JP2000187608A - Storage device sub-system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute data movement between new and old storage devices by file units without any interference of a host computer. SOLUTION: A new storage device sub-system 3 is provided with two interfaces 31 and 32, and connected with a host 1 and an old storage device sub-system 2. The new storage device sub-system 3 copies the data of a volume in the old storage device sub-system 2 to its own volume in the order of the head, and records the position at which the copy is completed in a copy pointer 43. Also, the new storage device sub-system 3 accepts access from the host 1 as if it is access from the old storage device sub-system 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a storage system in an information processing system or the like, and more particularly to a computer system having a function of transferring data between storage systems.

[0002]

2. Description of the Related Art A computer system used in a bank, a company, or the like handles an extremely large amount of data and requires a large-scale storage device. Furthermore, since such data increases daily, it is necessary to add storage devices or replace them with storage devices of larger capacity. When replacing a storage device, the usual means is to make a backup of the storage device (hereinafter referred to as the old storage device) on a magnetic tape, etc.
It is realized by restoring to.

However, in order to make a backup, it is necessary to temporarily stop the I / O of the host, and the stop time is generally long. The stop time increases as the capacity of the storage device increases. Furthermore, in the large-scale system as described above, 24-hour non-stop operation is premised, and this is the same as in the case of replacing the storage device. Therefore, the method of temporarily stopping the I / O on the host side and taking a backup is There's a problem.

In US Pat. No. 5,680,640, a method is described in which data is transferred online from an old storage device to a new storage device. Here, each address (track) of the volume in the old storage device is stored in the new storage device.
A table is provided for each track to determine whether data transfer from the old storage device to the new storage device has been completed.
If there is an I / O request from the host during migration, the operation is determined by referring to the table. For example, if there is a read request, it is checked by referring to the table whether the record (block) for which the read request has been transferred to the new storage device, and if the data from the old storage device has not been transferred, the old Read data from the storage device. If there is data in the new storage device, the data is read from the new storage device. When there is a write request, data is written to the new storage device, and the table is updated.

[0005]

In the method shown in U.S. Pat. No. 5,680,640, a table must be maintained for every track in the old storage device. Normally, a table obtained by making a semiconductor memory non-volatile by a battery backup or the like is used. However, a semiconductor memory having a very large capacity is not mounted on a storage device due to cost. When the data capacity of the old storage device is large, the table of the new storage device is correspondingly large, and the new storage device becomes very expensive.

In this method, since data is written only to the new storage device by writing from the host during data migration, if a failure occurs in one of the storage devices during the migration and the migration process stops, Both storage devices have data inconsistent states.

[0007]

In order to solve the above problems, a computer system according to the present invention has the following configuration.
A computer system according to the present invention includes a first storage subsystem, a second storage subsystem, and first and second storage subsystems.
And a host computer connected to the storage subsystem. The first storage subsystem has an interface for connecting to a host computer, and the second storage subsystem
Storage subsystem has a first storage subsystem and an interface for connecting to a host computer. The host computer is initially connected to the first storage and performs I / O processing. However, when performing data migration from the first storage subsystem to the second storage subsystem, the host computer once communicates with the host computer. The connection between the first storage subsystem is disconnected and the host computer is connected to the second storage subsystem.
Storage subsystem. This connection disconnection may be a state in which the connection is logically disconnected (offlined) instead of physically disconnecting the transmission path.

[0008] When the connection between the host computer and the first storage subsystem is cut off and the connection between the second storage subsystem is completed, the transfer is started. In the migration process, the second storage subsystem issues a read request to the first storage subsystem, and the read request is sent via the transmission path connecting the first storage subsystem and the second storage subsystem. To copy the data. In the second storage subsystem, there is a pointer that records how far the copy has been completed, so that the progress of the data migration can be known.

[0009] An I / O request from the host computer during migration is accepted by the second storage subsystem. When a read request is received from the host computer, the second storage subsystem refers to the copy pointer, checks whether the requested data already exists in the second storage subsystem, and checks the second storage subsystem. If so, it passes the data to the host computer. If not, the data is read from the first storage subsystem and passed to the host computer.
When a write request is issued, the second storage subsystem basically writes data to both the first and second storage subsystems. As a result, even if an I / O request is received from the host computer during the migration, the data in the storage subsystem does not become inconsistent, and the migration can be continued. If a write request is issued from the host computer during migration, data is duplicated in the first and second storage subsystems, and an error occurs in either storage subsystem. Even data is guaranteed.

