JPH08335146A - Disk system changing system - Google Patents

Abstract

PURPOSE: To change redundant arrays of inexpensive disks (RAID) configuration without stopping a system by changing inputted data into data suited to a second RAID disk system and redundant data. CONSTITUTION: This system is provided with an input means 100 for successively reading data based on a first RAID system 1, a changing means 106 for changing the data inputted by the input means 100 into the data suited to a second RAID disk system 5 and the redundant data, and an output means 102 for outputting the data and redundant data changed by the changing means 106. Then, the changing means 106 generates a redundant group so that the data of the first RAID disk system 1 can be suited to the second RAID disk system 5. Therefore, a disk system can be changed between different RAID levels. Besides, even when the configuration is different at the same RAID level, the disk system can be changed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk system in which a plurality of disk devices are connected, and more particularly to a disk system capable of increasing or decreasing the number of disk devices without stopping the system. Further, the present invention relates to a disk system that can easily realize a RAID level change and a RAID block size change without stopping the system when the disk system is a RAID disk system.

[0002]

[Prior art]

Conventional example 1. As a conventional invention for changing the RAID level, there is a "disk array device" in Japanese Patent Laid-Open No. 6-119121.
In the above invention, the optimum access format is selected according to the amount of data from the host device and the data is written in the disk array device. Then, after writing the data, by utilizing the free time of the disk array device, the optimum access format is selected to read the data written to the disk array device, and the read data is written to the disk device where the original data should be written. Write. FIG. 38 shows a principle explanatory diagram in the above invention. According to FIG. 38,
The above invention has a disk array 246 including a plurality of disk devices 232a to 232e for storing data and storing parity. When the write data amount instructed by the host device 218 is small, the first writing means 234 individually accesses the plurality of disk devices 232a to 232e to write the data divided into a predetermined length, for example, a sector unit. . When the write data amount instructed by the host device 218 is large, the second writing means 236 accesses the plurality of disk devices in parallel to write the data divided into sector units. Here, the first writing unit 234 performs writing in the individual access mode of FIG. 38 (b). In this individual writing mode, first, prior to the writing of data, the data of the planned writing position of the disk device to be written and the redundant information at the same position as the planned writing position stored in another disk device are written. Is read out and new redundant information is generated by exclusive OR. Next, the new sector data and parity,
Each of the disk devices for data storage and parity storage is individually accessed and written asynchronously. The second writing unit 236 writes in the parallel access mode of FIG. 38 (c). That is, a parity as new redundant information is generated from the exclusive OR of a plurality of sector data stored in parallel in the disk array direction. Next, the plurality of sector data and the parity are asynchronously written by accessing the plurality of disk devices in parallel. here,
The individual access mode corresponds to RAID5 (may be RAID4), and the parallel access mode is RAID.
Since it corresponds to No. 3, the above invention performs efficient data writing by making use of the advantages of RAID5 (or RAID4) and RAID3 according to the amount of data. Furthermore, R
After writing a large amount of data in the parallel access mode corresponding to AID3, by utilizing the free time of the disk array device, etc., the stored data is read out according to the parallel access mode and rewritten in the individual access mode to perform transaction. To be able to deal with individual reading and writing of the disk device during processing.

Conventional example 2. Further, as a conventional method for changing the RAID configuration, there is a method in which the system is once taken offline, backup is performed, and after the RAID configuration is changed, the disk device is initialized and the backed up data is reloaded.

[0004]

However, the conventional method of changing the RAID configuration in the disk system has a problem in that it cannot be added in a 24-hour operation system. Also,
There was a problem that a large capacity backup device was required.

The present invention has been made in order to solve the above-mentioned problems, and is an RAI without stopping the system.
Allows changing the D configuration.

[0006]

The present invention provides a first RA.
A disk system changing method for changing an ID (Redundant Arrays of Inexpensive Disks) disk system to a second RAID disk system is characterized by having the following elements. (A) input means for sequentially reading data based on the first RAID disk system, (b) changing means for changing the data input by the input means to data compatible with the second RAID disk system and redundant data, (c) ) Outputting means for outputting the data changed by the changing means and the redundant data.

Further, according to the present invention, a redundancy group is formed by N (N is an integer of 2 or more) disk devices.
In the disk system changing method of changing the RAID disk system of No. 1 to the second RAID disk system in which the redundant group is configured by N + M (M is an integer of 1 or more) disk devices, the following elements are provided. . (A) Input means for sequentially reading N + M-1 data from N disk devices; (b) Redundant data calculation means for obtaining redundant data by using N + M-1 data input by the input means as a new redundant group. , (C) N + M−1 pieces of data included in the new redundancy group and the redundancy data obtained by the redundancy data calculation means are N +
Output means for writing to M disk devices.

Further, according to the present invention, N × (N is an integer of 2 or more) disk devices forming a redundant group.
First with L disk devices (L is an integer of 1 or more)
From the RAID disk system of No. 2, a second R having N × (L + M) (M is an integer of 1 or more) disk devices.
The disk system changing method for changing to the AID disk system is characterized by having the following elements. (A) hot-line connecting means for hot-connecting N × M disk devices to the N × L disk devices of the first RAID disk system; and (b) connection by the hot-line connecting means. Initialization means for initializing the storage areas of the N × M disk devices to a predetermined value; (c) number of areas for increasing the maximum number of storage areas by the storage areas added by the N × M disk devices Change means.

According to the present invention, one of a RAID disk system based on RAID level 4 and a RAID disk system based on RAID level 5 is a first RAID.
In the disk system changing method for changing the RAID disk system from the first RAID disk system to the second RAID disk system, the D disk system is used and the other is the second RAID disk system,
It is characterized by having the following elements. (A) Input means for sequentially reading data belonging to the same redundancy group at one RAID level, (b) data read by the above-mentioned input means and redundant data for the other RAID level
Position determining means for determining the arrangement of data and redundant data so as to conform to the ID level, (c) Data of the redundant group in which the data and redundant data are read by the input means according to the arrangement determined by the position determining means. Output means for sequentially rewriting to the area where was recorded.

Further, according to the present invention, one of a RAID disk system based on RAID level 3 and a RAID disk system based on RAID level 4 or higher is used as a first RAID disk system, and a RAID disk system based on RAID level 3 is used. The first RAID
When the disk system is used, one of the RAID disk systems based on RAID level 4 or higher is set as the second RAID disk system, and the RAID level 4 is used.
If one of the RAID disk systems based on the above is the first RAID disk system, RA
Second RAID disk system based on ID level 3
R RAID disk system from the first RAID disk system to the second RAID disk system
The disk system changing method for changing the AID disk system is characterized by having the following elements. (A) Input means for sequentially reading data of one RAID level in units of the block size of blocks used at RAID level 4 or higher; (b) Data so that the data read by the input means conforms to the other RAID level. A format changing means for changing the format and changing to data and redundant data based on the new format, (c) according to the size changed by the size changing means, the redundant group of the data and the redundant data read by the input means. Outputting means for rewriting in order to the area where the data was recorded.

Further, according to the present invention, from the first RAID disk system based on RAID level 1 to the second RAID based on any of RAID levels 0, 3, 4, and 5.
The disk system changing method for changing the RAID level to the disk system is characterized by having the following elements. (A) Input the data of the first RAID disk system to be used for one group of the second RAID disk system and record it in the area overwritten to record the data of the group of the second RAID disk system An input means for inputting the data of the first RAID disk system and the other data included in the group of the second RAID disk system to which the data belongs, and (b) the data read by the input means as the second data.
Level changing means for changing to data of the RAID disk system, (c) output means for sequentially writing the data changed by the level changing means in the area where the data read by the input means was recorded, (d) the above A change management table for managing the position of data whose RAID disk system has been changed by the input means, the level changing means, and the output means, (e) referring to the change management table,
A rearrangement means for changing the arrangement of data.

The present invention also provides RAID level 3,
A disk system changing method for changing a RAID level from a first RAID disk system based on any one of 4 and 5 to a second RAID disk system based on RAID level 0 is characterized by having the following elements. (A) Any one of RAID levels 3, 4, and 5 RA
Input means for sequentially inputting the data of the ID level, (b) level changing means for changing the data inputted by the input means to data of RAID level 0, and (c) an output means for writing the data changed by the level changing means.

Further, according to the present invention, a RAID having a first block size is defined as a block which is an access unit for each disk device accessed by the RAID disk system.
From the first RAID disk system at any of levels 0, 3, 4, 5 to the RA having the second block size
Second RA of ID level 0, 1, 3, 4, 5
The disk system changing method for changing to the ID disk system is characterized by having the following elements. (A) The first block size, the number of disk devices connected to the first RAID disk system, and the first
RAID based on the RAID disk system level of
While calculating the group size of the ID disk system, the second block size, the number of disk devices connected to the second RAID disk system, and the second
Second RAID based on the RAID disk system level of
A group size of the ID disk system is calculated, and based on the group size of the first RAID disk system and the group size of the second RAID disk system, a group boundary according to the group size of the first RAID disk system, An input unit that calculates the least common multiple area size in which the group boundaries according to the group size of the second RAID disk system match, and sequentially reads the data of the first RAID disk system in units of the calculated least common multiple area size. b) size change means for changing the data read by the input means to data based on the second block size, and (c) data read by the input means for the data changed by the size change means. An output means for sequentially rewriting to the area.

Further, according to the present invention, the above-mentioned disk system changing system further comprises a locking means for prohibiting access to the data read by the input means until the data is output by the output means. Is characterized by.

Further, the present invention is characterized in that the disk system changing method further has a completion pointer for storing a position where writing by the output means is completed.

Further, in the present invention, the above-mentioned disk system changing method further refers to the completion pointer, executes access to data after the completion pointer based on the first RAID disk system, and executes the access before the completion pointer. It is characterized in that it comprises an access means for accessing data based on the second RAID disk system.

Further, according to the present invention, the above-mentioned disk system changing method further employs a detecting means for detecting an area of data recorded by the first RAID disk system, and a second RAID disk system. When the area in which is recorded is smaller than the area detected by the detection means, a remaining area initialization means for initializing the remaining area with a predetermined value is provided.

Further, according to the present invention, in the case where the disk system changing method further stores the maximum value of the data recording area, and the area for recording data is increased by taking the second RAID disk system. The maximum value is updated only for the increased area.

[0019]

In the disk system changing method of the present invention, the changing means creates the redundant group so that the data of the first RAID disk system is adapted to the second RAID disk system. Therefore, different RA
The disk system can be changed between the ID levels. Alternatively, at the same RAID level, the disk system can be changed even when the configuration is different.

In the disk system changing method of the present invention, when the redundant data calculating means adds a disk, the redundant group is created so as to include the added disk device. According to the present invention, the redundant data calculation means creates a new redundant group when a disk is added in the horizontal direction.

In the disk system changing method of the present invention, when a disk is added in the vertical direction, the hot disk connecting means hot-connects the added disk. The initialization means initializes the added disk device to make it usable. The area number changing means notifies the system of the area in which the usage state becomes possible.

In the disk system changing method of the present invention, the change is made between RAID level 4 and RAID level 5. RAID level 4 and RAID level 5 are
Since the position of the redundant data is different, the position determining means rearranges the position of the redundant data before the change to the position of the redundant data after the change, whereby RAID level 4 and R
Change AID level 5.

In the disk system changing method of the present invention, the RAID level 3 and the RAID level 4 or higher are changed. RAID level 3 and RAID level 4 and above are recorded in units of data of different formats. R
Since the data format of the data format used in AID level 4 or higher is larger, the data is input in units of the data size of the data format used in RAID level 4 or higher.
When changing from RAID level 3 to RAID level 4 or higher, the format changing means changes the format of the input data to a new RAID level 4 or higher data format and creates new data. When changing from RAID level 4 or higher to RAID level 3, the data of RAID level 4 or higher is divided to create data in the format of RAID level 3.

In the disk system changing method of the present invention, RAID level 1 to RAID levels 0, 3,
Make changes 4 and 5. RAID level 1 stores data in duplicate. On the other hand, RA
The ID levels 0, 3, 4, and 5 are for generating groups and recording data. Therefore, when changing the data of the RAID level 1 to the RAID levels 0, 3, 4, and 5, the data of the group in which the data based on the RAID level 1 is generated based on any of the RAID levels 0, 3, 4, and 5. Will be overwritten by. In order to prevent this overwriting, R recorded in the area to be overwritten
The AID level 1 data is saved in advance by the input means. The input means also inputs in advance data to be saved for overwriting and other data of the group to which the data should belong. The level changing means creates data based on the changed level from the input data. The output means writes the created data to the disk device,
Since the data to be overwritten by the writing has already been saved, there is no problem even if the data is overwritten.
Groups are created from the saved data and output in order. In this way, if the RAID level is changed in order, the data will be overwritten. Therefore, the data is saved in advance and the saved data is written in no particular order. A change management table exists to manage such out-of-order output. The rearrangement unit refers to the change management table and rearranges the data output in random order.

In the disk system changing method of the present invention, the RAID level of any of RAID levels 3, 4, and 5 is set.
Change from D level to RAID level 0. RA
ID level 0 has a structure without redundant data. Therefore, the RAID level data of any one of the RAID levels 3, 4, and 5 is input in order, the redundant data is removed and the RAID level is changed and output.

The disk system changing method of the present invention is
The RAID level is changed when the block size is different. When the block sizes are different, the input means determines each group size from the number of disk devices connected to the disk system, the block size before the change and the block size after the change. Then, the data is read with the size at which the boundary between the group size before the change and the group size after the change matches as one unit. For example, when the group size before the change is 6 Kbytes and the group size after the change is 9 Kbytes, 18 Kbytes are read as one unit. The size changing means divides the read 18 Kbyte data into two and creates two 9 Kbyte groups. However, redundant data is not included in 9 Kbytes. In this way, the block size can be changed.

The lock means in the present invention prohibits access to the data read by the input means until the data is output by the output means.

Further, a completion pointer is provided, and the completion pointer stores the position where writing is completed by the output means.

When accessing the data after the completion pointer, the access means uses the access method before the change. When accessing the data before the completion pointer, the modified method is used.
Therefore, the disk system can be changed in real time.

The remaining area initializing means initializes the remaining area when the area where the changed data is stored is smaller than the area where the pre-change data is stored. By initializing the area, it is possible to easily create redundant data when recording the subsequent data.

The disk system changing method of the present invention is
When the data recording area is increased by adding the disk device, the maximum value of the accessible area is updated. In this way, disks can be added in real time.

[0032]

【Example】

Example 1. In this embodiment, N units (N is an integer of 2 or more)
First R forming a redundant group of disk devices
A disk system changing method for changing from the AID disk system to the second RAID disk system in which a redundant group is constituted by N + M (M is an integer of 1 or more) disk devices will be described below. FIG. 1 is a system configuration diagram showing an example in which disks are added in the horizontal direction. In FIG. 1, the first RAID disk system 1
Is composed of a RAID controller 2 and five disk devices 3 connected to the RAID controller 2 by seven ports 4 of ports a to g. Then, while moving the system, the disk device 3a is added to each of the vacant ports f and g of the RAID controller 2. Then, the first RAID disk system 1 is set to the READY state. After that, a new data structure is completed step by step from the top for the second RAID disk system 5 including the added disk device 3a. At this time, the access to the already completed stage is performed by the RAID configuration in the second RAID disk system 5. The stage for which the configuration is to be changed is locked and the access to the stage is made to stand by. Then, the data necessary for configuring the second RAID disk system is
The data is read based on the data structure in the first RAID disk system 1. Then read the second RAI
The data is written in a predetermined position in the data structure of the D disk system 5. After writing the data, the lock is released for the previously locked data. Access to data that has not undergone a configuration change or data that has not been locked is performed based on the RAID configuration in the first RAID disk system 1. When the configuration change up to the final data based on the data configuration in the first RAID disk system 1 is completed, "0" or "" is set so that parity matching can be performed in the area of the disk device after the final data is stored. Write "1".

