JP2007293486A - Disk device, disk control method and program - Google Patents

Abstract

[PROBLEMS] To change the configuration of a disk drive group associated with the construction, enhancement, or change of a new system has an impact on customer operations, costs due to the system being shut down, and human costs for configuration design and change. Without actually operating the system, the validity of the set configuration could not be verified.
A disk control unit 4 of a controller 2 counts the number of accesses from a host 1 and compares the number of data accesses within a predetermined period with a threshold for each stripe to determine a high load or low load area. For the stripe determined to be a high load area or a low load area, the copy destination is a stripe size area of the same capacity so that data can be moved to a disk drive prepared in advance for the data update processing after the determined time point. Means for securing, and means for directly writing data to be written in the copy destination area in units of blocks to the write processing to the stripe.
[Selection] Figure 1

Description

  The present invention includes a disk drive group having a plurality of disk drives, a controller that is connected to a host and is operated, and controls the disk drive group, and a host that performs interface control with the host in the controller The present invention relates to a disk device, a disk control method, and a program that include an interface control unit, a storage unit that stores information, and a disk control unit that performs data read / write control on the disk drive group.

(Description of conventional example)
§1: Description of Conventional Example 1 FIG. 12 is an explanatory diagram of Conventional Example 1. Hereinafter, Conventional Example 1 will be described with reference to FIG.

  Conventional example 1 shown in FIG. 12 is an example of a disk device. This disk device is composed of a controller 2 and a disk drive group 6 connected to the controller 2, and is a device that is operated by being connected to the host 1.

  The controller 2 includes a host interface control unit (hereinafter referred to as a “host I / F control unit”) for performing interface control with a plurality of hosts 1-1, 1-2, and 1-3. 3, a disk control unit 4 for performing various disk controls including data read / write control on the disk drive group 6, a storage unit 5 for storing various information (or data), and the like. Yes.

  The disk drive group 6 includes a plurality of disk drives D # 0 to D # n. Each of the plurality of disk drives D # 0 to D # n includes a plurality of disk drives therein to constitute a RAID device (RAID0 to RAID5). Hereinafter, the disk drives constituting these RAID devices will be described simply as “disk drives D # 0 to D # n”.

  The operation of the disk device having the above configuration will be described below.

  (a): Before starting normal operation, the relationship between the host 1-1, 1-2, 1-3 and the number of the disk drive used by the host 1-1, 1-2, 1-3 is At the time of initial introduction, connection relations are set in advance in the storage unit 5 by manual operation by an operator or the like (stored as setting information in the storage unit 5).

  (b): The disk device is connected to the hosts 1-1, 1-2, and 1-3 to start operation. In this case, the host I / F control unit 3 of the controller 2 in the disk device is connected by a path and starts operation.

  (c): During operation, the disk control unit 4 performs data access (data read or write) to the disk drive group 6 by performing control based on information set in the storage unit 5.

  (d): After the operation is started, a load may be concentrated on a specific disk drive from the plurality of hosts 1-1, 1-2, 1-3 due to the data access. In such a state, the following problems occur.

  1. When the system performance cannot be achieved, the processing speed of the system is reduced. As a result, utilization efficiency decreases and maintenance operation costs increase.

  2. When a failure occurs in the disk drive group 6, the range of influence on the connected host is widened.

  (e): In order to solve the above-mentioned problems, it is considered to construct a new system or to change the configuration of the disk drive group 6 in accordance with the enhancement or change of the system. In this case, it is necessary to redesign after examining the operation method and performance requirements for the existing system, and after changing the configuration based on the design after stopping the host system.

  However, in that case, there will be an impact on customer operations, costs due to the system being shut down, and human costs for configuration design and changes, and the validity of the configuration that was set if the system was not actually operated. Cannot be verified.

§2: Description of Conventional Example 2 Hereinafter, Patent Document 1 will be described as Conventional Example 2. Conventional Example 2 describes the following contents.