In the above method, when an I / O request is received from the host to an area where data is not copied from the first storage subsystem during the migration processing, access to the first storage subsystem is restricted. Because of this, the processing speed does not increase.Therefore, the access frequency is counted for each disk of the second storage subsystem, statistics are obtained on which areas are accessed by the host. There is also a method of executing the transfer process preferentially from a high area. As a result, the copy of the frequently accessed area is completed early,
Even if an I / O request comes, access to the first storage subsystem can be reduced.

[0011] Further, in some cases, it is sufficient to migrate not only the entire volume, but also a specific area, or even a specific file. In that case,
There is also a method of migrating only an area where a specific file is recorded by referring to file management information in the second storage subsystem.

[0012]

(First Embodiment) FIG. 1 shows a configuration example of a computer system to which the present invention is applied. The computer system includes a host 1, an old storage subsystem 2, and a new storage subsystem 3. The old storage subsystem 2 is an interface 2 for connecting to the host 1
1, an interface 22 for connecting to the new storage subsystem 3, and a disk 23. The new storage subsystem 3 includes an interface 31 for connecting to the host 1, an interface 32 for connecting to the old storage subsystem 2, a disk 33, a cache memory 34, a control device 35, and a setting console 36. You. A plurality of disks 23 and disks 33 may be present in each device, respectively, or a disk array may be composed of a plurality of disks. Also, disk 2
3. The disk 33 does not necessarily have to be the same type of disk. For example, SCSI (Small Computer System Int.)
(Erface) In the case of a disk such as a disk that handles a fixed-length data format, the disk 33 only needs to have a larger capacity than the disk 23. In the case of a disk that handles data in the CKD (count key data) format, it is only necessary that the capacity of the disk 33 is larger than that of the disk 23 and that the formats (3380/3390 format, etc.) match.

The control unit 35 is generally composed of a CPU and a memory, and has a new storage subsystem 3 in the memory.
Exists and is executed, but is omitted in the drawings of the present invention. The main components of the present invention in the control device 35 include an I / O processing unit 41, a transition unit 42,
A copy pointer 43 and a command queue 44 exist.
The I / O processing means 41 receives a read / write request from the host 1 and performs processing. The migration unit 42 performs a process of moving data from the old storage subsystem 2 to the new storage subsystem 3. Copy pointer 43 is transferred to transfer means 42
To record how far the data migration process has progressed. The command queue 44 is sent from the host 1
Used to contain I / O requests. I / from host 1
The O request enters here once and is processed by the I / O processing means 41. The setting console 36 changes various settings of the new storage subsystem 3. For example, disk 33
Is made offline so that the host 1 can use it, and a processing instruction for migrating data from the old storage subsystem 2 to the new storage subsystem 3 is issued.

Referring to FIG. 2, a data migration process from the old storage subsystem 2 to the new storage subsystem 3 will be described.

First, the old storage subsystem 2 is connected to the host 1 via the transmission line 51, and the I / O with the host is performed.
It is carried out. When migrating data from the old storage subsystem 2 to the new storage subsystem 3, I / O of the host 1 is temporarily stopped and the connection of the transmission path 51 is cut off. Then, the new storage subsystem 3 sends a # to the host 1 via the transmission line 52.
The old storage subsystem 2 is connected by transmission 13 #,
State. In this operation, the connection between the host 1 and the old storage subsystem 2 is established by the host 1 and the new storage subsystem 3.
Since only the connection to the host is switched, I / O from the host only needs to be stopped for a short time.

With the configuration shown in FIG. 1, the transition processing can be started. First, the copy pointer 43 in the control device 35 is initialized to 0 (step 1001). In step 1002, the command queue is referred to and it is checked whether a read / write processing request has been received from the host 1 (step 1).
003). If there is a processing request, the processing in the queue is executed (step 1004). When there are a plurality of processing requests, various methods can be considered, such as executing all of the processing requests or executing only one of them. Here, only one is executed. This processing will be described later.

In step 1005, a request to read the area indicated by the copy pointer 43 is sent to the old storage subsystem 2. The read unit is FBA for disks 23 and 33
In the case of the format, it may be a multiple of one sector, and in the case of a CKD format disc, it may be a multiple of the track. In this embodiment, the disk is in FBA format, and the read unit is 1
Set for each sector.