Next, using a specific RAID configuration, the RAID configuration change after the disks are added in the horizontal direction will be described. FIG. 2 is a diagram showing an example of a disk extension procedure in RAID level 5. In FIG. 2, the first RAID disk system 1 is configured by RAID level 5. The disk device 3 is connected to each of the ports a to e of the RAID controller 2. In this embodiment, port a to port e
The device numbers of the disk devices 3 connected to are sequentially numbered from D0 to D4. Then, the cables and the disk devices 3a having the device numbers D5 and D6 are connected to the vacant ports f and g, respectively. After the disks are connected, the I / O bus is reset from the ports f and g to put the added disks D5 and D6 into the READY state. At this time, the I / O bus connected to port f and port g is SC
If it is an SI bus, the RAID controller 2
After sending the SI bus reset to each disk unit, ST
ART UNIT command, TEST UNIT RE
Send the ADY command to each connected disc,
Put each disk in READY state. Next, the data structure is changed. To change the data configuration, the data collection of one row in the second RAID disk system 5 is changed to one.
Each redundancy group is constructed one by one in order from the top. FIG. 2 shows that the data configuration up to the upper two stages has been completed. For this reason, a value indicating the completion up to the second stage is stored in the completion pointer.

A case of creating a third-stage redundancy group in the second RAID disk system 5 will be described with reference to FIG. First of all, the input means 100 uses the second R
Data to be stored in the third row of the AID disk system 5 "12, 13, 14, 15, 16, 17"
Lock. Then, the input unit 100 replaces the data of the data “12, ..., 17” with the first RAID.
The data is read based on the data structure based on the disk system 1, and the read data is transmitted to the output unit 102 and the redundant data calculation unit 101. Redundant data calculation means 101
Is the data “12, ...
., 17 ”data is calculated based on exclusive OR to calculate redundant data. The calculated redundant data is output to the output unit 1.
Send to 02. The output means 102 writes each data at a predetermined position based on the data “12, ..., 17” received from the input means 100 and the redundant data received from the redundant data calculation means 101. The data scheduled to be written in the third stage of the second RAID disk system 5 is “12, ..., 14, P (redundant data), 15,
..., 17 ". Then, the output means 102 advances the completion pointer to the next stage by one stage, and after setting it to three stages, the data “12, ...
.., unlock the 17 ".

From the host system to the second RAID
When an access request is made to the disk system 5 and if the area is before the completion pointer, the data is accessed based on the data configuration in the second RAID disk system 5. If the access request is for data stored in the area after the completion pointer, the first RAID disk system 1
The data access is executed based on the data structure in. Further, the access request for the locked area is the same as the processing for the access request for the area after the completion pointer. The data block size in the host system and the data block size in the RAID are generally different. Therefore, when an access request is made by designating a block address and the number of blocks from the host system, an address conversion means (realized by software) in the RAID (not shown) performs the conversion to calculate the block address and the number of blocks in the RAID. And then R
Access is performed inside the AID. For example, assuming that the block size of the host system = 2 KB and the block size inside the RAID = 4 KB, and there is a READ request from the host system with a block address of 10 and a block number of 7, the block address within the RAID is 5, It is converted into READ with 4 blocks and accessed, and the part required after READ, that is, 14K from the beginning
Transfer B to the host system. Since this conversion is simply calculated, it is omitted below, and the block means a block in the RAID after conversion.

By the procedure as described above, the second RAI
When the value of the completion pointer reaches the final address in the first RAID disk system 1, "0" is stored in the data non-storing area of the additional disk 3a as the data structure changing process in the D disk system 5 is performed.
By storing “0”, parity matching can be obtained. Further, the output unit 102 rewrites the final address in the first RAID disk system 1 stored in the RAID controller 2 to the final address in the second RAID disk system 5. The rewritten value is the RAID that the RAID controller 2 has.
The host system is notified through the management tool, and this notification is used as a completion report to complete the disk expansion process.

Next, a disk extension procedure in RAID levels 3 and 4 will be described. Figure 3 shows the RAID
It is a figure which shows an example of the disk addition procedure in level 3 or 4. The disk expansion in FIG. 3 is performed by the RAI in FIG.
The disk expansion in the D level 5, the port number for connecting the expanded disk, and the data configuration in the first RAID disk system 1 are the same. However, RA
Since the ID levels are different, redundant data is stored in the device number D4 in FIG. In the second RAID disk system 5, the value of the completion pointer indicates the completion up to the second stage. Then, the data structure of the third stage in the second RAID disk system 5 is changed. At this time, the type of data read by the input means 100 from the first RAID disk system 1 and the type of data to be locked are the same as in FIG. Input means 10
0 transmits the read data to the redundant data calculation means 101 and the output means 102. Redundant data calculation means 101
Calculates redundant data by the same method as in FIG.
It is transmitted to the output means 102. The output means 102 includes the data received from the input means 100 and the redundant data calculation means 10.
Change the data structure based on the redundant data received from 1.
Data is written in the third stage in the second RAID disk system 5. The type of data at this time is the same as that in FIG. 2, but in FIG. 3, since the RAID level is 3 or 4, the arrangement of the data to be written is “12, ..., 17,
P (redundant data) ". The processing after this is the same as in FIG.
The same process as the process described above is performed.

In the description of the disk expansion procedure in RAID levels 3, 4, and 5, it is explained that the output means 102 receives data from the input means 100 and the redundant data calculation means 101. However, the output unit 102 may receive the data and the redundant data only from the redundant data calculation unit 101.

Next, a disk extension procedure in RAID level 0 will be described. FIG. 4 is a diagram showing an example of a disk extension procedure in RAID level 0. R
The difference between the AID level 0 and the RAID levels 3, 4, and 5 described above is that redundant data does not exist. Therefore, one of the first RAID disk system 1 and the second RAID disk system 5 in FIG.
The data stored in one redundancy group is shown in Fig. 2 above.
The data arrangement is slightly different from that of RAID levels 3, 4, and 5 in FIG. For example, in the second RAID disk system 5 of FIG. 4, the data stored in the first level is
“0, ..., 6”, which is the second in FIGS. 2 and 3.
The data stored in the first stage of the RAID disk system 5 is “0, ..., 5”. For this reason,
When changing the data structure of the third level in RAID level 0, the input means 100 uses the data "1" to be stored in the third level of the second RAID disk system 5.
4, ..., 20 ”is read based on the data configuration in the first RAID disk system 1. And
Since the input means 100 does not need to calculate redundant data, it sends the read data only to the output means 102. The output unit 102 writes the data received from the input unit 100 in the third stage of the second RAID disk system 5. The input unit 100 locks and reads data to be read in the same manner as the RAID levels 3, 4, and 5, and the output unit 102 unlocks the data after writing, and further, Change the value of the completion pointer. And the data structure change is the first RA
When the final address in the ID disk system 1 is reached, there is no redundant data in RAID level 0, so there is no need to write "0" in the data non-storage area of the additional disk. Then, the final address stored in the RAID controller 2 is changed to a new final address. Further, the notification of the value of the final address to the host system is performed in the same manner as in the above RAID levels 3, 4, and 5.

In the above description of FIG. 4, the input means 10 is used.
0 reads data from the first RAID disk system 1 and sends the read data to the output means 102,
The data received by the output means 102 was written in the second RAID disk system 5. However, since the redundant data is not included in the read data and the write data, the input unit 100 may write the read data as it is in the second RAID disk system 5.

Next, a disk extension procedure in RAID level 1 will be described. FIG. 5 is a diagram showing an example of a disk extension procedure in RAID level 1. In FIG. 5, since the first RAID disk system 1 and the second RAID disk system 5 are RAID level 1, the data has a mirror configuration and the copy data of the disk device D0 connected to the port a is the port b. Is stored in the disk device D1 connected to. The copy data of the data stored in the disk device D2 connected to the port c is stored in the disk device D3 connected to the port d. Then, the disks to be added are connected to the ports e and f. RAI
Since the data structure of D level 1 is the mirror structure as described above, the data structure is not changed even when a new disk is added in the horizontal direction. Also, R
Since the AID level is 1, the output unit 102 does not need to store the initial value in the data storage area of the added disk. In addition, the data lock and the change of the value of the completion pointer explained in the disk extension procedure in the RAID levels 1, 3, 4, and 5 change the final address stored in the RAID controller 2, and the host system To notify the end of the expansion process.

As described above, in the RAID configuration in which the redundant data exists, the redundant group is created so as to include the disk device in which the redundant data calculating means is added. In this way, even if disk devices are added in the horizontal direction, the user can easily construct a data structure in a new disk device environment without stopping the system as in the conventional case. Further, since it is not necessary to back up data, the cost of preparing a backup disk device can be reduced.

Example 2. In this embodiment, an example of adding disks in the depth direction (vertical direction) will be described below. FIG. 6 is a system configuration diagram showing an example in which disks are added in the vertical direction. In FIG. 6, the first RAID disk system 1 includes a RAID controller 2 and a RAID
The ID controller 2 includes five ports 4 and a disk 3 connected to a cable extending from the port 4. Port 4 is port a to port e
2 disks 3 are vertically connected to each port 4. Further, the connection of each disk 3 to the cable is assumed to be live-line connected by the live-line connecting means 103. Further, the RAID controller 2 is
The maximum storage area of the disk 3 connected to the ID controller 2 is stored in the memory. First RA in FIG. 6
In the ID disk system 1, the maximum storage capacity is "10
0xnx2 "(" 100 "is the maximum storage capacity of each disk 3, and" n "is the number of horizontal valid data disks connected to the RAID controller 2. For example, horizontal disk When the number of devices is N, RAID
At level 0, n = N, at level 1, n = N / 2, and at levels 3, 4, and 5, n = N-1. "2" indicates that the disk devices 3 are connected in two columns in the vertical direction. ) Is stored in the memory of the RAID controller 2.

The first RAI configured as described above
A case where the second RAID disk system 5 is constructed by adding five disk devices 3a in the third stage to the D disk system 1 will be described. First, the disk device 3a is hot-connected to the table output from each port 4 by the hot-line connecting means 103. And the connection is S
If connected via CSI bus, START UNI
The T command and the TEST UNIT READY command are transmitted to each disk device 3a on the SCSI bus to bring the disk device 3a into the READY state. Next, the value "0" or "1" is written in all the areas of the disk device 3a added by the initialization means 104 so that parity consistency can be obtained. However, the disk device 3 to be added in advance
If a is initialized to the above value, the writing of the above value by the initialization means becomes unnecessary. Finally, the area changing unit 105 changes the maximum number of storage areas stored in the RAID controller 2 of the second RAID disk system 5. The maximum storage capacity is “100 × n × 3” (“100”)
Is the maximum storage capacity of each disk 3, and “n” is RAI.
This is the number of horizontal valid data disks connected to the D controller 2. For example, when the number of disk devices in the horizontal direction is N, n = N for RAID level 0, n = N / 2 for level 1, n = N-1 for levels 3, 4 and 5.
Becomes "3" indicates that the disk devices 3 are connected in three columns in the vertical direction. ) Is written in the memory of the RAID controller 2. Then, the new maximum storage capacity number is notified to the host system through the RAID management tool provided in the RAID controller 2, and this notification completes the disk device addition processing.

As described above, in the vertical disk extension, the horizontal disk extension in the first embodiment and the data stored in the disk device 3 constituting the first RAID disk system are re-created. The difference is that there is no need to organize. Then, the hot disk connection means hot-connects the added disk, and the initialization means initializes the added disk device to make it usable. Further, the area changing means informs the system of the area in which the usage state becomes possible. Therefore, the user who uses the disk system can effectively and easily use the disk system even when the disk device is added in the vertical direction.

Example 3. In this embodiment, an example of adding disks in the horizontal and vertical directions will be described below. Figure 7
[Fig. 3] is a system configuration diagram showing an example in which disks are added in the horizontal and vertical directions. According to FIG. 7, the first RAID disk system 1 includes a RAID controller 2, seven ports 4, five disk devices 3 respectively connected to the ports a to e, and further, the disk device 3 described above. One disk device 3 is connected to each of them in the vertical direction. For example, with respect to the first RAID disk system 1 configured as described above, two disk devices 3a are arranged horizontally and two units are vertically arranged in the vacant ports f and g.
It is also possible to build a second RAID disk system 5 by adding a total of four units. In this way, when the disk device 3a is added, the data reconstruction can be performed by the same method as in the first embodiment. Further, the cable of the disk device 3a in the vertical direction can be connected by the hot line connecting means 103, and the maximum storage area number can be updated by the area changing means 105.

Example 4. In this embodiment, data reorganization associated with a change in RAID level will be described below. In particular, an example will be described in which the data read by the input unit and the redundant data are reorganized by using the position determining unit that determines the arrangement of the data and the redundant data so that the data and the redundant data match the other RAID level.

First, a method for realizing the disk system changing method according to the present invention will be described. FIG.
FIG. 3 is a diagram showing an example of realizing the disk system changing method in the present invention. According to FIG. 8, the input means 10
0 sequentially reads data from the disk devices that make up the first RAID disk system 1. At this time, the input means 100 uses the R of the first RAID disk system.
Enter data based on AID level. Next, the changing unit 106 changes the data input by the input unit 100 into data that conforms to the RAID level of the second RAID disk system and redundant data. Then, the output unit 102 outputs the data changed by the changing unit 106 and the redundant data to the disk device constituting the first RAID disk system 2, and this is used as the second RAID disk system. The first RAID disk system 1 and the second RAID disk system 5 are not separate systems, but the first RAID disk system 1
The second RAID disk system 5 is obtained by reorganizing the data and rewriting the data.

Next, an example in which the first RAID disk system is composed of RAID level 4 and is changed to the second RAID disk system which is composed of RAID level 5 will be described. FIG. 9 shows RAID level 4 to R.
It is a figure which shows the procedure of changing to AID level 5. RAI
The first RAID disk system configured by D level 4 and the second RAID disk system configured by RAID level 5
When changing to another RAID disk system, RAID
The difference between level 4 and RAID level 5 is that the position of parity (indicated by “P” in FIG. 9) that is redundant data in one group is different. The one group is a group of data that can be input / output by one access from the RAID controller 2 stored in the disk device 3 connected in the horizontal direction. For example, in the first RAID disk system 1 of FIG. 9, the data “0, 1, 2, P” at the top is set as group 0. In FIG. 9, the data reorganization of the second RAID disk system 5 is completed up to the group 1, and the completion pointer indicates the completion up to the second stage. And the second RAI
A procedure for reorganizing the third stage data of the D disk system 5 will be described. First, the input means 100 uses the first RAI.
Input "6, 7, 8, P" from the third stage (group 2) of the D disk system 1. And the input means 10
0 transmits the data input to the position determination means 107. The position determination means 107 rearranges the data received from the input means 100 so as to conform to RAID level 5. The writing position of the parity of the third stage is to write to the disk number D2 of the third unit. Therefore, the position determining means 107 rearranges the data into “6, 7, P, 8” and transmits it to the output means 102. Output means 102
Sends the data received from the position determining means 107 to the second R
Rewriting is performed on the third stage of the AID disk system 5. Then, the output means 102 changes the completion pointer to a value indicating completion up to the third stage. The input means 100
Locks the input data when the data is input from the first RAID disk system 1.
Also, the host system waits for access to the above data. Then, the output means 102 outputs the second RAID
After rewriting the data to the disk system 5,
unlock.

