculated.

(2) The present invention also provides a first storage device having a master LU that executes necessary processing based on a request from a master-side host computer, a first backup LU that remotely replicates data of the master LU, and In a computer system comprising a second storage device at a remote site having a second backup LU for replicating data of a first backup LU,
When the master LU data is remotely replicated to the first backup LU and the data of the first backup LU is replicated to the second backup LU, the first backup LU is blocked due to a failure. Then, in the case of subsequent recovery, an all data copy step for copying all the data of the second backup LU to the recovered first backup LU, and the first backup at the time when the first backup LU is closed Based on the step of transmitting the second difference information obtained from the data update position of the LU and the data copy position of the second backup LU to the first storage device, and the recovery of the first backup LU, The data update position of the master LU and the master LU before the block The first difference information with respect to the copy position remote-replicated to the first backup LU is calculated, and the first difference information is merged with the second difference information sent from the second storage device. The step of remotely replicating the supplementary data from the master LU to the first backup LU based on the information of the processing result, and the supplementary data of the master LU sent by this remote replication as the first backup A remote replication method having a copy supplement to an LU.

  This remote replication method also has the same effect as the computer system (1).

  According to the present invention, when data is recovered after being blocked due to a failure of the first backup LU, all data of the second backup LU is copied, so the amount of data transferred from the master LU to the backup LU can be reduced, and the transmission line It is possible to provide a computer system and a remote replication method that can reduce the burden.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a configuration diagram showing an embodiment of a computer system according to the present invention. This computer system includes a master host computer 10a and a backup host computer 10b for executing various applications at a remote site from the viewpoint of avoiding data loss due to disasters such as earthquakes, fires, and floods. . These host computers 10a and 10b are connected to storage devices 20a and 20b as external storage devices, respectively. Both storage apparatuses 20 a and 20 b are connected by a transmission line 30.

  The storage device 20a connected to the master-side host computer 10a includes a disk device 21a and a storage controller 22a. The disk device 21a is composed of, for example, one logical disk device 21a-1. The logical disk device 21a-1 is configured using a disk storage device, for example, a hard disk drive (HDD). A logical unit (logical unit) is allocated to the logical disk device 21a-1. The logical unit assigned to the logical disk device 21a-1 is a master logical unit (hereinafter referred to as a master LU) from which data is read / written according to the application when the master-side host computer 10a executes the application. Used. Therefore, hereinafter, the logical disk device 21a-1 is referred to as a master LU. This master LU can also be said to be a master LU for remote replication.

  On the other hand, the storage device 20b connected to the backup-side host computer 10b is composed of a disk device 21b and a storage controller 22b, like the storage device 20a. The disk device 21b includes, for example, two logical disk devices 21b-1 and a logical disk device 21b-2. The logical unit assigned to the logical disk device 21b-1 is used as a backup logical unit (hereinafter referred to as a first backup LU) that replicates and holds the data of the master LU 21a-1 by remote replication. Therefore, it can be said that the first backup LU is a remote replication backup LU.

  The logical unit assigned to the logical disk device 21b-2 is a backup logical unit (hereinafter referred to as a second backup backup unit) that uses the first backup LU as a master LU and replicates and holds the data of the first backup LU by replication. Called LU).

  Next, the master storage controller 22a performs access control to the master LU 21a-1 in response to a request from the master host computer 10a. The storage controller 22a is provided with a microprocessor 23a that constitutes a main control unit of the controller 22a, firmware 24a that is a control program executed by the microprocessor 23a, and a difference table 25a. The microprocessor 23a includes a backup task for performing backup (snapshot collection) by remote replication from the master LU 21a-1 to the storage device 20b connected to the backup side host computer 10b in accordance with the firmware 24a. In the figure, 27a and 28a are ports.

  The difference table 25a is a bitmap format table composed of bits associated with each logical block address area of the master LU 21a-1. In the difference table 25a, the bit associated with the logical block address area when data is updated only in the master LU 21a-1 is set to flag ON.