The read unit is set to the console 3 for setting.
6 may be specified. This makes it possible to increase the read unit when the migration process is to be performed with a higher priority than the I / O process coming from the host, or conversely, to reduce the delay in the normal I / O process. Can be reduced.

At step 1006, read data is received from the old storage subsystem 2, and at step 1007, data is written to the new storage subsystem 3 at the relevant address. In step 1008, 1 is added to the copy pointer 43. At step 1009, it is checked whether the copy pointer 43 has reached the maximum number of blocks of the disk 23,
If it has not reached yet, the process returns to step 1002 and the process is repeated. If the maximum number of blocks has been reached, the transfer processing ends.

Next, the flow of processing in the new storage subsystem 3 when a read request is received from the host 1 during the migration processing will be described with reference to FIG. First, the read destination address from the host 1 is compared with the copy pointer 43 held in the new storage subsystem 3 (step 1).
101). If the copy pointer 43 is larger than the read destination address, the data is already copied from the old storage subsystem 2 to the new storage subsystem 3, so the data is read from the new storage subsystem 3 (step 1102), read processing can be performed by sending data to the host.

Conversely, at step 1101, copy pointer 4
If 3 is equal to or less than the read destination address, the new storage subsystem 3 does not reflect the contents of the old storage subsystem 2 and sends a read request to the old storage subsystem 2 (step 1103). ), And receives the read data from the old storage subsystem 2 (step 11).
04) The read process is completed by sending the data received from the old storage subsystem 2 to the host (step 1105).

Next, the flow of processing in the new storage subsystem 3 when a write request is received from the host 1 during the migration processing will be described with reference to FIG. In step 1201, first, a write request is issued to the old storage subsystem 2 to write the write data received from the host 1 to the old storage subsystem 2, and the data is written. In step 1202, the write destination address from the host 1 is compared with the copy pointer 43 held in the new storage subsystem 3. If the write destination address is smaller than the copy pointer 43, the write data from the host 1 is written to the disk 33. If the write destination address is equal to or greater than the copy pointer 43, the data is copied from the old storage subsystem 2 to the new storage subsystem 3 by migration processing later. There is no need to write, and the write processing ends without doing anything to the new storage subsystem 3.

(Second Embodiment) FIG. 5 shows a configuration example of a computer system to which the present invention is applied. This example also has the same configuration as that of the first embodiment, and includes a host 1, an old storage subsystem 2, and a new storage subsystem 3. In this example, there are a plurality of disks 23 and disks 33 in the old storage subsystem 2 and the new storage subsystem 3.

The control unit 35 is generally composed of a CPU and a memory, and has a new storage subsystem 3 in the memory.
Exists and is executed, but is omitted in the drawings of the present invention. The control device 35 differs from the first embodiment in that there are a plurality of copy pointers 143 and a plurality of access counters 145 corresponding to the copy pointers 143. FIG. 6 illustrates the copy pointer and the access counter. In this example, the copy pointer, the access counter, the start address of the area, and the end address are managed as one record. The access counter 145 indicates in which area of the disk 33 the host 1
Used to count whether there has been an I / O request from.
This is generally used to check which area is accessed most and determine the order in which to start the data migration process. For example, each of the plurality of disks 33 is divided into n regions from the beginning, and the number of I / O requests from the host for each region is recorded in the access counter 145. In this way, by observing the tendency of which area of the disc is frequently accessed, the migration process is preferentially performed from the area considered to be accessed heavily. The number of accesses may be totaled for all disks 33 or specified from the setting console 36. N copy pointers 143 are also prepared and are associated with the access counter 145. Each copy pointer takes a value from the beginning to the end of the area.

Referring to FIG. 7, the data transfer process from the old storage subsystem 2 to the new storage subsystem 3 will be described. The process up to the connection between the host 1 and the new storage subsystem 3 before the migration process is the same as in the first embodiment.

First, the copy pointer 1 in the control device 35
43 is initialized (step 2001). In the initialization, the head values of the plurality of copy pointers are set to the head addresses of the respective areas. Next, refer to the access counter of each area,
A plurality of copy pointers are rearranged in descending order of access frequency (step 2002). For example, when the area of the disk 33 is divided into three and the access frequency is as shown in FIG. 6, the area is rearranged as shown in FIG. Step 2003
Then, the transfer means selects the first copy pointer, that is, the one with the highest access frequency. In step 2004, the command queue 44 is checked to see if a request for read / write processing has been received from the host 1 (step 2005). If there is a processing request, the processing in the queue is executed (step 2006).