The second R from the host computer
As for the access request to the AID disk system 5, the access to the area (the area indicated by the value smaller than the value set in the completion pointer) for which the data reorganization has already been completed is the RAI in the second RAID disk system 5.
Processing is performed based on the D level. Further, among the uncompleted areas (areas for values larger than the value set in the completion pointer), the areas locked by the input means 100 are made to wait for access. Then, when the data reorganization is completed and the lock is released, the data access processing is immediately performed based on the RAID level in the second RAID disk system 5. Further, among the uncompleted areas, the areas that are not locked are the RAID in the first RAID disk system 1.
Access based on the D level. After performing the data reorganization on all the data stored in the disk device 3 constituting the first RAID disk system 1, the output unit 102 outputs the final data stored in the memory of the RAID controller 2. Address to the second RA
Rewrite to the new final address in the ID disk system 5. This completes the second RAID disk system 5
Complete the construction of.

Next, an example in which a second RAID disk system composed of RAID level 4 is newly constructed from a first RAID disk system composed of RAID level 5 will be described. FIG. 10 is a diagram showing a procedure for changing from RAID level 4 to RAID level 5. FIG. 10 shows that the data reorganization up to the second stage of the second RAID disk system 5 has been completed. Therefore, 3 of the first RAID disk system 1
An example of reorganizing the stage data will be described. As described above with reference to FIG. 9, RAID level 4
In RAID level 5, the type of data included in one group is the same, but the position of parity is different. Therefore, the position of the data transmitted from the input unit 100 is reorganized by the position determining unit 107. FIG.
At 0, the input means 100 inputs the data “6, 7, P, 8” in the third row of the first RAID disk system 1,
The input data is transmitted to the position determining means 107. The position determining means 107 sets the data sequence to "6, 7, 8, P".
Then, the rearranged data is transmitted to the output means 102. The output means 102 outputs the received data to the second RAI.
Rewriting is performed on the third stage of the D disk system 5. Further, the lock and unlock of the input data, the change of the value of the completion pointer, and the change of the final address are performed in the same manner as the procedure of reorganizing the data from RAID level 4 to RAID level 5 in FIG. 9 above. To do. Input means 100, position determination means 107, output means 102
Is realized by software included in the RAID controller 2.

RAID level 4 and RAID level 5
As described above, when performing data reorganization between RAID levels having different positions of redundant data, the position determining unit rearranges the position of the redundant data before the change to the position of the redundant data after the change. Therefore, the user can easily make a system change between RAID level 4 and RAID level 5.

Example 5. In this embodiment, an example will be described in which a first RAID disk system configured by RAID level 3 is newly constructed to a second RAID disk system configured by RAID level 4 or higher. RAID level 3 and RAID level 4
The difference is that the data to be stored is byte interleaved at RAID level 3 and is block interleaved at RAID level 4 or higher.
Therefore, in this embodiment, the format changing means 108 changes the data storage format from byte interleave to block interleave and vice versa.

First, an example of constructing a second RAID disk system 5 configured by RAID level 5 from a first RAID disk system configured by RAID level 3 will be described. FIG. 11 shows a RAID
It is a figure which shows the procedure which changes from the level 3 to the RAID level 5. FIG. 11 shows that the data reorganization has been completed up to the second stage of the second RAID disk system. Here, the data in the third row of the first RAID disk system 1 is transferred to the second RAID disk system 5
An example of performing reorganization will be described. The input means 100 is the first
The data of the third stage of the RAID disk system 1 is accessed and 12 pieces of data B24 to B25 are input. It should be noted that BXX represents the byte number of the data. In addition, 11 blocks are usually 4K.
Although it is in the B position, in FIG. 11, 4 bytes are set as one block for ease of explanation. For example, B24, B2
4 bytes of 7, B30, B33 are one block. The input unit 100 locks the three blocks from the first RAID disk system 1 to B24 to B35, and inputs the above data. Then, the input means 100
Sends the data to the format changing means 108. The format changing unit 108 reorganizes the data received from the input unit 100 so that the data conforms to RAID level 5. Since RAID level 5 is block interleave, it is necessary to change the byte interleave data input from RAID level 3 to block interleave. Here, the byte interleave data from B24 to B35 is changed to block interleave. The data to be written in the device number D0 of the disk device 3 is B24, B25, B26, B27 in order from the top.
And this is the new data “6”. Next, the data to be written in the device number D1 is B28, B29,
B30 and B31, which are set as new data “7”.
Further, redundant data is written in the device number D2. The calculation of redundant data will be described later. Finally, the data to be written in the device number D3 is B32, B33, B34, B35, which is the new data "8". Then, the parity that is redundant data is calculated from the data "6", "7", and "8". As a method of calculating the parity, the exclusive OR of the three data is obtained, and this is used as the parity. The format changing unit 108 transmits the data reorganized by the above method to the output unit 102. The output means 102 is the second R
The format changing means 10 is provided in the third stage of the AID disk system 5.
The data “6, 7, P, 8” received from 8 is rewritten. After that, the value of the completion pointer is changed to a value indicating the completion up to the third stage, and the lock applied to B24 to B35 (data “6” to “8” in the second RAID disk system 5) is released. . Also, the second RAI
The access from the host system to the D disk system 5 corresponds to the access request from the host system described in the fourth embodiment in the same manner. The change is completed when the completion pointer reaches the final address in the first RAID disk system 1.

Further, the RAID level 4 and the RAID level 5 are the same in the type of data in one group, and are different only in the position of parity which is redundant data. Therefore, the procedure for reorganizing data from RAID level 3 to RAID level 4 is the same as the procedure shown in FIG. 11 described above, and a description thereof will be omitted.

Next, an example of constructing the first RAID disk system 1 having the RAID level 5 to the second RAID disk system 5 having the RAID level 3 will be described. FIG. 12 is a diagram showing a procedure for changing from RAID level 5 to RAID level 3. As explained in FIG. 11 above, RAID
The difference between level 5 and RAID level 3 is the difference between block interleaving and byte interleaving. Therefore, the format changing unit 108 converts the data, which is the block interleave, input from the first RAID disk system 1 configured by RAID level 5 into
In order to reorganize the data in the second RAID disk system 5 composed of RAID level 3, the format is changed to byte interleave. FIG. 12 shows the second RAI
It shows that the data organization up to the second stage of the D disk system 5 is completed. Here, the second RAID
An example of rebuilding the third stage of the disk system 5 will be described. The input unit 100 locks three blocks of data “6”, data “7”, and data “8” from the third stage of the first RAID disk system 1 and then inputs them. The input unit 100 transmits the input data to the format changing unit 108. The format changing unit 108 converts the received data into byte interleave. In this embodiment, one block has 4 bytes, as described above. Therefore, the data “6” received from the input means 100 is converted into “B24 to B27”. Then, the data “7” is converted into “B28 to B31”. Further, the data “8” is converted into “B32 to B35”. Further, the format changing unit 108 converts the converted data into a format in which the data is written in the four disk devices 3 forming the second RAID disk system 5. The device number D2 of the disk device 3 is "B24, B
27, B30, B33 ”is written. Also, device number D
In the disk device of No. 1, "B25, B28, B31, B
34 ”is written. "B26, B29, B32, B35" are written in the disk device 3 having the device number D2.
Further, redundant data is written in the disk device 3 having the device number D3. Redundant data is 3 from device number D0 to D2.
The exclusive OR is calculated from the data to be written in one disk device, and is calculated. For example, "B24", "B25",
The exclusive OR of the data of "B26" is calculated and used as the parity which is the redundant data. Next, the format changing means 1
08 transmits the format-modified data to the output unit 102. The output unit 102 rewrites the data received from the format changing unit 108 in the third stage of the second RAID disk system 5. After the writing, the lock applied to the data is released as described with reference to FIG. Then, the value of the completion pointer is changed to a value indicating completion up to the third stage. The access from the host system to the second RAID disk system 5 is handled by the same method as in the above-mentioned fourth embodiment. First RA
The change is completed when the data reorganization is completed up to the final address of the ID disk system 1.

Also, as explained above, the RAI
The D level 4 and the RAID level 5 have the same type of data in one group and are redundant data, but have different parity positions. Therefore, the procedure for reorganizing data from RAID level 4 to RAID level 3 is similar to the procedure shown in FIG. 12 described above. Therefore, RA
Description of the change from the ID level 4 to the RAID level 3 is omitted.

As described above, RAID level 3 and RAI
D level 4 and above are recorded in units of data of different formats. This is the difference between byte interleaving and block interleaving, as explained above. In the above example, the data size of the data format used in RAID level 4 or higher is larger. Therefore, RAID
Data is input from RAID level 3 with the data size of the data format used in level 4 or higher as a unit. The format changing means newly changes the format of the input data to RA when changing from RAID level 3 to RAID level 4 or higher.
Change to a data format of ID level 4 or higher. When changing from RAID level 4 or higher to RAID level 3, R
Data in the RAID level 3 format is created by dividing blocks of AID level 4 or higher. As a result, a user who uses the disk system according to the present invention can easily change the level without being aware of the difference in data formats such as byte interleaving and block interleaving.

Example 6. In this embodiment, from the first RAID disk system configured with RAID level 1, R of RAID level 0, 3, 4, 5
An example of rebuilding the second RAID disk system configured by the AID level will be described. The change of the RAID level is characterized by using a change management table. A detailed description of the change management table is given below.

First, an example of reconstructing a second RAID disk system configured by RAID level 5 from a first RAID disk system configured by RAID level 1 will be described. FIG. 13 is a diagram showing a procedure for changing from RAID level 1 to RAID level 5. FIG. 14 is a diagram showing a continuation of the procedure for changing to RAID level 5 in FIG. FIG. 15 is a diagram showing a change management table in changing from RAID level 1 to RAID level 5. FIG. 16 is a flowchart showing the procedure for creating the change management table in FIG. FIG. 13 shows that the data reorganization up to the second level of the second RAID disk system 5 has been completed, and a procedure for performing the data reorganization up to the second level will be described as an example. First
The RAID disk system 1 of 1 is configured by RAID level 1. Therefore, the data has a mirror configuration. The first RAID disk system 1 in FIG. 13 includes four disk devices 3. The redundant data of the disk device number D0 is stored in the disk device of the device number D1, and the redundant data of the data stored in the device number D2 is stored in the disk device 3 of the device number D3. The data to be stored in the first stage of the second RAID disk system 5 is the data “0, 1, 2”.
And the parity “P”. The above four data are
When rewriting to the RAID disk system 1 of
The data “100” written in the device number D2 of the first RAID disk system 5 is destroyed. Therefore, the input unit 100 calculates the group including the data “100” in the second RAID disk system 5. In the second RAID disk system 5, the group including the data “100” is the data “99,10”.
0, 101 ". Thereby, the input means 100
The first RAID disk system 1 is accessed four times. In the first access, the data “0,100” is input. In the second access, data "1,101"
Enter In the third access, enter "2",
The data “99” is input at the fourth access. In addition,
The input unit 100 locks data to be input before inputting. The input unit 100 transmits the input data to the level changing unit 109 after the input of the necessary data is completed. Further, the input unit 100 updates the change management table 110 based on the input information. Here, the input means 100
The update of the change management table 110 by means of will be described.

When inputting data from the first RAID disk system 1 for the first time, the input means 100 sets a new level of data in the change management table 110. FIG.
In No. 3, the second RAID disk system 5 is RAID
Since the level 5 is used, the arrangement of data stored in each stage is as shown in FIG. According to FIG. 15, the data to be stored in the first stage of the second RAID disk system 5 is “0 to 2”, and the data to be stored in the second stage is “3 to 5”. In this way, the numbers of the data to be stored in all the stages of the second RAID disk system 5 are set as the new-level data numbers of the change management table 110 (S1 in FIG. 16). Furthermore,
The input means 100 can also be used for locked data.
The value of the lock state of the change management table 110 is “OFF”
To "ON" (S2 in FIG. 16). Input means 1
The update processing of the change management table 110 by 00 has been described above.

Next, the processing of the level changing means 109 will be described. The level changing unit 109 changes the arrangement of the data stored in the first and second stages of the second RAID disk system 5 based on the data received from the input unit 100. In the first stage of the second RAID disk system 5,
Based on the data “0, 1, 2” and the above three data, the parity which is the redundant data for which the exclusive OR is calculated is stored. Furthermore, in the second row, the data “99, 100, 10
1 ”and the exclusive OR from the above three data,
This is written as parity which is redundant data. The level changing means 109 uses the rearranged data “P,
0, 1, 2 ”and“ P, 99, 100, 101 ”are transmitted to the output means 102. Further, the level changing means 109 is
The “changed” in the change management table 110 is updated. FIG.
5, the level changing means 109 sets the new level data “0 to 2” and “changed” of “99 to 101” to “N”.
Update from "O" to "YES". Level changing means 109
The above is the processing.

Next, the processing of the output means 102 will be described. The output unit 102 rewrites the data received from the level changing unit 109 into the corresponding disk device of the second RAID disk system 5. In the first writing, the data “P, 0, 1, 2” is written in the first stage of each of the disk device numbers D0 to D3. In the second writing, the data “P, 99,100,101”
Are written in the second stages of the disk device numbers D0 to D3, respectively. Further, the output unit 102 updates the change management table 110 based on the output information. The output unit 102 is the “location” of the change management table 110 in FIG.
At which step of the second RAID disk system 5 the data has been written is set. Here, "first stage" is set to the "location" of the new level data "0-2". Furthermore, "second stage" is set to the "location" of the new level data "99 to 101" (S4 in FIG. 16). The "lock state" of the new level data "0-2" and "99-101" is changed from "ON" to "OFF". Then, the locks on the data “0 to 2” and the data “99 to 101” are released. The above is the update processing of the change management table 110 by the output means 102.

Next, referring to FIG. 14, the data reorganization processing in the second and subsequent stages in the second RAID disk system 5 will be described. The input unit 100 reorganizes the data in the second RAID disk system 5 for the data having the smallest address among the unprocessed data in the first RAID disk system 1. Second RAI
The D disk system 5 has completed the processing up to the second stage. Therefore, consider the data stored in the third and subsequent stages. The data scheduled to be stored in the third row is "3, 4,
5 ". If the above data is overwritten on the third level of the second RAID disk system 5, the data “102” will be destroyed. Therefore, the group including the data “102” is stored in the fourth row of the second RAID disk system 5. The group including the data “102” is “10
2, 103, 104 ". Therefore, the input means 100
Accesses the first RAID disk system 1 four times and inputs necessary data. In the first access,
Input the data “102”. In the second access,
Input the data “3,103”. In the third access, the data “4,104” is input. In the fourth access, data "5" is input. Then, the inputted information is transmitted to the level changing means 109. Further, the change management table 110 is updated based on the input information as described above. The update process is the same as the process content described above, and here, the change management table 1
The update contents of 10 are omitted. The level changing means 109 is
The data received from the input unit 100 is rearranged in the order of the data in the second RAID disk system 5.
Further, the level changing means 109 has the change management table 11
0 is updated based on the change information. The update processing is the same as the processing content described above, and thus the description thereof is omitted here. Next, the level changing unit 109 transmits the rearranged data to the output unit 102. The output unit 102 writes the data based on the data received from the level changing unit 109, as in the third and fourth rows shown in FIG. Then, the completion pointer is changed to a value indicating completion of the fourth stage. And the output means 1
02 performs update processing on the change management table 110 based on the output information. The change process is the same as the process content described above, and the description of the update process content is omitted here. Further, the data locked by the input means 100 is unlocked.