(a): “A logical volume to be migrated for data migration for system load balancing is selected from access information to physical drives and logical drives under the disk array controller without increasing the management load of the disk array controller. Refer to the summary column.
(b): “a storage means in which each piece of information is stored across a plurality of control means, a means for checking the access status of each storage means connected to the control means and connected to each control means for each piece of information; A storage system comprising: "... see claims section
(c): “An object of the present invention is to provide a system and method for collecting short-term volume information without increasing the load on the entire system and automatically instructing the control apparatus to perform migration. "... see paragraph number [0013].

§3: Description of Conventional Example 3 Hereinafter, Patent Document 2 will be described as Conventional Example 3. Conventional Example 3 describes the following contents.

(a): “In conventional tiered storage devices, data blocks that are not normally accessed are stored in a low-speed storage medium. However, data that is not normally accessed is accessed periodically. However, since these data are not normally accessed, they are moved to a low-speed storage medium, and at the time of access, access having a long response time must be performed once. Provides a hierarchical storage device that eliminates the need for the
(b): “Accurate access fluctuation cycles are obtained by continuously accumulating accesses per unit time of data blocks. With reference to these fluctuation cycles, data blocks are stored in storage with appropriate performance in advance between storage tiers. Move it. "... See the summary column

§4: Description of Conventional Example 4 Hereinafter, Patent Document 3 will be described as Conventional Example 4. Conventional Example 4 describes the following contents.

  (a): “Providing a method and apparatus for adaptively localizing data that is frequently accessed and randomly addressed in DASD to improve system access performance.” See summary section. (b): “At the selected sampling interval, the DASD storage controller analyzed the data access pattern based on the frequency of access, identified the remapping algorithm to remap the logical group to the physical group, and was identified Move physical group according to remapping algorithm.

  Data reordering on DASD places frequently accessed data in close proximity, thereby minimizing seek time. The adaptive data localization method is performed periodically by the storage controller, but is transparent to the host file system. Reordering of data on DASD is performed during periods of low system data transfer activity. Refer to the summary column.

§5: Description of Conventional Example 5 Hereinafter, Patent Document 4 will be described as Conventional Example 5. Conventional Example 6 describes the following contents.

(a): “Provide a storage device that improves the response time.” Refer to the summary column. (b): “The file name, file size, and head of the file information area in the disk device format. The cluster number is read out, the access information for each file recorded in the access information area is referred to, the file update order map is generated by arranging in order of the number of access times of each file, and stored in the RAM. Change the order of file information in the file.In other words, files with a higher access count are placed higher in the update order map. "... See the summary column
JP 2001-337790 A JP 2003-216460 A JP 10-91362 A JP-A-10-105445

  (1): In the case of Conventional Example 1, it is considered to construct a new system or to change the configuration of the disk drive group in accordance with the system enhancement or change. In this case, it is necessary to redesign after examining the operation method and performance requirements for the existing system, and after changing the configuration based on the design after stopping the host system.

  However, in that case, there will be an impact on customer operations, costs due to the system being shut down, and human costs for configuration design and changes, and the validity of the configuration that was set if the system was not actually operated. Cannot be verified.

  (2): Conventional example 2 indicates that “data migration migration target logic for balancing system load from access information to physical drives and logical drives under the disk array controller without increasing the management load of the disk array controller. “Select volume”. However, as in the present invention, the conventional example 2 does not describe “preventing that data with a heavy load is concentrated on the local area before data recorded in the disk drive” as in the present invention. The content of the background art.

  (3): Conventional example 3 indicates that “the access fluctuation cycle is obtained by continuously accumulating accesses per unit time of the data block. With reference to this fluctuation cycle, the data block is appropriately set between the storage tiers in advance. Move to a storage with good performance. " However, as in the present invention, the conventional example 3 does not describe “preventing that data with a large load is concentrated on a local area before data recorded in a disk drive” as in the present invention. The content of the background art.