In step 2007, a request to read the area indicated by the copy pointer is sent to the old storage subsystem 2. In step 2008, read data is received from the old storage subsystem 2, and in step 2009, data is written to the new storage subsystem 3 at the relevant address.
In step 2010, 1 is added to the copy pointer. In step 2011, it is checked whether or not the copy pointer 43 has exceeded the final address of the migration target area. If it has not reached yet, the process returns to step 2004 and repeats the processing. If the maximum number of blocks has been reached, it is checked whether there is a copy pointer to be processed next (step 2012). If there is no copy pointer to be processed, the migration process ends. If there is, the next copy pointer, that is, the pointer of the next most frequently accessed area, is selected, and the process returns to step 2004.

Next, the processing in the new storage subsystem 3 when a read request is received from the host 1 during the migration processing will be described (FIG. 9).

First, the request address coming from the host 1 is
It is checked whether or not the area has already been copied (step 2101). In the checking method, the plurality of copy pointers 143 are searched from the beginning, and a range of the copy pointers is checked. Whether or not the copy pointer range has already been copied can be determined only by determining whether the copy pointer to which the request address belongs is before or after the copy pointer that is currently performing the migration process. If the area has already been copied, read data is read from the new storage subsystem 3 (step 210).
2) Send data to the host (step 2105). If not, a read request is sent to the old storage subsystem 2 (step 2103), and read data is received from the old disk subsystem 2 (step 210).
4).

In the check in step 2101,
If the requested address from the host 1 matches the area currently being migrated, it is considered that copying has not been completed, and step 2103 and subsequent steps are performed. Alternatively, the processing shown in FIG. 3 may be performed to check whether the read request address corresponds to a copy completed portion in the area, and select the processing.

Next, the processing when a write request comes will be described with reference to FIG. In step 2201, it is checked whether or not the requested address from the host 1 is an area where copying has already been completed. In the checking method, a plurality of copy pointers 143 are searched from the head as in the case of the read processing,
Find out which range it belongs to.

If the area has already been copied, the write data from the host 1 is written to the disk 33 of the new storage subsystem 3 (step 2202).
If the copy has not been completed, a write request is issued to the old storage subsystem 2 to write the write data received from the host 1 to the old storage subsystem 2 (step 220).
3).

At the time of writing to the new storage subsystem 3, data is copied from the old storage subsystem 2 to the new storage subsystem 3 by migration processing later. Need not be written, and the write processing ends without doing anything to the new storage subsystem 3. Alternatively, as in the first embodiment, a method of writing data to the old storage subsystem 2 side before step 2201 to make it a double-write state may be used to improve reliability.

In the method of the present embodiment or the first embodiment, if an I / O request for an uncopied area is received during the migration process, an access to the old storage subsystem 2 occurs, so that the processing speed is affected. Get out. Therefore, in the method of the present example, data migration can be completed at an early stage in a high-access area, and a reduction in processing speed is suppressed for I / O during migration processing.

(Third Embodiment) FIG. 11 shows a configuration example of a computer system to which the present invention is applied. The computer system includes a host 1, an old storage subsystem 2, and a new storage subsystem 3. The basic configuration is the same as that of the first embodiment.

The difference from the first embodiment is that the control unit 35 in the new storage subsystem 3 has a file analysis means 46. The file analysis means 46 recognizes the writing position of a specific file in the old storage subsystem 2 and enables data movement of only one file. This includes the file system used by host 1,
It is necessary that the file systems recognizable by the file analysis unit 45 match, but here it is assumed that they match. Further, there is area information 147 for managing the position of the file. This will be described later.

Referring to FIG. 12, a file management system on a disk device will be described. The file management method described here is an example, and if the file analysis unit 46 can recognize other file systems, the file management method can be applied to other file systems.

A volume label 71 is provided at a specific position on the disk.
2 (Volume Table Of Contents) is written. The VTOC 72 records, as management information of the file written on the disc, a file name, a data format of the file, information on a position where the file is stored, and the like. Here, each piece of the position information is called DSCB.