Then, the input means 100 uses the second RAI.
Data reorganization of the fifth and subsequent stages in the D disk system 5 is performed. The data scheduled to be stored in the fifth level of the second RAID disk system 5 is “6, 7, 8”. In the fifth stage of the second RAID disk system 5, the above "6.
Even if the data of 8 ”is rewritten, there is no data destroyed in the first RAID disk system 1. Therefore, the input unit 100 locks the data “6”, “7”, and “8” by inputting the first RAID disk system 1 three times and then inputs the data. The input data is transmitted to the level changing means 109. Then, the change management table 110 is also updated by the procedure described above. Level changing means 109
Converts the sequence of data to be written in the second RAID disk system 5 based on the data received from the input means 100. Furthermore, the data “6”, “7”, “8”
The exclusive OR of is calculated and this is taken as parity which is redundant data. Further, the level changing unit 109 updates the change management table 110 based on the change information according to the procedure described above. The level changing unit 109 transmits the rearranged data to the output unit 102. The output unit 102 transfers the data received from the level changing unit 109 to the second RAID disk system 5 as shown in FIG.
Rewrite as shown in (B). And
The value of the completion pointer is changed to a value indicating that the processing up to the fifth stage has been completed. Further, the output unit 102 updates the change management table 110 based on the output information in the procedure as described above. Then, the data lock is also released.

When the reorganization of the data from the first RAID disk system 1 to the second RAID disk system 5 is completed by the above procedure, the data swap operation is performed as the next step. In the rewriting of the data, the data is not stored at the position where the data should be originally stored in consideration of the destroyed data. Therefore, a data swap operation is required. The swap operation is performed by the rearrangement unit 114 with reference to the change management table 110 from the smallest address in the second RAID disk system 5. For example, data "0-2"
Is stored in the correct location, so no swap operation is required. Next, since the data “3 to 5” is the data that should originally be placed in the second stage, the data is swapped from the third stage to the second stage. At this time, the data "99-10
1 ”is saved in the buffer. After that, the data “99 to 101” is swapped from the buffer to the 34th stage, which should be originally written. The data written in the 34th row is stored in the buffer before "99 to 101" is overwritten. In this way, data swap is performed in order from the smallest address of the second RAID disk system 5. In the swap operation, the redundant data (P) stored in each stage is also rewritten to the correct position (disk device). It should be noted that the data to be swapped is locked after the swap processing, and the lock is released after the swap processing. When the swap operation is completed for all stages, the rearrangement means 11
4 is the RA in the second RAID disk system 5.
The final address stored in the ID controller 2 is
Change to the new final address value. Then, the changed address is notified to the host system, and this is used as the completion notification of the changing operation.

When the host system accesses the second RAID system during the above-mentioned change operation,
It corresponds as follows. If the value of “updated” in the change management table 110 of the data to be accessed is “NO”, the data is accessed based on the RAID level in the first RAID disk system 1. If “changed” in the change management table is “YES”, the second position is determined based on the position set in “location”.
The data is accessed based on the RAID level in the RAID disk system 5. Even when the data is locked, access is performed by the same method as when the value of “changed” described above is “NO”. Note that the lock management is performed and the change management table 11 is performed even when the relocation unit 114 is performing the swap processing.
0 is updated based on the swap information. Therefore, even if the host system accesses the second RAID disk system during the swap processing, the data can be read from the position where the data is stored by referring to the change management table 110.

The RAID level 4 is the same as the RAID level 5 in the type of data constituting one group, and is different only in the position of parity which is redundant data. Therefore, from the first RAID disk system configured from RAID level 1 to RAID level 4
The procedure for reconstructing the second RAID disk system configured by the above is similar to the method for reconstructing the second RAID disk system configured by RAID level 5 described above. You can

Next, an example of reconstructing a second RAID disk system configured by RAID level 3 from a first RAID disk system configured by RAID level 1 will be described. FIG. 17 is a diagram showing a procedure for changing from RAID level 1 to RAID 3. FIG. 18 is a diagram showing a continuation of the procedure for changing to RAID level 3 in FIG. FIG. 19 is a diagram showing a change management table in changing from RAID level 1 to RAID level 3.

First, the procedure for reorganizing the data up to the second stage of the second RAID disk system 5 will be described with reference to FIG. In RAID level 1, data is configured by block interleaving.
The D level 3 is composed of data by byte interleaving. Therefore, when data is reorganized from the first RAID disk system 1 configured by RAID level 1 to the second RAID disk system 5 configured by RAID level 3, data is transferred from block interleave to byte interleave. Need to be organized. The input unit 100 predicts the data to be stored in the first stage of the second RAID disk system 5. As a result of the prediction, the data stored in the first row is the data “0”, “1”, and “2” (however, the second RAID
The disk system 5 is composed of RAID level 3. Therefore, in reality, "B0", "B1", ...
Data is converted into byte interleave and stored as shown in. ). The above data “0”, “1”, and “2” are changed to the second
The data "100" is destroyed by storing the data in the first stage of the RAID disk system 5. Therefore, the redundancy group including the data “100” is stored in the second row. The redundancy group containing the data “100” is
These are "99", "100", and "101". For this reason,
The input unit 100 is the first RAID disk system 1
Access 4 times and input data. Data “0” and “100” in the first access, data “1” and “101” in the second access, data “2” in the third access, and data “9” in the fourth access.
Enter "9" after locking it. Then, the input unit 100 further updates the change management table 110 based on the input information. The input means 100 is the second RAID
When the data to be stored in the first stage of the disk system is input, the "new level data number" of the change management table 110 is set. For the setting, set the data numbers that are planned to be stored in order from the top, and then step down one by one.
Set the data numbers to be stored in the third and third rows.
In FIG. 19, the setting of the "new level data number" is "B0 to B11", "B12 to B23", ...
Set as “B788 to B799”. In addition, "B x
“X” indicates the number of bytes of data in RAID level 3. In this example, one block in RAID level 1 (one block is a unit in which the RAID controller 2 accesses one disk device) is 4 bytes. Further, the input means 100
After the setting of the "new level data number" is completed, the data in which the "lock state" of the change management table 110 is input is changed from "OFF" to "ON".

The data "0" in the first RAID disk system 1 corresponds to the data in the second RAID disk system 5
Then, it corresponds to the data “B0, B1, B2, B3”.
Then, the data “1” is the data “B4, B5, B6,
"B7". Also, the data “2” is the data “B
8, B9, B10, B11 ”. Furthermore, the data "99" is "B396, B397, B398, B39.
9 ”, the data 100 corresponds to the data“ B400, B
401, B402, B403 ”and data“ 101 ”
Is the data “B404, B405, B406, B40
7 ". Therefore, in the “lock state” in the change management table 110 of FIG. 19, if the value set in the “new level data number” matches the value converted into the above byte interleave, it is changed from “OFF” to “ON”. "
Change to

Further, the input means 100 uses the first RAID
The data input from the disk system is transmitted to the level changing means 109. The level changing unit 109 converts the block interleaved data into the byte interleaved data so that the data received from the input unit 100 conforms to the RAID level of the second RAID disk system. Further, the level changing means 109 uses the second RA.
The arrangement of the data is converted so that the data can be output according to the configuration of the disk device that constitutes the ID disk system 5. In this example, the data "B0, B3," is output as the data to be output to the first stage of the device number D0 of the disk device.
B6, B9 ", data" B1, B4, B7, B10 "as data to be output to the first stage of device number D1 and data" B2, B5, "to be output to the first stage of device number D2. B8, B11 ”data sequence is converted. Then, the parity that is the redundant data stored in the device number D3 is calculated. The parity is calculated by obtaining an exclusive OR based on the data stored in the disk device numbers D0 to D2 and using this as the parity. For example, the data “B
The data is arranged so that the exclusive OR of "0", "B1", and "B2" is obtained and stored in the uppermost stage of the disk device number D3. Level changing means 109 is input means 100
The received data is the data for two stages in the second RAID disk system 5. For this reason, the level changing unit 109 converts the sequence of the data to be stored in the second stage. Further, the parity is also calculated, and the calculated parity is added to the data.

The level changing means 109 converts the arrangement of the data as described above, and sends the converted data to the output means 102. An image of the data transmitted to the output means 102 is shown in FIG.
And "data to be output in the second row".

Further, the level changing means 109 sets "changed" in the change management table 110 to "OF" based on the change information.
Update from "F" to "ON". The update is output means 102.
"New level data number" that matches the data sent to
Do against. In FIG. 19, the value of “B0 to B11” is changed from “OFF” to “ON”.

The output means 102 is the level changing means 109.
The more received data is rewritten in the second RAID disk system 5. Figure 17 shows the rewritten image
Shown in Then, the value of the completion pointer is changed to a value indicating the end of the second stage. Moreover, the lock of the data applied by the input means 100 is released. Furthermore, the output means 10
2 sets the output position in the “location” of the change management table 110 based on the output information. Then, the "lock status" for the unlocked data is changed from "ON" to "O".
FF ”. In this example, the "new level data number" is set to "B0 to B11" as "first stage". Then, "B396 to B407" are set as "second stage".
Further, the value of the "locked state" of the "new level data number" similar to the above is set to "OFF".

Then, the input means 100 uses the second RAI.
The data to be stored in the third stage of the D disk system 5 is predicted. As a result of the prediction, the data scheduled to be stored in the third row are the data “2”, “3”, and “4”. Will be destroyed. Therefore, it is considered that the input unit 100 stores the redundant group including the data “102” in the fourth row. Data "10
The redundancy group including “2” is data “102, 103,
104 ". For this reason, the input means 100 uses the RAI
Access the D disk system 1 four times and input the necessary data. In the first access, data "102"
Enter In the second access, the data “3,10
Enter 3 ”. In the third access, data "4
Enter "104". In the fourth access, data "5" is input. The input unit 100 locks the input data in advance. After that, the input unit 100 transmits the input data to the level changing unit 109. In addition, the input unit 100 updates the “lock state” of the change management table 110 based on the input information. The updating method is the same as the method described above.

Next, the level changing means 109 converts the data received from the input means 100 from block interleave to byte interleave. Then, the converted data is transferred to the third stage and the fourth stage of the second RAID disk system 5.
Converts the data stored in the second row. Further, the parity to be stored in the disk device having the device number D3 is calculated, and this is also added to the data array. And the level changing means 1
09 transmits the changed data to the output means 102.
Further, the level changing unit 109 updates “changed” in the change management table 110 based on the change information. The updating method is the same as the method described above.

Next, the output means 102 rewrites the data received from the level changing means 109 to the third and fourth stages of the second RAID disk system 5. Further, the value of the completion pointer is changed to a value indicating that the reorganization up to the fourth stage is completed. Then, the output means 102
Updates the “location” and “lock state” of the change management table 110 based on the output information. The updating method is the same as the method described above. Then, the output unit 102 unlocks the data. Second RA above
FIG. 18A is a diagram showing an image of the data reorganization of the third and fourth stages of the ID disk system 5.

Then, the input means 100 uses the second RAI.
The data to be stored in the fifth stage of the D disk system 5 is predicted. As a result of the prediction, it can be seen that the data scheduled to be stored in the fifth level of the second RAID disk system 5 is the data “6, 7, 8”. The above data is second
When stored in the fifth level of the RAID disk system 5 of No. 1, no data is destroyed. Therefore, in the data reorganization of the fifth row, only the data for creating the fifth row need be input. Therefore, the input unit 100 accesses the first RAID disk system 1 three times.
In the first access, data "6" is input. In the second access, data "7" is input. At the third access, data "8" is input. Input means 10
With 0, the input data is locked in advance. After that, the input means 100 changes the level changing means 109.
Send the data entered in. Further, the input means 100
Updates the “lock state” of the change management table 110 based on the input information. The updating method is the same as the method described above.

Next, the level changing means 109 converts the data received from the input means 100 from block interleave to byte interleave. Then, the converted byte interleaved data is converted into an array of data to be output to the second RAID disk system 5. Further, the parity stored in the disk device number D3 is also calculated. The calculated parity is added to the data sequence. Then, the level changing unit 109 transmits the rearranged data to the output unit 102. Further, the level changing means 109
Updates "changed" in the change management table 110 based on the change information. The updating method is the same as the processing described above.

Next, the output means 102 rewrites the data received from the level changing means 109 in the fifth stage of the second RAID disk system 5. Further, the value of the completion pointer is changed to a value indicating that the data reorganization up to the fifth stage has been completed. Furthermore, the output unit 102 updates the “location” and “lock state” of the change management table 110 based on the output information. The updating method is the same as the processing described above. Then, the data lock is released. FIG. 18B is a diagram showing an image in which the fifth stage data reorganization of the second RAID disk system 5 is performed.
Shown in

According to the procedure described above, the first R
After completing the reorganization for all the data of the AID disk system 1, the relocation means 114 sets the second RAID
The data is swapped to a position where the data should have been originally written in the disk system 5. Swap is performed by referring to the change management table 110. After the swap processing is completed, the rearrangement unit 114 changes the final address stored in the memory of the RAID controller 2 to the value of the final address in the second RAID disk system 5. Then, this final address is notified to the host system, and this is used as a completion notification. Furthermore, when an access request is issued from the host system to the second RAID disk system 5 during the data reorganization, the first RA configured by the RAID level 1 described above is used.
The same measures are taken as in the case of reorganizing data from the ID disk system 1 to the second RAID disk system 5 configured by RAID level 5.

Next, an example of reconstructing a second RAID disk system configured by RAID level 0 from a first RAID disk system configured by RAID level 1 will be described below. The change management table 110 is also used in this reconstruction. FIG. 20 shows RAID
It is a figure which shows the procedure which changes from level 1 to RAID level 0. FIG. 21 is a diagram showing a continuation of the procedure for changing to RAID level 0 in FIG. FIG. 22 shows RAID
It is a figure which shows the change management table in the change from level 1 to RAID level 0.

In this example, a procedure for reorganizing data from the first stage of the second RAID disk system 5 shown in FIG. 20 will be described. First, the input means 10
0 predicts the data to be stored in the first stage of the second RAID disk system 5. As a result of the prediction, the data to be stored in the first stage is the data “0, 1, 2, 3”
It can be seen that it is. Further, if the above data is rewritten in the first stage, the data “100” in the first RAID disk system 1 will be destroyed. For this reason,
The second tier in the second RAID disk system 5 is also created at the same time. 2 of the second RAID disk system 5
In the tier, “100, 101, including data“ 100 ”,
102, 103 ”is stored. Therefore, the input means 10
0 accesses the first RAID disk system 1 four times and inputs necessary data. In the first access, the data “0,100” is input. In the second access, the data “1,101” is input. At the third access, the data "2, 102" is input. In the fourth access, the data “3,103” is input. The input unit 100 locks the input data in advance. After that, the input unit 100 transmits the data input to the level changing unit 109. Further, the input means 100 uses the input information to change management table 110.
"Lock status" of is updated. The updating method is the same as the processing described above.

Next, the level changing means 109 converts the arrangement of the data based on the data received from the input means 100 so as to match the arrangement of the disk devices 3 constituting the second RAID disk system 5. . Then, the converted data is transmitted to the output means 102. Further, the level changing unit 109 updates “changed” in the change management table 110 based on the change information. The updating method is the same as the processing described above.

Then, the output means 102 rewrites the data received from the level changing means 109 in the first and second stages of the second RAID disk system 5. Then, the value of the completion pointer is changed to a value indicating the completion up to the second stage. It also unlocks the data. Furthermore, the output unit 102 uses the change management table 110 based on the output information.
"Location" and "Lock status" of are updated. The updating method is the same as the above. In FIG. 20, the second RAI
The figure which performed the data reconstruction of the 1st step of the D disk system 5 and the 2nd step is shown.

Next, the input means 100 uses the second RAID
The data to be stored in the third stage of the disk system 5 is predicted. As a result of the prediction, the data to be stored in the third row are "4, 5, 6, 7". By storing the above data in the third tier of the second RAID disk system 5, the data of the first RAID disk system 1 is not destroyed. Therefore, the input means 100
The first RAID disk system 1 is accessed four times and necessary data is input. In the first access, data "4" is input. In the second access, data "5" is input. In the third access, data "6" is input. In the fourth access, data "7" is input. The input unit 100 locks the input data in advance. Then, the data input to the level changing means 109 is transmitted. Furthermore, the input means 100 uses the change management table 110 based on the input information.
"Lock status" of is updated. The updating method is the same as the above.