  On the other hand, the backup-side storage controller 22b performs access control to the backup LU 21b-1 and the backup LU 21b-2 in response to requests from the master-side host computer 10a and the backup-side host computer 10b. The storage controller 22b is provided with a microprocessor 23b constituting a main control unit of the controller 22b, firmware 24b which is a control program executed by the microprocessor 23b, and a difference table 25b. Note that the microprocessor 23b includes a backup task for performing backup (collecting a snapshot) from the first backup LU 21b-1 to the second backup LU 21b-2 according to the firmware 24b. Reference numerals 27b and 28b denote ports.

  The difference table 25b is composed of a bit map format table composed of bits associated with each logical block address area of the first backup LU 21b-1. In the difference table 25b, the table bit associated with the data update position in the logical block address area where the data of the first backup LU 21b-1 has been updated is set to ON.

The master side storage controller 22a is functionally configured as shown in FIG.
That is, in the storage controller 22a, the microprocessor 23a performs synchronization processing between the master LU 21a-1 and the first and second backup LUs 21b-1 and 21b-2 according to the firmware 24a based on an application request from the master-side host computer 10a. Synchronous processing means 221a to be executed, and data read / write execution means 222a to read / write data to the master LU by executing an application of the master host computer 10b after the first backup LU 21b-1 is blocked due to failure During the above-described remote replication processing from the master LU 21a-1 to the first backup LU 21b-1, and after the first backup LU 21b-1 is blocked, data is written from the master-side host computer 10b to the master LU 21a-1. A history data recording unit 223a for recording the data update position of the master LU 21a-1 with the bit flag ON and a difference information merging unit 224a are provided in the difference table 25a based on the data to be stored. Accordingly, the first difference information can be calculated from the bit flag position of the difference table 24a based on the recording by the history data recording means 223a. The difference information merge processing means 224a merges the second difference information, which will be described later, transmitted from the backup side storage apparatus 20b after the recovery of the first backup LU 21b-1, and the merged first and first information 2 has a function of remotely replicating the data of the master LU 21a-1 to the first backup LU 21b according to the difference information of 2.

On the other hand, the storage controller 22b on the backup side is functionally configured as shown in FIG.
On the basis of an application request from the backup-side host computer 10b, the microprocessor 23b executes synchronization processing between the master LU 21b-1 and the first and second backup LUs 21b-1, 21b-2 according to the firmware 24b, Based on the data written to the first backup LU 21b-1 at the time of the replication process, the data update position of the first backup LU 21b-1 is recorded by the bit flag ON in the difference table 25b, and the second in the copy completion table 26b. When the history data recording means 222b for recording the copy position of the backup LU 21b-2 and the recovery or replacement of the first backup LU 21b-1 are recovered, all the data of the second backup LU 21b-2 is restored to the first backup LU 21b-2. All data copy means 223b for copying to the copy LU 21b-1 and data supplement recovery means 224b are provided. The data supplement recovery means 224b calculates the second difference information from the data update position of the difference table 25b and the copy position of the copy completion table 26b, and transmits the second difference information to the master storage controller 22a. Based on the difference information obtained by merging the second difference information and the first difference information, the first backup LU is supplemented with data that is remotely replicated from the master LU 21a-1.

  Next, the operation of the computer system as described above and the remote replication method according to the present invention will be described with reference to FIGS.

(1) About synchronization processing.