The file is placed in an area called an extent. An extent is a continuous area in which a plurality of tracks are gathered. One extent may exist for one data set, or a plurality of extents may exist. When there are a plurality of extents, each extent need not exist in a continuous area. Data on the file is recorded sequentially from the beginning of the extent described in the VTOC.

The area information 147 will be described with reference to FIG. The area information 147 holds the start address and the end address of the extent of the migration target file. When there are a plurality of extents, a plurality of information is held.

The data migration process from the old storage subsystem 2 to the new storage subsystem 3 will be described with reference to FIGS. This processing includes two parts: a part for interpreting VTOC information written in the old storage subsystem 2 (file analysis processing) and a processing for actually copying data (data migration processing). To start the processing, a request is issued from the host 1 or the setting console 36 to the new storage subsystem 3. At this time, a volume name and a file name in the old storage subsystem 2 to be migrated are specified.

The file analysis process will be described with reference to FIG. First, a read request is issued to the old storage subsystem 2 to read the volume label and place it on the cache memory (step 3001). Since the reading position of the volume label is a fixed position, it can be realized by reading an address at a certain fixed position. In step 3002, the volume label is written to the disk 33. Step 30
In 03, the position of the VTOC is extracted from the volume label, and the VTOC is read into the cache memory. Next, the VTOC read from the cache memory is read from the head, a target file is searched, and its position information is read (step 3004). In step 3005, the extent position is checked from the position information of the target file, and the start position and the end position of each extent are recorded in the area information 147. When there are a plurality of extents, a plurality of extent information is sorted and arranged in ascending order of address. In step 3006, the information of the target file of the VTOC is written to the disk 33.

FIG. 15 illustrates the transition process. Step 3
In 101, the copy pointer 43 in the control device 35 is initialized to the extent start address. Step 3102
To read the command queue and read /
It is checked whether a write processing request has been received (step 3103). If there is a processing request, the processing in the queue is executed (step 3104). In step 3105, a request to read the area indicated by the copy pointer 43 is sent to the old storage subsystem 2.

In step 3106, read data is received from the old storage subsystem 2, and in step 3107, data is written to the new storage subsystem 3 at the relevant address. In step 3108, 1 is added to the copy pointer 43. In step 3109, it is checked whether the copy pointer 43 has exceeded the last address of the extent. If not, the process returns to step 3102 to repeat the processing. If the maximum number of blocks has been reached, the area information 147
To check if the next extent exists (step 3110). If the processing has been completed for all the extents, the migration processing ends, and if there is a next extent, the copy pointer is set to the next extent start address (step 3111).

Next, the flow of processing in the new storage subsystem 3 when a read request is received from the host 1 during the migration processing will be described with reference to FIG. First, the read destination address coming from the host 1 is compared with the information on the extent of the file to be copied held in the new storage subsystem 3, and the read destination address falls within the extent of the area information 147. It is checked whether it is (Step 3201). If the address is within the extent, the copy pointer 43 is compared with the read destination address (step 3202). If the copy pointer 43 is larger than the read destination address, the data is already copied from the old storage subsystem 2 to the new storage subsystem 3, so the data is read from the new storage subsystem 3 (step 3203), read processing can be performed by sending data to the host.

Step 3201 If the file is not within the extent, the read request from the host is not an access to the file to be migrated this time, so the data is read from the old storage subsystem 2. A read request is sent to the old storage subsystem 2 (step 3204), and read data is received from the old storage subsystem 2 (step 3205). Alternatively, even if it is determined in step 3202 that the copy pointer 43 is equal to or less than the read destination address, the old storage subsystem 2
Will be read from the memory. Finally, the read processing is completed by sending the data to the host (step 3206).

Here, when there are a large number of extents, it takes overhead to check whether or not the read destination address in the step is within the extent. Therefore, in the case of the read processing, the implementation may be such that the read from the old storage subsystem 2 is executed unconditionally regardless of the value of the read destination address.

Next, the flow of processing in the new storage subsystem 3 when a write request is received from the host 1 during migration processing will be described with reference to FIG. At step 3301,
First, the write destination address coming from the host 1 is compared with the information on the extent of the copy target file held in the new storage subsystem 3, and the write destination address falls within the extent of the area information 147. Check if it is. If the data is in the extent, the write request is issued to the old storage subsystem 2 and the write data received from the host 1 is written because the data is to be written to the file to be migrated (step 3302). Next, the copy pointer 43
Is compared with the write destination address (step 330).
3). If the copy pointer 43 is smaller than the write destination address, data is copied from the old storage subsystem 2 to the new storage subsystem 3 by migration processing later. No, the write processing ends without doing anything to the new storage subsystem 3. Otherwise, or if it is determined in step 3301 that the write destination address is outside the target extent, the new storage subsystem 3
Data is written to the disk 33 in the disk (step 330
4).