Subsequently, the level changing means 109 rearranges the data received from the input means 100 in the order of the data to be stored in the second RAID disk system 5. Then, the rearranged data is transmitted to the output means 102. Further, the level changing unit 109 updates “changed” in the change management table 110 based on the change information. The updating method is the same as the above.

Subsequently, the output means 102 stores the data received from the level changing means 109 in the third stage of the second RAID disk system 5. Then, the value of the completion pointer is changed to a value indicating that the processing up to the third stage has been completed. Then, the output means 102 is the input means 100.
Unlock the data applied by. Furthermore, the output means 1
02 updates the “location” and “lock state” of the change management table 110 based on the output information. The updating method is the same as the above. In FIG. 21, the second RAI
The image which performed the data rearrangement of the 3rd stage of D disk system 5 is shown.

When the data reorganization is completed for all the data in the first RAID disk system 1 by the procedure as described above, the relocation means 114 makes the second RAI.
In the D disk system 5, swap processing is performed so that data is rewritten to a position where data should be originally stored. The swap process is similar to the process described above. After the swap process is completed, the rearrangement unit 114 changes the value of the final address stored in the memory of the RAID controller 2 to the value of the final address in the second RAID disk system 5. Then, the value of the final address is notified to the host system, and this is used as a completion notification. Further, the correspondence when the host system makes an access request to the second RAID disk system 5 during data reorganization corresponds to the RAI described above.
The same method as that for changing from D level 1 to RAID level 5 is used.

As described above, in this embodiment, the RAID system is changed by the disk system changing method according to the present invention.
It is explained that the level 1 is changed to the RAID levels 0, 3, 4, and 5. This embodiment is characterized in that RAID level 1 duplicates data and stores the data in a mirror configuration. In contrast, RAID
Levels 0, 3, 4, and 5 are for generating groups and recording data. Therefore, when changing the data of the RAID level 1 to the RAID levels 0, 3, 4, and 5, a group in which the data based on the RAID level 1 is generated based on any of the RAID levels 0, 3, 4, and 5. Will be overwritten by the data. In order to prevent this overwriting, the RA stored in the overwritten area
The data of ID level 1 is saved in advance by the input means. The input unit also inputs in advance the data to be saved for overwriting and other data of the group to which the data should belong. The level changing means creates data based on the changed level from the input data. If the output means writes the created data in the disk device, the data overwritten by the writing has already been saved, so there will be no problem even if it is overwritten. Groups are created from the saved data and output in order. In this way, the RAID level is changed in order, but since data that is overwritten occurs, the data is saved in advance and the saved data is written in random order. A change management table exists to manage such out-of-order output. The rearrangement unit refers to the change management table and rearranges the data output in random order. By providing this change management table, if the disk system change method according to the present invention is used, R
The system can be easily changed from AID level 1 to RAID levels 0, 3, 4, and 5.

Example 7. In this embodiment, from the first RAID disk system configured with any of the RAID levels 3, 4 and 5 to RAID
An example of rebuilding the second RAID disk system configured from level 0 will be described below. Although the level changing means 111 appears in the following description,
This means is the level changing means 10 in the sixth embodiment.
9 may be a different means or a similar means. However, in the case of using the same means, a new function is to have a function of converting the arrangement of data based on any of the RAID levels 3, 4 and 5 into the arrangement of data based on RAID level 0.

First, from the first RAID disk system 1 configured with RAID level 5, the RAID
An example of rebuilding the second RAID disk system 5 configured from level 0 will be described. FIG. 23 shows RAI.
It is a figure which shows the procedure which changes from D level 5 to RAID level 0. The second RAID disk system 5 of FIG.
Indicates that the data organization up to the second stage has been completed. Therefore, in this example, a procedure for performing the data reorganization of the third stage of the second RAID disk system 5 will be described. According to FIG. 23, the first RAID disk system and the second RAID disk system have the same disk device configuration. Therefore, in order to reorganize the data from RAID level 5 to RAID level 0, the input unit 100 may input data in blocks of 4 from the first RAID disk system. Block is 1
The RAID system is a unit for accessing one disk device 3. Therefore, the input means 100
From the first RAID disk system, data “8,
"9,10,11" is input by one access.
Lock the input data in advance. After that, the input unit 100 transmits the input data to the level changing unit 111. The level changing unit 111 converts the arrangement of data from RAID level 5 to RAID level 0 based on the data received from the input unit 100. The converted data is transmitted to the output means 102. The output unit 102 rewrites the data “8, 9, 10, 11” received from the level changing unit 111 to the third stage of the second RAID disk system 5. Then, the value of the completion pointer is changed to a value indicating the completion of the third stage. After that, the output unit 102 unlocks the data.

The first RAID disk system 1 composed of RAID level 4 is different from the first RAID disk system 1 composed of RAID level 5 only in the position of parity data. Therefore, the first RA configured from RAID level 4
When data is reorganized from the ID disk system 1 to the second RAID disk system 5 configured by RAID level 0, the timing of the data read by the input unit 100 is configured by RAID level 5 described above. The access to the first RAID disk system 1 is the same. Therefore, here, R
The procedure for constructing the second RAID disk system 5 configured by RAID level 0 from the first RAID disk system 1 configured by AID level 4 is omitted.

Next, an example of reconstructing the second RAID disk system configured by RAID level 0 from the first RAID disk system configured by RAID level 3 will be described. Figure 24 shows RAI
It is a figure which shows the procedure which changes from D level 3 to RAID level 0. According to FIG. 24, the first RAID disk system 1 and the second RAID disk system 5 are composed of four disk devices 3. For this reason, the input unit 100 uses the second RAID in the same manner as the conversion from RAID level 5 to RAID level 0 described above.
In order to create one group of the disk system 5, it is sufficient to input 4 blocks of data with one access from the first RAID disk system 1. However, the first R
The AID disk system 1 is composed of RAID level 3. Therefore, in the input means 100, the data of four blocks input from the first RAID disk system 1 is continuous data of "B0 to B12" and discontinuity of "B15", "B18" and "B21". Data will be entered. The data to be stored in the first group of the second RAID disk system 5 is “B
It is continuous data from 0 to B15 ”. Therefore, as it is, the data of "B13" and "B14" are lost. Therefore, the input unit 100 uses the first RAID
The disk system 1 must be accessed twice, data "B0" to "B11" must be input in the first access, and data "B12 to B23" must be input in the second access. In this way, unnecessary data is also input. In this example, the second RAID
A procedure for performing the third stage data reorganization in the disk system 5 will be described.

In order to create the third stage of the second RAID disk system 5, the input means 100 uses the first RAID
A block including data "B32 to B47" is input from the disk system 1. That is, the input unit 100 accesses the first RAID disk system 1 twice,
Data "B24 to B35" is input at the first access. Then, at the second access, the data “B36 to B4
Enter 7 ”. The input data is locked in advance. And lock it. And the input means 10
0 extracts the data necessary for creating the third stage of the second RAID disk system 5 from the input data. The data to be extracted are "B32 to B35" and "B3".
6 to B47 ”. The retrieved data is transmitted to the level changing means 111. The level changing unit 111 converts the data received from the input unit 100 from the byte interleave format to the block interleave format. For example,
Data "8" is generated from the data "B32, B33, B34, B35". Then, the level changing means 11 transmits the rearranged data to the output means 102. The output means 102 outputs the data “8, 9, 10, 11” as the second data.
Rewriting is performed on the third stage of the RAID disk system 5. The output unit 102 rewrites the data in the second RAID disk system 5, and then the input unit 1
The data locked by 00 is unlocked. When the data reorganization up to the final address in the first RAID disk system 1 is completed, the output unit 102 sets the final address stored in the memory of the RAID controller 2 to the final address in the second RAID disk system 5. change. And R
Through AID management tool, etc. (RAID management tool is R
The host system (which is software executed on the memory of the AID controller 2) is notified of a new final address, and this notification is used as a completion notification.

As described above, in this embodiment, an example of changing the RAID level of any of the RAID levels 3, 4 and 5 to the RAID level 0 in the disk system changing method of the present invention has been described. . As described above, RAID level 0 has a configuration that does not have redundant data. Therefore, RAID level 3,
The RAID level data of either 4 or 5 is input in order, the redundant data is removed, and the RAID level is changed and output. By doing so, the user using the disk system using the disk system changing method according to the present invention can easily change the system from RAID level 3, 4, 5 to RAID level 0.

Example 8. In this embodiment, a RAID level change accompanied by a block size change will be described below. The RAID levels that can be changed are as follows. 1. R of RAID level 0, 1, 3, 4, 5
A second RAID disk system configured by RAID level 0 is reconstructed from a first RAID disk system configured by AID level. 2. A first RAI composed of RAID level 1
A second RAID disk system configured by RAID level 1 is rebuilt from the D disk system. 3. RAI of RAID level 1, 3, 4, or 5
From the first RAID disk system configured at the D level to the second RAID level 3 configured at the RAID level 3
Rebuild the RAID disk system. 4. RAI of RAID level 1, 3, 4, or 5
From the first RAID disk system configured at the D level to the second RAID level 4 configured at the RAID level 4
Rebuild the RAID disk system. 5. RAI of RAID level 1, 3, 4, or 5
From the first RAID disk system configured at the D level to the second RAID level 5 configured at the RAID level 5
Rebuild the RAID disk system.

A feature of this embodiment is that the first R
The size changing means 112 is used to convert the data of the AID disk system into the data of the second RAID disk system. The size changing unit 112 includes the number of disk devices connected to the first RAID disk system 1 and the first RAID disk system 1
The group size in the first RAID disk system 1 is calculated based on the block size of
Group size). Then, the group size is calculated based on the number of disk devices connected to the second RAID disk system and the block size of the second RAID disk system 5 (this is referred to as the second group size). Then, based on the first group size and the second group size, the group boundary in the first RAID disk system 1 and the second R
The least common multiple area size where the boundaries of the groups in the AID disk system 5 match is calculated. Then, using the calculated least common area size as a unit, the first RAI
The input means 100 reads data from the D disk system. This embodiment will be described below using a specific RAID level and block size.

First, from the first RAID disk system of 1 block 2 KB configured with RAID level 0, 1 configured with RAID level 0 as well
An example of rebuilding the second RAID disk system of block 3 KB will be described. FIG. 25 is a diagram showing a level change procedure involving a size change from RAID level 0 to RAID level 0. In FIG. 25, the data “Nxx” stored in the second RAID disk system 5 is
The newly resized data is shown.

The input means 100 calculates the first group size and the second group size before inputting data from the first RAID disk system 1. Further, the least common multiple area size is calculated based on the first group size and the second group size. First, a method of calculating the first group size will be described. The first group size is obtained from the number of disk devices 3 configuring the first RAID disk system 1 and the block size of the first RAID disk system 1. According to FIG. 25,
Since four disk devices are connected to the first RAID disk system 1 and one block is 2 KB, the first group size is 4 KB × 8 KB to 8 KB.
Is required. Also, four disk devices 3 are connected to the second RAID disk system 5, and one block size is 3 KB. Therefore, the second group size is calculated to be 12 KB from 4 units × 3 KB.
Further, when the least common multiple is obtained from the first group size “8” and the second group size “12”, the least common multiple range size can be calculated as “24 KB”. Therefore, the input unit 100 inputs a block of 24 KB, which is the least common multiple area size, from the first RAID disk system 1. In FIG. 25, one block of the first RAID disk system 1 is 2 KB, so 12 blocks should be input. Therefore, the input means 1
For 00, block data “0”, block data “1”, block data “2”, ..., Block data “10”, block data “11” are input from the first RAID disk system 1. Lock the input data in advance. After that, the input means 100
Sends the input data to the size changing unit 112.

The size changing means 112 is the input means 100.
Based on the received data, the block size of the second RAID disk system 5 is changed. Figure 25
Then, since the block size of the second RAID disk system 5 is 3 KB, 3 blocks from the beginning of the input data.
Data "N0" is created by using KB as new one-block data. As described above, when the data matched with the block size of the second RAID disk system 5 is created in order, the created data becomes the data of 8 blocks from “N0” to “N7”. When the data is rearranged so as to be stored in the four disk devices that constitute the second RAID disk system 5, the data "N0, N1, N2, N3" and "N4, N
5, N6, N7 ". After that, the size changing unit 112 transmits the rearranged data to the output unit 102.

The output means 102 is a size changing means 112.
The more received data is rewritten in the second RAID disk system 5. When the second RAID disk system 5 is created by the method described above, two groups (groups are
RAID system is the second RAID disk system 5
Is the smallest unit to access. In FIG. 25, one group includes data “N0, N1, N2, N
3 ”. ) You are creating it. The output unit 102 rewrites the data in the second RAID disk system 5 and then changes the value of the completion pointer to a value indicating that the second stage reorganization process is completed. Then, the data lock is released. By repeating the above processing, when the processing is completed up to the final address in the first RAID disk system 1, the host system is notified of the completion.

Next, an example of reconstructing a second RAID disk system of 1 block 3 KB composed of RAID level 0 from a first RAID disk system of 1 block 2 KB composed of RAID level 3 explain. FIG. 26 shows RAID level 3 to RAI.
It is a figure which shows the level change procedure which changes a size to D level 0.

Also in this example, the input means 10 is used as in the above example.
In 0, the first group size and the second group size are obtained, and further, the least common multiple area size is calculated based on the first group size and the second group size. According to FIG. 26, the block size of the first RAID disk system 1 is 2 KB, and the first RAID disk system 1 is composed of four disk devices 3. For this reason, the first group size is 4 units x 2 KB,
KB ”is required. Also, the second RAID disk system 5 has a block size of 3 KB,
The D disk system 5 is composed of four disk devices 3. Therefore, the second group size can be calculated as “12 KB” from 4 units × 3 KB. Further, when the least common multiple area size is calculated from the first group size “8 KB” and the second group size “12 KB”, it is calculated as “24 KB”. Therefore, the input means 10
0 is 24K from the first RAID disk system 1
B, that is, 12 blocks of data are input. Figure 2
6, the data inputted by the input means 100 are the block data “B0, B3” and the block data “B1, B
4 ”, ..., Block data“ B19, B22 ”, and block data“ B20, B23 ”for 12 blocks. Note that the parity data (in FIG. 26, the parity data is indicated by “P”) is not the target of reading. The input unit 100 locks the input data in advance. Then, the data input to the size changing unit 112 is transmitted.

Subsequently, the size changing means 112 rearranges the data received from the input means 100 so that one block has 3 KB. The data created by the rearrangement are eight block data of block data “N0”, block data “N1”, ..., Block data “N6”, and block data “N7”. 8 above
One block data is group data “N0, N1, N2, N in the second RAID disk system 5.
3 "and group data" N4, N5, N6, N7 ". The size changing unit 112 then outputs the output unit 102.
Send group data to.

Then, the output means 102 rewrites the data received from the size changing means 112 in the second RAID disk system 5. This completes the data reorganization of the upper two stages of the second RAID disk system 5 of FIG. The output unit 102 changes the value of the completion pointer to a value indicating completion up to the second stage. Further, the data lock is released. The data is reorganized by the procedure as described above, and when the completion pointer reaches the final address, the output means 102 sets the value of the final address stored in the memory of the RAID controller 2 to the second value. To the value of the final address in the RAID disk system 5. Then, the host system is also notified of the value of the final address, and this is used as a reorganization completion notification.

Next, an example of reconstructing a second RAID disk system of 1 block 6 KB composed of RAID level 0 from a first RAID disk system of 1 block 3 KB composed of RAID level 4 explain. FIG. 27 shows RAID level 4 to RAI.
It is a figure which shows the level change procedure which changes a size to D level 0.