  The master-side host computer 10b makes the contents of the master LU 21a-1 and the contents of the backup LUs 21b-1, 21b-2 coincide with the microprocessors 23a, 23b of the storage apparatuses 20a, 20b by executing the application. Requests control, that is, synchronization processing. Here, the storage apparatus 20a executes the synchronization process shown in FIG. 4 (S1). Further, both the backup-side host computer 10b and the storage apparatus 20a receive a request for synchronization processing via the microprocessor 23b, and issue an instruction for synchronization processing to the microprocessor 23b as shown in FIG. A synchronization process is executed (S21). These steps S1 and S21 correspond to synchronization processing means 221a and synchronization processing steps. In synchronization processing, data writing to the master LU 21a-1 → data writing to the first backup LU 21b-1 → second backup LU 21b-2 in response to data writing from the master-side host computer 10a to the disk devices 21a and 21b. Write data to the. The write data is written to the same logical block address of each LU 21a-1, LU 21b-1, and LU 21b-2. When the data writing to the second backup LU 21b-2 is completed, the write completion data is sent to the master side host computer 10a through the microprocessors 23b and 23a. This synchronization processing is continued for an arbitrary time based on the business administrator's business and other requests. (A) in FIG. 6 represents the synchronization processing state. However, in the present invention, the synchronous process is not necessarily an essential requirement, and the master LU 21a-1, the backup LUs 21b-1, 21b-2 and the asynchronous process may be involved.

(2) About replication processing due to the abolition of synchronous processing.

  When the master-side host computer 10a issues an instruction to collect snap shot data by executing an application based on the business administrator's business or other requests, the microprocessor 23a determines that the synchronization processing is abolished, and the snap shot data Is collected (S2). The microprocessor 23a transmits the snap shot collection instruction data to the backup side host computer 10b through the transmission line 30 based on the collection instruction (S3).

  On the other hand, when the backup-side host computer 10b receives the snap shot collection instruction data, the backup-side host computer 10b determines that the second backup LU 21b-2 is to be disconnected (S22), and the first backup LU 21b-1 receives the first backup LU 21b-1 via the microprocessor 23b. A process for disconnecting the second backup LU 21b-2 is executed (see (b) of FIG. 6).

  On the other hand, the storage controller 22b of the backup-side host computer 10b executes data writing to the first backup LU 21b-1. (S23). At this time, the storage controller 22b turns on the bit of the difference table 25b corresponding to the logical block address area when data is updated only to the first backup LU 21-1 by remote replication (S24).

(3) Regarding blocking of the first backup LU 21b-1.

  By the way, when the remote replication process and the remote replication process are executed as described above, if the first backup LU 21b-1 is blocked due to a failure as shown in FIG. 6C, the first backup LU 21b-1 is concerned. Will not be available. When the storage controller 22b determines that the LU 21b-1 is blocked (S25), the replication process is terminated, the LU block is notified to the master storage controller 22a through the transmission line 30, and the process is terminated. At this time, the difference table 25b holds the bit ON state when the LU 21b-1 is closed. Here, restoration work such as repair and replacement of the first backup LU is performed by maintenance personnel.

  On the other hand, the master-side host computer 10a continues to read / write data to / from the master LU 21a-1 (S4: corresponds to the data read / write execution means 223a). As a result, each time data is updated to the master LU 21a-1, the storage controller 22a sets the bit of the difference table 26a corresponding to the logical block address area of the master LU 21a-1 in which the data is written to the flag ON. (S5). The copy completion table 27a holds the bit ON state when the LU 21b-1 is closed for reading / writing of data to the master LU 21a-1. Step S9 corresponds to the history data recording means 224a.

(4) Data recovery processing to LU 21b-1.

  When the first backup LU 21b-1 of the storage device 20b is recovered by restoration or replacement, the backup-side host computer 10b executes, for example, a process as shown in FIG. 6B. That is, the backup side host computer 10b requests the storage controller 22b to start the operation. The storage controller 22b determines whether or not the first backup LU 21b-1 can be normally used according to the control program of the firmware 24b (S31). When it is determined that the storage controller 22b can be used, the LU recovery data is transferred to the master via the transmission line 30. This is notified to the host computer 10a (S32). Thereafter, as shown in FIG. 6D, the storage controller 22b copies all the data stored in the second backup LU 21b-2 to the first backup LU 21b-1 (S33: all data copy). Means 223b, corresponding to all data copying step). Therefore, when data is restored to the first backup LU 21 b-1, a part of the data of the second backup LU 21 b-2 can be recovered by the internal copy process without using the transmission line 30.