[0049]

According to the computer system of the present invention, data can be migrated from the old storage subsystem to the new storage subsystem while receiving processing from the host computer. In addition, migration can be performed in file units, and only the data to be migrated can be selectively migrated. Also, since the data being migrated is duplicated in the old storage subsystem and the new storage subsystem, the migration process continues even if a failure occurs in either storage subsystem during migration. it can.

[Brief description of the drawings]

FIG. 1 shows a configuration example of a computer system to which the present invention is applied.

FIG. 2 shows a flow of a data migration process in the computer system of the present invention.

FIG. 3 shows a flow of a read process during a data migration process of the new storage subsystem 3;

FIG. 4 shows a flow of a write process during a data migration process of the new storage subsystem 3;

FIG. 5 shows a configuration example of a computer system according to a second embodiment to which the present invention is applied.

FIG. 6 shows a management method of a copy pointer and an access counter according to the present invention.

FIG. 7 shows a flow of a data migration process in the computer system of the present invention.

FIG. 8 shows a state in which copy pointers are rearranged in the data migration processing of the present invention.

FIG. 9 shows a flow of a read process during a data migration process of the new storage subsystem 3;

FIG. 10 shows a flow of a write process during a data migration process of the new storage subsystem 3;

FIG. 11 shows a configuration example of a computer system according to a third embodiment to which the present invention is applied.

FIG. 12 shows a file management method according to the third embodiment.

FIG. 13 shows the contents of area information in the third embodiment.

FIG. 14 shows a flow of a file analysis process according to the third embodiment of the present invention.

FIG. 15 shows a flow of a data migration process according to the third embodiment of the present invention.

FIG. 16 shows a flow of a read process during a data transfer process according to the third embodiment of the present invention.

FIG. 17 shows a flow of a write process during a data migration process according to the third embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Host, 2 ... Old storage subsystem, 3 ... New storage subsystem, 21 ... Interface, 22 ... Interface, 23 ... Disk, 31 ... Interface, 32 ... Interface, 33 ... Disk, 34
... cache memory, 35 ... control means, 36 ...
Setting console, 41: I / O processing means, 42: migration means, 43: copy pointer, 44: command queue, 51: transmission path, 52: transmission path,
143: copy pointer, 145: access counter,
147 area information.

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Akira Kurano 2880 Kozu, Kozuhara-shi, Kanagawa Pref., Ltd.Storage Systems Division, Hitachi, Ltd. 5B065 BA01 CE11 CE21 EA31 5B082 AA00 CA10 DE04 DE06 HA05 in Storage System Division

Claims (3)

[Claims] 1. A computer system comprising: a first storage subsystem; a second storage subsystem; and a computer connected to the first and second storage subsystems, wherein the first storage subsystem comprises a first storage subsystem, a second storage subsystem, and a computer connected to the first and second storage subsystems. The storage subsystem of the second
And the second storage subsystem is connected to the first storage subsystem and the computer, and (A) a specific storage of the first storage subsystem is connected to the first storage subsystem. Means for transferring the data of the area to the corresponding storage area of the second storage subsystem in order from the top (B) location information storage means for recording to which area the transfer has been performed by the transfer means (A) (C) upon receiving a read request from the computer, determine from the location information storage means whether data has been transferred to the second storage subsystem, and if data has been transferred, The data in the second storage subsystem is transferred, and if not, the data is transferred from the corresponding storage area of the first storage subsystem to the second storage subsystem. Means for reading the data and transferring data to the computer; (D) receiving a write request from the computer, the corresponding storage area of the second storage subsystem and the first storage subsystem. Means for writing data to the first storage device subsystem and executing data migration from the first storage device subsystem to the second storage device subsystem without stopping processing of the computer. Equipment subsystem. 2. The storage subsystem according to claim 1, wherein an area having a high access frequency in said specific storage area is transferred preferentially. 3. The storage device according to claim 1, wherein said specific storage area is a file managed by a computer connected to said first or second storage device subsystem. sub-system.