In this example also, the input means 100 calculates the first group size and the second group size, and calculates the least common multiple area size from the first group size and the second group size. According to FIG. 27, the first RAID disk system is 3 KB per block and is composed of four disk devices 3. Also, the second RAI
The D disk system is 6 KB in one block, and is composed of four disk devices 3. However, the first R
Since the AID disk system 1 is RAID level 4, one of the four disk devices 3 is a disk device 3 for parity which is redundant data. Therefore, the first
The group size is 3 units x 3 KB. Than this,
The first group size is “9 KB”. And
The second group size is 4 units x 6KB "24KB"
Can be calculated. Furthermore, the first group size "9KB"
When the least common multiple area size is calculated from the second group size “24 KB”, it can be calculated as “72 KB”. Therefore, the input means 100 inputs 24 blocks of data corresponding to 72 KB from the first RAID disk system 1. That is, the input means 100 from the first RAID disk system 1 of FIG. 27 has 24 blocks including block data “0”, block data “1”, ... Enter the data of. Lock the input data in advance. After that, the input unit 100 transmits the input data to the size changing unit 112.

Subsequently, the size changing means 112 changes the data received from the input means 100 into 1 block of 6 KB. The data obtained by changing to 6 KB are block data “N0” and block data “N
1 ”, ..., Block data“ N10 ”, and block data“ N11 ”are 12 block data. Then, the 12 pieces of block data are rearranged so as to be compatible with the second RAID disk system 5. The data obtained by rearranging is the group data “N0,
N1, N2, N3 "..., Group data" N8,
N9, N10, N11 ". This is the second RAI
This corresponds to data for three stages in the D disk system 5. The size changing unit 112 transmits the rearranged data to the output unit 112.

Then, the output means 102 uses the group data received from the size changing means 112 as the second RAID.
Rewrite to the disk system 5. The data image after writing is shown in the upper three rows of the second RAID disk system 5 in FIG. After that, the output unit 102 changes the value of the completion pointer to a value indicating the completion up to the third stage. Further, the lock of the data applied by the input means 100 is released. When the data reorganization process is performed by the above procedure and the value of the completion pointer reaches the final address in the first RAID disk system, the output unit 102 outputs the final address stored in the memory of the RAID controller 2. The value is changed to the final address in the second RAID disk system 5. And
The host system is also notified of the final address, and this is used as the notification of the data reorganization processing completion.

The RAID level 4 and the RAID level 5 are the same in the type of data forming one group, and are different only in the position of parity. Therefore, R
The first RAID disk system configured by AID level 5 is replaced with the second R disk configured by RAID level 0.
The method for rebuilding the AID disk system is the first R configured by RAID level 4 described above.
This can be performed similarly to the method of reconstructing the second RAID disk system configured by RAID level 0 from the AID disk system.

Next, the procedure for reconstructing the second RAID disk system 5 configured by RAID level 0 from the first RAID disk system 1 configured by RAID level 1 will be described below. FIG. 28 shows
It is a figure which shows the procedure which changes the size from RAID level 1 to RAID level 0. First RAI in FIG. 28
The D disk system 1 is composed of four disk devices 3. And one block is 2 KB. Therefore, when the first group size is calculated, 4 units x 2K
B can be calculated as “8 KB”. But first
The RAID disk system 1 is composed of RAID level 1. Therefore, the stored data has a mirror structure. Therefore, although the four disk devices 3 are connected, the data stored in one group is "0,0" when the upper data is taken as an example.
COPY, 100, 100 COPY "
It can be seen that two data "0" and "100" are stored in this one group. That is, the data is stored in the two disk devices 3. Therefore, the first group size can be calculated as “4 KB” from 2 units × 2 KB.

Then, the second group size is obtained.
The second RAID disk system 5 is composed of four disk devices 3. And one block is 3KB
Is. Therefore, the second group size is 4 units x 3K
From B, it can be calculated as “12 KB”.

Next, the least common multiple area size is obtained based on the first group size and the second group size calculated above. The least common multiple area size can be obtained as “12 KB” from the least common multiple of the first group size “4 KB” and the second group size “12 KB”. Therefore, it is considered that the input unit 100 inputs “12 KB” from the first RAID disk system 1 in units. However, the first RAID disk system 1 is
It is composed of AID level 1. Therefore, RAI
The data of the first RAID disk system 1 required to create the first block data “N0” of the second RAID disk system 5 configured by the D level 0 is stored in the disk device number D0. It is stored in the vertical direction as "0", "1", .... Therefore, in order to create the data of “N0” to “N3” which is the data of the first group of the second RAID system 5, the data “0 to 5” of the disk device number D0 of the first RAID disk system 1 is created. Must be entered. Further, by storing the first group of the second RAID disk system 5, the data “100, 101” stored in the device number D2 of the first RAID disk system 1 will be destroyed. In order to prevent the above data destruction, it is considered to store the group including the data to be destroyed in the second stage of the second RAID disk system 5. However, if a group including data “100, 101” is stored in the second row of the second RAID disk system 5, the data “102” stored in the disk device number D2 of the first RAID disk system 1 will be stored. , Will be further destroyed. Therefore, the group including the data 102 must also be saved. The data necessary for creating a group including the data “100, 101” is the data “96-
101 ”. Further, the data necessary for creating the group including the data “102” is the data “102 to 107”. Therefore, the input means 100
Consider inputting the data required to create the three stages of the second RAID disk system 5 at once. However, when the second RAID disk system 5 is created in three stages, the boundary of the group of the first RAID disk system 1 and the boundary of the group of the second RAID disk system 5 do not match. This is because the first RAID disk system 1 is composed of 1 block and 2 KB, and the second RAID
This is because the D disk system 5 is composed of 1 KB and 3 KB. Creating three stages of the second RAID disk system 5 means reconstructing three blocks of the disk device number D0 of the second RAID disk system 5, that is, 9 KB. In "9KB",
It is not a multiple of 2KB.

For this reason, 4 of the second RAID system 5 is used.
Consider creating columns at the same time. First, the second RAI
"N0-N stored in the first stage of the D disk system 5"
The data of "3" is composed of "0 to 5" of the disk device number D0 of the first RAID disk system 1.
The data “N4 to N7” stored in the second stage of the second RAID disk system 5 is the data “6” of the disk device number D0 of the first RAID disk system 1.
~ 11 ". In addition, the data “N68 to N68” stored in the third stage of the second RAID disk system 5
71 "is created from the data" 96 to 99 "of the disk device number D0 of the first RAID disk system 1 and the data" 100 to 101 "of the disk device number D2. Furthermore, the data “N68 to 71” stored in the fourth level of the second RAID disk system 5 is the first RAI.
It is created from the data “102 to 107” of the disk device number D2 of the D disk system 1.

Therefore, the input means 100 uses the first RAI.
The D disk system 1 is accessed 16 times to obtain the necessary data. In the first access, the data "0, 1
Enter "00". Hereinafter, the input means 100 will be the first R
The AID disk system 1 is sequentially accessed from the top to acquire the data “11,111”. Further, the data “96” is accessed by the thirteenth access, and thereafter, the data “97”, the data “98”, and the data “99” are sequentially input. The input means 100 responds to input data by
Lock in advance. Then, the input data is transmitted to the size changing means 112. Also, the input means 100
Is based on the input information, creates the change management table 110 described in the sixth embodiment, and reorganizes the data of the first RAID disk system 1 into the data of the second RAID disk system 5. Record the condition.

Subsequently, the size changing means 112 changes the size of the data received from the input means 100 so that one block is changed from 2 KB to 3 KB. Furthermore,
The order of data output to the second RAID disk system 5 is changed. Then, the changed data is output means 1
Send to 02. Further, the size changing unit 112 updates the change management table 110 based on the change information by the same method as the level changing unit 109 described in the sixth embodiment.

Then, the output means 102 rewrites the data received from the size changing means 12 in the second RAID disk system 5. Then, the input means 100
Unlock the data applied by. Also, the value of the completion pointer is changed to a value indicating the completion of the fourth stage. Further, the change management table 110 is updated based on the output information by the same method as the method described in the sixth embodiment.

By the above procedure, the first RAI
Data is reorganized from the D disk system 1 to the second RAID disk system 5. After the data reorganization processing is completed up to the final address in the first RAID disk system 1, although not shown in FIG. Method to swap data.
When the swapping process is completed for all the data by the relocation unit 114, the relocation unit 114 executes the RAID
The final address value stored in the memory of the controller 2 is rewritten to the final address in the second RAID disk system 5. Then, the final address is also notified to the host system, and this is used as a completion notification.

Next, from the first RAID disk system 1 configured by RAID level 1,
An example of reconstructing the second RAID disk system 5 having the ID level 1 will be described. FIG. 29 shows R
It is a figure which shows the procedure of resizing AID level 1.
In FIG. 29, the first RAID disk system 1 has 4
It is composed of one disk device 3. And one block is 2 KB. The second RAID disk system 5 is also composed of four disk devices 3. And one block is 3 KB. But,
Since the first RAID disk 1 and the second RAID disk system 5 are configured by RAID level 1, the number of disk devices that store data can be considered to be two for the same reason as above. Therefore, the first
The group size can be calculated as “4 KB” from 2 units × 2 KB. The second group size is 2 units x 3K
From B, it can be calculated as “6 KB”. Then, the maximum common multiple area size can be calculated as “12 KB” from the first group size “4 KB” and the second group size “6 KB”.

In this example, the RAIDs of the first RAID disk system 1 and the second RAID disk system 5 are
The level is 1. Therefore, the data reconfiguration can be performed by dividing it into one mirror configuration configured by the disk device numbers D0 and D1 and one mirror configuration configured by the disk device numbers D2 and D3. That is,
Consider reconfiguring data for 12 KB that constitutes a mirror configuration of one data. For example, the input means 10
0 is data of 12 KB of the second RAID disk system 5. Data to be stored in the first and second stages of the disk device numbers D0 and D1 are input from the disk device number D0 of the first RAID disk system 1 by three accesses. The data “0”, the data “1”, and the data “2” to be input are locked and then input, and the input data is transmitted to the size changing unit 112.

Then, the size changing means 112 transfers the data received from the input means 100 from 1 block to 2 KB.
Change to KB. Then, the order of the data to be rewritten in the second RAID disk system 5 is changed. The changed data sequence is the data "N0, COPY N
0 ”and data“ N1, COPY N1 ”. Then, the data is transmitted to the output means 102.

The output means 102 outputs the received data to the second
The RAID disk system 5 is rewritten. Then, the lock of the data applied by the input means 10 is released. Further, the value of the completion pointer is changed to a value indicating completion up to the second stage. With the above procedure, when the value of the completion pointer reaches the final address in the first RAID disk system 1, the completion is notified to the host system.

Next, the procedure for reconstructing the second RAID disk system 5 composed of RAID level 3 from the first RAID disk system 1 composed of RAID level 1 will be described. Figure 30 shows RAI
It is a figure which shows the procedure which changes the size from D level 1 to RAID level 3. In FIG. 30, the first RAID disk system 1 is composed of four disk devices. And one block is 2 KB. Also, the second
The RAID disk system 5 is also composed of four disk devices. And one block is 3 KB. However, the first RAID disk system 1
Since the RAID level 1 is used, the stored data has a mirror structure, and the data required by the input unit 100 is stored in two disk devices. Therefore, the first group size is
"2 KB x 2 KB" can be calculated as "4 KB". Also, the second
The group size of can be calculated from 4 units × 3 KB. However, the second RAID disk system 5
It is composed of RAID level 3. Therefore, one of the four disk devices 3 is a parity disk device. Therefore, the number of disk devices in which data is stored is three, and the second group size can be calculated as “9 KB” from 3 units × 3 KB. As a result, the maximum common area size is the first group size “4 KB”.
And the second group size "9KB" to "36KB"
Can be calculated.

The input means 100 inputs 36 KB worth of data in the second RAID disk system 5 (corresponding to four stages in the second RAID disk system 5 of FIG. 30). At this time, since the first RAID disk system 1 is composed of RAID level 1,
It is conceivable that some data will be destroyed if four levels of data are created. In the second RAID disk system 5, the data that is destroyed in the first RAID disk system 1 is the data "100 to 105" in the disk device number D2 by reorganizing the data of four stages from the beginning. is there. The above data “10
0 to 105 ”corresponds to data for two stages in the second RAID disk system 5. The data for two stages corresponds to the data “99 to 107” in the first RAID disk system 1. Therefore, the second RAI
In creating four stages in the D disk system 5, data “0,100”, data “1,101”, ..., Data “5,105”, data “6” from the first RAID disk system 1. , ..., Data “8” and data “99” are input in order. Lock the input data in advance. After that, the input unit 100 transmits the input data to the size changing unit 112. Further, the input means 100 updates the change management table 110 based on the input information by the same method as in the sixth embodiment described above.

Then, the size changing means 112 changes one block from 2 KB to 3 KB. Also, the second RA
Since the ID disk system 5 is composed of the RAID level 3, the size changing means 112 changes the data into the data in the byte interleave format. Furthermore, the size changing unit 112 also calculates the parity, which is redundant data, by using the exclusive OR when changing the arrangement of the data. Then, the data to be output to the second RAID disk system 5 is also converted including the parity and the data, and the data including the converted parity is transmitted to the output means 102. Further, the size changing unit 112 updates the change management table 110 based on the change information by the same method as the format changing unit 109 described in the sixth embodiment.

Then, the output means 102 rewrites the data received from the size changing means 112 in the second RAID disk system 5. The rewritten data is stored in the first to fourth stages (NB0 to NB17, NB198 to) in the second RAID disk system 5 of FIG.
NB215). Further, the output means 102 is
The data locked by the input means 100 is unlocked. Then, the value of the completion pointer is changed to a value indicating that the data organization of the fourth row is completed. Further, the change management table 110 is updated based on the output information by the same method as the method described in the sixth embodiment.

When the data reorganization up to the final address in the first RAID disk system 1 is completed by the above procedure, although not shown in FIG. 30, the relocation means 114 will be described in the sixth embodiment. The data swap processing is performed by the same method as the above method. After swap processing is completed for all data,
The relocation unit 114 uses the value of the final address stored in the memory of the RAID controller 2 as the second RAID
Change to the value of the final address in the D disk system 5. Then, the final address is also notified to the host system, and this is used as a completion notification.

Next, an example of reconstructing the second RAID disk system configured by RAID level 4 from the first RAID disk system 1 configured by RAID level 1 will be described. FIG. 31 shows RA
It is a figure which shows the procedure which changes the size from ID level 1 to RAID level 4. In FIG. 31, the first RAID disk system 1 is composed of four disk devices 3. And one block is 2 KB. Also,
The second RAID disk system 5 is also composed of four disk devices 3. And 1 block is 3
It is KB. However, the first RAID disk system 1 is composed of RAID level 1. Therefore, the number of disk devices that store data is two. In addition, the second RAID disk system 5 is R
Since the AID level is 4, one of the four disk devices is a disk device for storing parity, which is redundant data. Therefore, there are three disk devices that store data. Therefore, the first group size is calculated as “4 KB” from 2 units × 2 KB. Also,
The second group size is "9KB" from 3 units x 3KB
Can be asked. Then, the maximum common multiple area size is calculated as "36 KB" from the first group size "4 KB" and the second group size "9 KB". The second RAID disk system in this example is RA
The second RAID disk system in the above example has an ID level of 4, and has a RAID level of 3. RAID level 3 and RAID level 4
Has a difference in whether the data format is block interleave or byte interleave. However, since the types of data forming one group are the same, the second RAID disk system 5 is
Data for four stages is created in the data reorganization process of one time. Further, the data necessary for creating four layers is also read by the input means 100 from the first RAID disk system 1 as in the above example.