  Further, the storage controller 22b calculates difference information D2 from the bit flag position of the difference table 25b and the bit flag position (copy completion position) of the copy completion table 26b (S34). The difference information D2 is the data write update of the first backup LU 21b-1 after the second backup LU 21b-2 is disconnected from the first backup LU 21b-1 until the LU 21B-1 is closed based on the abolition of the synchronization process. It can be said that this is the logical block address area until the data copy to the second backup LU 21b-2 is completed. For example, when represented by a bitmap table, as shown in FIG. 7F, a black bit in the bitmap table can be represented by a flag ON.

  Therefore, when the storage controller 22b calculates the difference information D2, the storage controller 22b transmits the difference information D2 to the storage controller 22a through the transmission line 30 (corresponding to S35: difference information transmission step).

  On the other hand, the storage controller 22a receives the LU recovery data from the storage controller 22b and determines whether or not the first backup LU 21b-1 is resynchronized (S6), and then the difference information D2 is sent from the storage controller 22b. Wait for it to come (S7). Here, when receiving the difference information D2, the storage controller 22a calculates the difference information D1 from the bit flag position of the bit flag position of the difference table 25a (S8). If this difference information D1 is expressed by, for example, a bitmap table, as shown in FIG. 7G, black bits in the bitmap table can be expressed by a flag ON.

  When the storage controller 22a acquires the difference information D2 and D1 in this way, the storage controller 22a merges the difference information D2 and D1 (S9), and then performs a synchronization process (S10).

  The storage controller 22b on the backup side copies the data corresponding to the difference information (D2 + D1) of the master LU 21a-1 as supplementary data to the first backup LU 21b-1 as shown in FIG. 6 (e) (S36: Data Equivalent to a replenishment step).

  Therefore, data (h) corresponding to the sum of the bit flags (f) and (g) in FIG. 7 is copied to the first backup LU 21b-1. That is, all the data of the master LU 21a-1 is copied to the first backup LU 21b-1, and resynchronization processing is possible.

  Therefore, according to the embodiment as described above, when the first backup LU 21b-1 is closed due to a failure and then recovered by repair or replacement, all data already copied to the second backup LU 21b-2. Is copied to the first backup LU 21b-1 and is not copied to the second backup LU 21b-2, but the data written to the master LU 21a-1 is the same as the first difference information D1. Since the amount of data that has been merged with the second difference information D2 may be remotely replicated from the master LU 21a-1 to the first backup LU 21b-1, the amount of data transferred from the storage controller 22a can be set to the second backup. All data already copied to LU21b-2 can be reduced The Therefore, the burden on the transmission line 30 can be reduced, and the transmission line 30 can be used efficiently.

  In addition, this invention is not limited to the said embodiment, In the range which does not deviate from the summary, various deformation | transformation can be implemented. Moreover, each embodiment can be implemented in combination, and in that case, the effect of the combination can be obtained.

The block diagram which shows one Embodiment of the computer system which concerns on this invention. The functional block diagram which shows an example of the storage controller of the master side. The functional block diagram which shows an example of the storage controller of the backup side. The figure explaining operation | movement of the storage controller of the master side, and the remote replication method concerning this invention. The figure explaining operation | movement of the storage controller of a backup side, and the remote replication method concerning this invention. The figure which shows the change log | history of the computer system and remote replication method which concern on this invention. The figure explaining the data supplement example of the simulated bitmap as a result of merging processing of the 1st difference information D1 and the 2nd difference information D2.

Explanation of symbols

  10a ... Master side host computer, 10b ... Backup side host computer, 21a, 21b ... Disk device, 21a-1 ... Master LU, 21b-1, 21b-2 ... Backup LU, 22a, 22b ... Storage controller, 23a, 23b ... Microprocessor, 25a, 25b ... difference table, 221a ... synchronization processing means, 222a ... data read / write execution means, 223a ... history data recording means, 224a ... both difference information merge processing means, 221b ... synchronization processing means, 222b ... history data recording Means, 223b, all data copy means, 224b, data supplement recovery means.