Therefore, as in the above example, the input means 100 receives data "0," from the first RAID disk system 1.
100, ..., Data “7,107”, .., Data “8” and Data “99” are input. Input means 10
With 0, the input data is locked in advance. Then, the input data is transmitted to the size changing means 112. Further, the input unit 100 updates the change management table 110 based on the input information by the same method as the method described in the sixth embodiment.

Then, the size changing means 112 changes one block from 2 KB to 3 KB. In this change, the data received from the input means 100 is once converted into a byte interleaved format and reorganized so that one block has 3 KB. Then, the reorganized data is used as the second RAI.
The data of one group in the D disk system 5 is rearranged in accordance with the arrangement. At this time, the parity, which is redundant data, is also calculated using the exclusive OR, and the data is added to the list. The size changing unit 112 transmits the rearranged data to the output unit 102. Further, the size changing unit 112 uses the change information to change management table 11 based on the change information.
0 is updated by the same method as the level changing unit 109 described in the sixth embodiment.

Then, the output means 102 outputs the second RAI.
Data of four stages is rewritten in the D disk system 5 (“N0 to N5” and “N100 to N105” in the second RAID disk system 5 of FIG. 31). Furthermore,
The value of the completion pointer is changed to a value indicating the completion up to the fourth step. Then, the output unit 102 updates the change management table 110 based on the output information by the same method as the method described in the sixth embodiment. And the input means 10
Releases the lock on the data by 0.

When the data reorganization up to the final address value in the first RAID disk system 1 is completed by the above procedure, although not shown in FIG. 31, the rearrangement means 114 uses the same method as in the sixth embodiment. Performs data swap processing. After the swap processing is completed for all the data, a value of "0" or "1" is written to the remaining area so that the parity can be matched and initialized. The rearrangement unit 114 uses the value of the final address stored in the memory of the RAID controller 2 as the second RAID value.
Change to the value of the final address in the disk system 5. Then, the rearrangement unit 114 notifies the host system of the value of the final address, and uses this as a completion notification.

The RAID level 4 and the RAID level 5 are the same in the type of data forming one group, and are different only in the position of parity which is redundant data. Therefore, the first configured from RAID level 1
The method of reconstructing the RAID disk system of No. 2 into the second RAID disk system configured by RAID level 5 can be performed by the same method as the above example.

Next, the procedure for changing the block size from the first RAID disk system 1 configured by RAID level 3 to the second RAID system 5 configured by RAID level 3 will be described below. . FIG.
FIG. 2 is a diagram showing a procedure for changing the size of RAID level 3. In FIG. 32, the first RAID disk system 1 is composed of four disk devices 3. And one block is 2 KB. Also, the second RAI
The D disk system 5 is also composed of four disks. And one block is 3 KB. But,
In RAID level 3, one disk device is used for parity, which is redundant data. Therefore, the first RAID disk system 1 and the second RAID disk system 5 are
Data will be stored in three disk devices.
Therefore, the first group size is calculated as “6 KB” from 3 units × 2 KB. Further, the second group size can be calculated as “9 KB” from 3 units × 3 KB. Therefore, the maximum common multiple area size is calculated as "18 KB" from the first group size "6 KB" and the second group size "9 KB".

In the second RAID disk system 5, the maximum common size area size “18 KB” corresponds to two stages. Then, in the first RAID disk system 1, the maximum common multiple area size “18 KB” corresponds to three stages. This indicates that the group size boundaries match. Therefore, the input unit 100 inputs necessary data from the first RAID disk system 1 three times in order to reorganize two stages of data in the second RAID disk system 5. In the first access, the data “B0 to B5” is input. In the second access, the data “B6 to B1
Enter 1 ”. In the third access, the data "B1
2 to B17 ”is input. For the data to enter,
Lock in advance. Further, the input data is transmitted to the size changing means 112.

The size changing means 112 is the input means 100.
The input data is changed from 2 KB to 3 KB per block. Then, the second RAID disk system 5
Rearrange the order of the data to be output to. The rearranged data is transmitted to the output means 102.

The output means 102 is the size changing means 112.
The more received data is rewritten to the second RAID disk system 5. “B0 to B17” in the second RAID disk system 5 of FIG. 32 is the data after rewriting. The output unit 102 releases the data locked by the input unit 100 after the writing is completed. Then, the value of the completion pointer is changed to a value at which the data reorganization up to the second stage is completed.

The data is reorganized by the above procedure, and when the reorganization is completed up to the final address in the first RAID disk system 1, the host system 2 is notified of the completion.

The RAID level 4 and the RAID level 3, and the RAID level 5 and the RAID level 3 are 1
The types of data making up the group are the same, and the only difference is in the data format: byte interleave or block interleave. Therefore, if the first RAID disk system 1 is configured with the same number of units and the same block size as the first RAID disk system 1, the RAID level 4 or 5 The second RAID disk system 5 composed of level 3 can be rebuilt by the same procedure.

Next, an example of reconstructing the second RAID disk system 5 configured by RAID level 4 from the first RAID disk system 1 configured by RAID level 3 will be described below. FIG.
FIG. 3 is a diagram showing a procedure for changing the size from RAID level 3 to RAID level 4. In FIG. 33, the first R
The AID disk system 1 is composed of four disk devices 3. And one block is 2 KB. The second RAID disk system 5 is composed of four disk devices 3. And one block is 3 KB. However, RAID level 3 and R
In AID level 4, since one disk device is used for parity which is redundant data, the number of disk devices to which data is written is three. Therefore, the first group size is calculated as “6 KB” from 3 units × 2 KB. In addition, the second group size is "9K
B ”is required. Further, the maximum common area size is the first
It is possible to obtain "18 KB" from the group size "6 KB" and the second group size "9 KB". In the second RAID disk system 5, the maximum common multiple area size “18 KB” corresponds to two stages. Further, in the first RAID disk system 1, it corresponds to three stages. This is the first RAID disk system 1
And the second RAID disk system 5 indicate that the boundaries of the group sizes match.

Therefore, the input means 100 uses the first RAI.
"B0" which is data for three stages from the D disk system 1
~ B17 "is input by three accesses. The input data is locked in advance. Then, the data input to the size changing unit 112 is transmitted.

Subsequently, the size changing means 112 changes the data received from the input means 100 from 2 KB to 3 KB in one block. Further, the second RAID disk system 5 is composed of RAID level 4. Therefore, the data format is converted from the byte interleave format to the block interleave format. After the conversion, the size changing unit 112 rearranges the data arrangement of the second RAID disk system 5. At this time, the parity, which is the redundant data in one group, is also calculated using the exclusive OR, and the data is added to the group. The size changing unit 112 transmits the rearranged data to the output unit 102.

Then, the output means 102 rewrites the data received from the size changing means 112 in the second RAID disk system 5. After rewriting,
The value of the completion pointer is changed to a value indicating the completion up to the second stage. Then, the lock of the data applied by the input means 100 is released.

By the above procedure, when the data reorganization up to the value of the final address in the first RAID disk system 1 is completed, the completion notification is sent to the host system.

The RAID level 5 is the same as the RAID level 4 in the type of data constituting one group, and is different only in the position of parity which is redundant data. Therefore, the first configured with RAID level 3
RAID level 5 from RAID disk system 1 of
Second RAID disk system 5 composed of
The procedure for reconstructing can be performed in the same manner as the procedure described above.

Next, the procedure for changing the block size from the first RAID disk system 1 configured with RAID level 4 to the second RAID disk system 5 configured with RAID level 4 will be described below. I do. FIG. 34 is a diagram showing a procedure for resizing the RAID level 4. In FIG. 34, the first RAI
The D disk system 1 is composed of four disk devices 3. And one block is 2 KB.
The second RAID disk system 5 is also composed of four disk devices. And one block is 3 KB. However, RAID level 4 uses one disk device to store parity, which is redundant data. Therefore, the data is stored in the three disk devices. Therefore, the first group size is 3
"6 KB" is calculated from the unit x 2 KB. Further, the second group size can be calculated as “9 KB” from 3 units × 3 KB. Therefore, the maximum common multiple area size is calculated as "18 KB" from the first group size "6 KB" and the second group size "9 KB". Second RA
In the ID disk system 5, the maximum common area size “18 KB” corresponds to two stages. Also, the first RA
The ID disk system 1 corresponds to three stages. This indicates that the boundaries of the groups of the first RAID disk system and the second RAID disk system match. Therefore, the input means 100
Inputs the data of three stages of data "0" to "8" by three accesses from the first RAID disk system 1. The input data is locked in advance. After that, the input unit 100 transmits the input data to the size changing unit 112.

The size changing means 112 is the input means 100.
The received data is changed from 2 KB to 3 KB per block. Then, the second RAID disk system 5
Sort the data in. At this time, the parity, which is redundant data, is also calculated using the exclusive OR, and the data is added to the list. After that, the resizing means 112
The data is transmitted to the output means 102.

The output means 102 rewrites the data received from the size changing means 112 to the second RAID disk system 5. The data “N0 to N5” in the second RAID disk system 5 of FIG. 34 is the data after rewriting. Furthermore, the output means 1
02 changes the value of the completion pointer to a value indicating completion up to the second stage. Then, the output means 102 is the input means 10
Unlocks the data locked by 0.

By the above procedure, when the data reorganization up to the final address in the first RAID disk system 1 is completed, the host system is notified of the completion.

The RAID level 5 is the same as the RAID level 4 in the type of data constituting one group, and is different only in the position of parity which is redundant data. Therefore, the method for reconstructing the second RAID disk system 5 configured by RAID level 5 from the first RAID disk system 1 configured by RAID level 5 is performed by the same method as the above example. be able to.

Next, a procedure for reconstructing the second RAID system 5 having the RAID level 5 from the first RAID disk system 1 having the RAID level 4 will be described. FIG. 35 shows the RAI
It is a figure which shows the procedure which changes the size from D level 4 to RAID level 5. In FIG. 35, the first RAID disk system 1 is composed of three disk devices. And one block is 2 KB. Also,
The second RAID disk system 5 is also composed of four disk devices 3. And 1 block is 3
It is KB. However, in RAID level 4 and RAID level 5, since one disk device 3 is used for parity, there are three disk devices to write data. Therefore, the first group device is "6K" from 3 units x 2KB.
B ”can be calculated. The second group size is 3
It can be calculated as "9 KB" from the unit x 3 KB. And the first
When the maximum common multiple area size is calculated from the group size “6 KB” of the second group and the second group size “9 KB”,
KB "can be calculated. In the second RAID disk system 5, the maximum common multiple area size “18 KB” corresponds to two stages. Further, in the first RAID disk system 1, it corresponds to three stages. This indicates that the group size areas of the first RAID disk system 1 and the second RAID disk system 5 match. Therefore, the input means 100 is the first RAID
Data "0" to "8", which are data for three stages, are input from the disk system 1 by three accesses. The input data is locked in advance. After that, the input unit 100 has the size changing unit 112.
Send data to.

The size changing means 112 is the input means 100.
The received data is changed so that one block is changed from 2 KB to 3 KB. Then, the data arrangement of the second RAID disk system 5 is rearranged. At this time, the parity which is the redundant data is also calculated by using the exclusive OR, and this is added to the data. Then, the size changing unit 112 transmits the rearranged data to the output unit 102.

The output means 102 reorganizes the data of two stages of the second RAID disk system 5. 36 shows data reorganized into data "P, N0 to N5" in the second RAID disk system 5 of FIG. Also,
The value of the completion pointer is changed to a value indicating the completion up to the second stage. Then, the lock of the data applied by the input means 100 is released.

According to the above procedure, when the data reorganization up to the final address in the first RAID disk system 1 is completed, the host system is notified of the completion.

The RAID level 4 and the RAID level 5 have the same type of data forming the group,
Only the position of parity, which is redundant data, is different.
Therefore, the method for reconstructing the second RAID disk system 5 configured by RAID level 4 from the first RAID disk system 1 configured by RAID level 5 is performed by the same method as described above. You can

As described above, according to the disk system changing method of the present invention, the RAID when the block size is different is used.
The level is changed. If the block sizes are different, the group size before the change and the group size before the change are respectively determined from the number of disk devices whose input means is connected to the disk system, the RAID level, the block size before the change, and the block size after the change. Find the later group size. Then, the data is read with the size at which the boundary between the group size before the change and the group size after the change matches as one unit. For example, the group size before change is 6 KB and the block size after change is 9 KB.
If it is, 18 KB is read as one unit. The size changing means creates two 9 KB groups from the read 18 KB data. However, 9K above
Redundant data is not included in B. In this way, the block size can be changed. Therefore, the user can easily realize the change to the system having the different block size.

Example 9. In this embodiment, the above-mentioned embodiment 6 is used.
Regarding the level change from the RAID level 1 to the RAID level 5 described in the above, the procedure for reorganizing data using the disk number table will be described. In the above embodiment, the rearrangement means 114 obtains the device number of the correct disk device that stores the redundant data during the swap processing, and writes the redundant data in the correct disk device. However, in this embodiment, the level changing means 10
When 9 sorts the data, it refers to the disk number table to obtain the correct disk device number for storing the redundant data, and sorts the data.

FIG. 36 shows RAID level 1 to RAID.
It is a figure which shows the procedure which changes to the level 5. FIG. 37 is a diagram showing a disk number table. In FIG. 36,
The first RAID disk system 1 is composed of RAID level 1. The second RAID level system 5 is composed of RAID level 5.
In this embodiment, the input means 100 inputs data from the first RAID disk system 1 and inputs a logical address for the data at the same time as in the sixth embodiment. Then, the level changing unit 109 refers to the disk number table 113 shown in FIG. 37, and refers to the disk number in which the redundant data of the data input in the second RAID disk system 5 should be stored, and the data storage in each disk. The table number indicating the position is calculated. The method of storing data at the positions of the disk number and the table number will be described below.

For example, in FIG. 36, in the second RAID disk system, data "0,1,2" and redundant data for data "0,1,2" and data "9" are used.
9,100,101 ”and the data“ 99,100,10 ”
A method of reorganizing redundant data for "1" will be described. The input means 100 uses the data “0, 1, 2” and the data “99,
100, 101 "is the first RAID disk system 1
Enter more. Further, the input means 100 uses the data “0,
1,2 ”and logical addresses“ 0,1,2 ”and“ 99,100,101 ”for data“ 99,100,101 ”
Also enter at the same time. Note that the data to be input is locked in advance. , Data “0, 1, 2”, data “99, 100, 101” and logical address “0, 1,
2 ”and“ 99, 100, 101 ”are level changing means 10
9

The level changing means 109 is the input means 100.
The logical address thus received is divided by 12 to obtain the remainder. Based on this remainder, the disk number table 1 of FIG. 37
13, the disk device number for storing the data is acquired. In this example, since the logical address for the data “0” is “0”, the remainder obtained from “0/12” is “0”. Referring to the disk number table 113 in FIG. 37, “0” is the disk device number “D
It is understood that the data is stored in "1". Since the logical address for "1" is "0", the remainder obtained from "1/12" is "1". Referring to the disk number table 113 in FIG. 37, it can be seen that "1" is stored in the disk device number "D2". In addition, data “2”
Since the logical address for "2/1" is "2/1"
The remainder obtained from "2" is "2". Referring to the disk number table 113 of FIG. 37, it can be seen that “2” is stored in the disk device number “D3”. Further, a method of calculating the storage device disk device number of the redundant data for the data “0, 1, 2” will be described.

The second RAID disk unit in this example is composed of four disk units 3 and
Since the ID level is 5, it is considered that one group of data including parity has four patterns as shown in FIG. The first pattern 0 is the group “P,
0, 1, 2 ”, the second pattern 1 is the group“ 3, P, 4, 5 ”, the third pattern 2 is the group“ 6, 7, P, 8 ”, the fourth Pattern 3 is a group “7, 10, 11, P”. Data "0, 1, 2"
The remainder for was 0,1,2. This is shown in Figure 3.
Corresponds to pattern 0 in 7. Therefore, referring to the pattern 0 in FIG. 37, it can be seen that the redundant data is stored in the disk device number “D0”. In addition, the data “9
The disk device numbers for storing data from the logical addresses “99”, “100”, and “101” for 9 ”,“ 100 ”, and“ 101 ”are set to“ logical address /
When calculated from the remainder of 12, the data “99” is stored in the disk device number “D0”, the data “100” is stored in the disk device number “D2”, and the data “101” is stored in the disk device number “D3”. I know what to do. The remainder for the data “99, 100, 101” is “3
4, 5 ". This corresponds to pattern 1 in FIG. Therefore, it can be seen that the redundant data is stored in the disk device number “D1”. As a result, the level changing means 109 makes the data “P, 0, 1, 2” and the data “99, based on the data received from the input means 100.
P, 100, 101 ". Then, the rearranged data is transmitted to the output means 102. In addition,
37, the logical address is divided by "12" when the disk device number for storing the data is obtained.
This is because the number of data becomes 12 in total when the parity (“P”) is removed from each pattern.

The output means 102 is the level changing means 109.
Based on the received data, the data is rewritten in the corresponding disk device of the second RAID disk system. The image after writing is the second RAI in FIG.
This is shown in the D disk system 5. The input means 100,
The update of the change management table 110 by the level changing unit 109 and the output unit 102 is performed in the same manner as in the sixth embodiment. When the data reorganization from the first RAID disk system 1 to the second RAID disk system 5 is completed by the above procedure, as in the sixth embodiment,
Swap processing is performed by the rearrangement unit 114. However,
In the swap processing of this embodiment, since the redundant data existing in each stage is stored in the correct disk device number, only the data rearrangement for each stage is performed.

By the above procedure, the input means 100
The data and the logical address of the data are input from the first RAID disk system 1, the disk device number for storing the data is calculated, and the data is rewritten. In the above method, the data in each stage is written in the position that should originally be written. Therefore, as in the sixth embodiment, the storage position of the redundant data existing in each stage in the swap process is calculated and the data is rearranged. Is unnecessary.

The formula for calculating the disk device number used above is the second RAID used in this embodiment.
The number of disk devices that make up the disk system and R
This is due to the AID level and the RAID level that constitutes the first RAID disk system, and the state of the data stored in the disk device. Therefore, from the first RAID disk system different from the above to the second R
When reconstructing the AID disk system, another calculation formula must be used.

Example 10. In addition, in the above-described first to third embodiments, the example of adding the disk device has been described.
However, it is possible to reduce the number of disk devices connected to the system by reducing the size of the system or changing the RAID level to store data efficiently. By using the disk system changing method according to the present invention, it is possible to easily reduce the number of disk devices. Specifically, this can be realized by reorganizing the data by the same method as the method described in the first to third embodiments, and then removing the unnecessary disk device from the system. In the above-described embodiment, the disk is added or the RAID level is changed while receiving the access from the host system. However, it is also possible to stop the access from the host system and perform the offline. In this case, the lock procedure is unnecessary.

[0168]

As described above, according to the present invention, the first
Second RAID of different RAID disk system configuration
It can be changed to an ID disk system.

Further, according to the present invention, the disk device can be effectively used even when it is added in the horizontal direction.

Further, according to the present invention, the disk device can be effectively used even when it is added in the vertical direction.

According to the present invention, the system can be changed between RAID level 4 and RAID level 5.

According to the present invention, the system can be changed between RAID level 3 and RAID level 4 or higher.

According to the present invention, the system can be changed from RAID level 1 to RAID levels 0, 3, 4, and 5.

According to the present invention, the system can be changed from RAID levels 3, 4, 5 to RAID level 0.

Further, according to the present invention, it is possible to change to a system having a different block size.

Further, according to the present invention, since the lock means is provided, it is possible to prohibit access to data during rewriting.

Further, according to the present invention, since the completion pointer is provided, the data before the change and the data after the change can be easily distinguished.

Further, according to the present invention, since the access means is provided, even if the data before and after the change exist in one system, the data can be accessed without contradiction.

Further, according to the present invention, since the remaining area initialization means is provided, even if the area of the data after the change is smaller than the area of the data before the change, the remaining area can be initialized. Therefore, it is possible to easily record the data thereafter.

Further, according to the present invention, data can be efficiently recorded even if a disk is added and the area for recording data is increased, and the disk can be used without waste.

[Brief description of drawings]

FIG. 1 is a system configuration diagram showing an example of horizontally expanding disks according to a first embodiment.

FIG. 2 is a diagram showing an example of a disk extension procedure in RAID level 5 according to the first embodiment.

FIG. 3 is a diagram showing an example of a disk extension procedure in RAID level 3 or 4 according to the first embodiment.

FIG. 4 is a diagram showing an example of a disk extension procedure in RAID level 0 according to the first embodiment.

FIG. 5 is a diagram showing an example of a disk extension procedure in RAID level 1 according to the first embodiment.

FIG. 6 is a system configuration diagram showing an example of vertically extending disks according to the second embodiment.

FIG. 7 is a system configuration diagram showing an example of expanding disks in the horizontal and vertical directions according to the third embodiment.

FIG. 8 is a diagram showing an example of realizing the disk changing method according to the present invention.

FIG. 9 is a diagram showing a procedure for changing from RAID level 4 to RAID level 5 according to the fourth embodiment.

FIG. 10 shows RAID level 3 to RAID level of the fourth embodiment.
The figure which shows the procedure to change to level 5.

FIG. 11 shows RAID level 3 to RAID level 5 of the fifth embodiment.
The figure which shows the procedure to change to level 5.

FIG. 12 shows RAID level 5 to RAID level of the fifth embodiment.
The figure which shows the procedure to change to level 3.

FIG. 13 illustrates RAID level 1 to RAID level 6 according to the sixth embodiment.
The figure which shows the procedure to change to level 5.

FIG. 14 is a diagram showing a continuation of the procedure for changing to RAID level 5 in FIG. 13;

FIG. 15 shows RAID level 1 to RAID level of the sixth embodiment.
The figure which shows the change management table in the change to level 5.

16 is a flowchart showing a procedure for creating a change management table in FIG.

FIG. 17 shows RAID level 1 to RAID level 6 according to the sixth embodiment.
The figure which shows the procedure to change to level 3.

FIG. 18 is a view showing a continuation of the procedure for changing to RAID level 3 in FIG. 17;

FIG. 19 is a diagram showing a change management table in changing from RAID level 1 to RAID level 3;

FIG. 20 illustrates RAID level 1 to RAID level 6 according to the sixth embodiment.
The figure which shows the procedure which changes to level 0.

FIG. 21 is a view showing a continuation of the procedure for changing to RAID level 0 in FIG. 20;

FIG. 22 is a diagram showing a change management table in changing from RAID level 1 to RAID level 0.

FIG. 23 shows RAID level 5 to R in Example 7.
The figure which shows the procedure which changes to AID level 0.

[FIG. 24] RAID level 3 to RAID level of Example 7
The figure which shows the procedure which changes to level 0.

FIG. 25 shows RAID level 0 to RAID level of the eighth embodiment.
The figure which shows the procedure of the level change which changes a size to level 0.

[FIG. 26] RAID level 3 to RAID level of Example 8
The figure which shows the procedure which changes a size to level 0.

[FIG. 27] RAID level 4 to RAID in Example 8
The figure which shows the procedure which changes a size to level 0.

28] RAID level 1 to RAID in Example 8
The figure which shows the procedure which changes a size to level 0.

FIG. 29 is a diagram showing a procedure for resizing the RAID level 1 according to the eighth embodiment.

FIG. 30 shows RAID level 1 to RAID level of Example 8.
The figure which shows the procedure which changes a size to level 3.

[FIG. 31] RAID level 1 to RAID level of Example 8
The figure which shows the procedure which changes a size to level 4.

FIG. 32 is a diagram showing a procedure for resizing the RAID level 3 according to the eighth embodiment.

FIG. 33 shows the RAID level 3 to the RAID level of the eighth embodiment.
The figure which shows the procedure which changes a size to level 4.

FIG. 34 is a diagram showing a procedure for resizing the RAID level 4 according to the eighth embodiment.

FIG. 35 shows RAID level 4 to RAID level of the eighth embodiment.
The figure which shows the procedure which changes a size to level 5.

FIG. 36 shows RAID level 1 to RAID level of the ninth embodiment.
The figure which shows the procedure which changes to the level 5.

FIG. 37 is a diagram showing a disk number table according to the ninth embodiment.

FIG. 38 is an explanatory view of the principle of the conventional invention.

[Explanation of symbols]

1 first RAID disk system, 2 RAID controller, 3 disk device, 4 ports, 5 second
RAID disk system, 100 input means, 10
1 redundant data calculation means, 102 output means, 103
Live line connection means, 104 initialization means, 107 position determination means, 108 format change means, 109 level change means,
110 change management table, 111 level change means,
112 size changing means, 113 disk number table, 114 rearranging means.

Claims (13)

[Claims] 1. A disk system changing method for changing a first RAID (Redundant Arrays of Inexpensive Disks) disk system to a second RAID disk system, which has the following elements: (A) input means for sequentially reading data based on the first RAID disk system, (b) changing means for changing the data input by the input means to data compatible with the second RAID disk system and redundant data, (c) ) Outputting means for outputting the data changed by the changing means and the redundant data. 2. A redundant group is composed of N + M (M is an integer of 1 or more) disk drives from a first RAID disk system in which a redundant group is composed of N (N is an integer of 2 or more) disk devices. Second RAID
In a disk system changing method for changing to a disk system, the disk system changing method is characterized by having the following elements: (a) input means for reading N + M-1 data in order from N disk devices; Redundant data calculation means for obtaining redundant data by using N + M-1 data input by the input means as a new redundant group, (c) N + M-1 data included in the new redundant group and the redundant data calculation means N + the calculated redundant data
Output means for writing to M disk devices. 3. A first RAID disk system comprising N × L (L is an integer of 1 or more) disk devices in which a redundant group is configured by N (N is an integer of 2 or more) disk devices, In a disk system changing method for changing to a second RAID disk system including N × (L + M) (M is an integer of 1 or more) disk devices,
Disk system changing method characterized by having the following elements: (a) Hot-line connection for hot-connecting N × M disk devices to N × L disk devices of the first RAID disk system Means, (b) initialization means for initializing the storage areas of the N × M disk devices connected by the hot-line connection means to a predetermined value, and (c) added by the N × M disk devices. A number-of-areas changing means for increasing the maximum number of areas by the amount of the stored storage area. 4. One of a RAID level 4 based RAID disk system and a RAID level 5 based RAID disk system is a first RAID disk system and the other is a second RAID disk system.
In a disk system changing method for changing a RAID disk system from a first RAID disk system to a second RAID disk system, the disk system changing method is characterized by having the following elements: (a) One RAID level is the same Input means for sequentially reading the data belonging to the redundant group, and (b) the data read by the input means and the redundant data are RA of the other.
Position determining means for determining the arrangement of data and redundant data so as to conform to the ID level, (c) Data of the redundant group in which the data and redundant data are read by the input means according to the arrangement determined by the position determining means. Output means for sequentially rewriting to the area where was recorded. 5. A RAID disk system based on RAID level 3, and a RAID based on RAID level 4 or higher.
Any one of the D disk systems is the first RAID disk system, and the RAID level 3 based RAID
When the D disk system is the first RAID disk system, RAID based on RAID level 4 or higher
One of the disk systems is the second RAID disk system, and RAID based on RAID level 4 or higher
One of the ID disk systems is the first RAID
When the disk system is used, the RAID disk system based on RAID level 3 is used as the second RAID disk system, and in the disk system changing method for changing the RAID disk system from the first RAID disk system to the second RAID disk system,
Disk system changing method characterized by having the following elements: (a) input means for sequentially reading data of one RAID level in units of block size of blocks used in RAID level 4 or higher; A format changing means for changing the data format so that the read data conforms to the other RAID level, and changing to data based on the new format and redundant data, (c) according to the size changed by the size changing means, Output means for sequentially rewriting the data and the redundant data in the area where the data of the redundant group read by the input means was recorded. 6. A first RAI based on RAID level 1
RAID level 0, 3, 4, from D disk system
In the disk system change method for changing the RAID level to the second RAID disk system based on any one of 5, the disk system change method is characterized by having the following elements: (a) One of the second RAID disk systems While inputting the data of the first RAID disk system used for the group, the data of the first RAID disk system recorded in the area overwritten to record the data of the group of the second RAID disk system, The input means for inputting other data included in the group of the second RAID disk system to which the data belongs, and (b) the data read by the input means as the second data.
Level changing means for changing to data of the RAID disk system, (c) output means for sequentially writing the data changed by the level changing means in the area where the data read by the input means was recorded, (d) the above A change management table for managing the position of data whose RAID disk system has been changed by the input means, the level changing means, and the output means, (e) referring to the change management table,
A rearrangement means for changing the arrangement of data. 7. A first RAID disk system based on any of RAID levels 3, 4, and 5 to RAID.
R to second RAID disk system based on level 0
In a disk system changing method for changing an AID level, a disk system changing method characterized by having the following elements: (a) Any one of RAID levels 3, 4 and 5 RA
Input means for sequentially inputting the data of the ID level, (b) level changing means for changing the data inputted by the input means to data of RAID level 0, and (c) an output means for writing the data changed by the level changing means. 8. A RAID level 0, 3, having a first block size, wherein a block is an access unit for each disk device accessed by the RAID disk system.
RAID level 0 having a second block size from the first RAID disk system of any one of
In the disk system changing method for changing to the second RAID disk system of any one of 1, 3, 4, and 5,
Disk system change method characterized by having the following elements: (a) a first block size, the number of disk devices connected to the first RAID disk system, and a first block size
RAID based on the RAID disk system level of
While calculating the group size of the ID disk system, the second block size, the number of disk devices connected to the second RAID disk system, and the second
Second RAID based on the RAID disk system level of
A group size of the ID disk system is calculated, and based on the group size of the first RAID disk system and the group size of the second RAID disk system, a group boundary according to the group size of the first RAID disk system, An input unit that calculates the least common multiple area size in which the group boundaries according to the group size of the second RAID disk system match, and sequentially reads the data of the first RAID disk system in units of the calculated least common multiple area size. b) size change means for changing the data read by the input means to data based on the second block size, and (c) data read by the input means for the data changed by the size change means. An output means for sequentially rewriting to the area. 9. The disk system changing method further comprises lock means for prohibiting access to the data read by the input means until output by the output means. The disk system changing method according to any one of claims 1, 2 and 4 to 8. 10. The disk system changing method according to claim 9, further comprising a completion pointer for storing a position where writing has been completed by the output means. 11. The disk system modification method further refers to a completion pointer, performs access to data after the completion pointer based on the first RAID disk system, and accesses data before the completion pointer. 11. The disk system changing method according to claim 10, further comprising access means for executing the step based on the second RAID disk system. 12. The disk system changing method further includes a detecting unit for detecting an area of data recorded by the first RAID disk system, and a second RA.
When the area for recording data is reduced from the area detected by the detecting means by using the ID disk system, the remaining area initializing means for initializing the remaining area with a predetermined value is provided. The disk system changing method according to claim 11. 13. The disk system changing method further stores a maximum value of a data recording area, and when the area for recording data is increased by adopting the second RAID disk system, the increased area is recorded. 13. The disk system changing method according to claim 3 or 12, wherein the maximum value is updated by the amount.