Claims (6)

A first storage device having a master logical unit that executes necessary processing based on a request from the master host computer, a first backup logical unit that remotely replicates data of the master logical unit, and the first backup logic In a computer system comprising a unit as a master and a second storage device at a remote site having a second backup logical unit that replicates data of the first backup logical unit,
The second storage device
When the data of the master logical unit is remotely replicated to the first backup logical unit and the data of the first backup logical unit is replicated to the second backup logical unit, the first backup logical unit All data copy means for copying all data of the second backup logical unit to the recovered first backup logical unit when the unit is blocked due to a failure and then recovered; The second difference information obtained from the data update position of the first backup logical unit at the time of blocking is transmitted to the first storage device, and as a result, the master logic returned from the first storage device Before unit replenishment data And a data supplement recovery means for copying supplement provided on the first backup logic unit,
The first storage device is
Based on the recovery of the first backup logical unit, first difference information is calculated from the data update position of the master logical unit, and the first difference information sent from the second storage device is included in the first difference information. And a difference information merging processing means for remotely replicating the supplementary data from the master logical unit to the first backup logical unit based on information obtained as a result of merging the difference information. . In the computer system according to claim 1,
The first storage device is
Synchronization processing means for executing synchronization processing between the master logical unit and the first backup logical unit, and execution of an application of the master side host computer after shutdown due to failure of the first backup logical unit, the master logical unit A data reading / writing executing means for reading / writing data with respect to the data, and a history data recording means for recording a data update position of the master logic unit at the time of executing the data reading / writing, A computer system capable of calculating one difference information. In the computer system according to claim 1 or 2,
The second storage device
Synchronization processing means for executing synchronization processing between the first backup logical unit and the second backup logical unit, and the master logical unit and the first backup after the synchronization processing of the synchronization processing means is abolished And a history data recording means for recording a data update position of the first backup logical unit accompanying a synchronization process between the logical units, wherein the second difference information can be calculated. . In the computer system according to claim 1 or 2,
Both the difference information merge processing means of the first storage device,
A logical unit recovery determination unit that determines recovery of the first backup logical unit after the first backup logical unit is blocked; and when the determination unit determines that the first backup logical unit has recovered, Difference information calculation means for calculating second difference information from an update position updated from the master-side host computer recorded by the history data recording means after the backup logical unit is closed, and the first storage device When the second difference information is received from the merge processing means for merging the first difference information with the second difference information, and the master logical unit based on the information of the result of the merge processing by the merge processing means The replenishment data to the first backup logical unit from a remote replica Computer system and having a Deployment to unit. In the computer system according to claim 1 or 3,
The data supplement recovery means of the second storage device is:
When the first backup logical unit is blocked due to a failure, the second difference from the update position where the data is updated from the master logical unit recorded by the history data recording means to the first backup logical unit by the synchronization process. Difference information calculating means for calculating information, and means for transmitting the second difference information after transmitting logical unit recovery data when the first backup logical unit is recovered; Means for copying the supplementary data remote-replicated from the master logical unit to the first backup logical unit based on the merged result information after transmitting the difference information to the first storage device; A computer system characterized by being provided. A first storage device having a master logical unit that executes necessary processing based on a request from the master host computer, a first backup logical unit that remotely replicates data of the master logical unit, and the first backup logic A computer system comprising a second storage device at a remote site having a second backup logical unit for replicating unit data;
When the data of the master logical unit is remotely replicated to the first backup logical unit and the data of the first backup logical unit is replicated to the second backup logical unit, the first backup logical unit If the unit is blocked due to a failure and then recovered, an all data copy step of copying all data of the second backup logical unit to the recovered first backup logical unit; Based on the step of transmitting the second difference information obtained from the data update position of the first backup logical unit to the second storage device at the time of blocking, and the recovery of the first backup logical unit, the master logic Unit data First master information obtained from the update position is calculated, and the master logic is based on information obtained by merging the first difference information with the second difference information sent from the second storage device. Remote replicating the supplementary data from a unit to the first backup logical unit; copying supplementary data of the master logical unit sent by this remote replication to the first backup logical unit; A remote replication method characterized by comprising:

Images