  (4): Conventional example 4 is "providing a method and apparatus for adaptively localizing data frequently accessed and randomly addressed in DASD to improve system access performance." is there. However, as in the present invention, the conventional example 4 does not describe “preventing the concentrated data from being concentrated on the data recorded in the disk drive in advance” as in the present invention. The content of the background art.

  (5): Conventional example 5 indicates that “a file update order map is generated by referring to the access information for each file recorded in the access information area, arranged in the order of the number of accesses of each file, and stored in the RAM. The file information arrangement order in the file information area is changed. That is, a file with a higher access count is arranged at the higher level of the update order map. However, as in the present invention, the conventional example 4 does not describe “preventing the concentrated data from being concentrated on the data recorded in the disk drive in advance” as in the present invention. The content of the background art.

  The present invention has been made to solve the conventional problems, and counts the number of accesses from the host, determines high load or low load based on the number of counts, and performs data movement based on the determination result, An object of the present invention is to prevent data with a heavy load from being concentrated locally with respect to data recorded on a disk drive.

  In order to achieve the above object, the present invention is configured as follows.

  (1): a host having a disk drive group having a plurality of disk drives and a controller which is connected to the host and operates and controls the disk drive group, and which controls the interface with the host in the controller. In a disk device comprising an interface control unit, a storage unit for storing information, and a disk control unit for controlling read / write of data with respect to the disk drive group, the disk control unit is connected from the host to the disk drive group. When the data read / write is accessed, the number of data read / write accesses is counted in stripe units, which is the data read / write unit, and the count data is stored in the storage unit. And access of data within a predetermined period of time. The number of scans is compared with a threshold set in advance for each stripe, and if the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area. If the number of accesses does not exceed the threshold, the load is low Load determination means for determining an area, and stripes determined as a high load area or a low load area by the load determination means with respect to a disk drive prepared in advance for data update processing after the determined time A copy destination securing means for securing a stripe size area having the same capacity as the disk drive to be moved as a copy destination so that data can be moved, and data to be written to the write processing for the stripe, the secured copy destination A copy method that writes directly to an area in units of blocks, which is the unit of writing in the stripe. When the host accesses the data written in the copy destination area, the disk drive simulates the state as if the data existed in the area where the data should have been originally written. A host access processing unit that performs processing to be recognized by the data processing unit, and a block that has not been updated in the write processing for the stripe and has not been updated, a predetermined period has elapsed since the determination by the load determination unit. After that, the data moving means for copying the block to the copy destination area as an internal process, and the space for releasing the area where the stripe originally existed as the free data area after the copy processing of all the blocks of the stripe is completed. And an area releasing means.

  (2): In the disk device of (1), for a stripe determined as a high load area by the load determination unit, a high performance prepared in advance for data update processing from the time when the stripe is determined as a high load area. In order to be able to move to the disk area, a high-load copy destination securing unit that secures a stripe size area with the same capacity as the moving stripe in the high-performance disk area to be moved, and write processing to the stripe And high-load copy means for directly writing the data to be written to the copy destination area in the secured high-performance disk area in units of blocks, which are write units in the stripe. To do.

  (3): In the disk device of (1) above, for the stripe determined as the low load area by the load determination means, the low performance prepared in advance for the data update processing after the point determined as the low load area In order to be able to move to the disk area, a low-load copy destination securing means that secures a stripe size area with the same capacity as the moving stripe in the low-performance disk area to be moved, and write processing to the stripe And low-load copy means for directly writing data to be written to the copy destination area in the reserved low-performance disk area in units of blocks, which are write units in the stripe. To do.

  (4): a host having a disk drive group having a plurality of disk drives and a controller that is connected to the host and operates and controls the disk drive group, and that controls the interface with the host in the controller. In a disk control method for a disk device comprising an interface control unit, a storage unit for storing information, and a disk control unit for controlling read / write of data with respect to the disk drive group, data is transferred from the host to the disk drive group. When the read / write access is performed, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the storage unit. Data access times within a predetermined period The number is compared with a threshold set in advance for each stripe, and if the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area. If the number of accesses does not exceed the threshold, the low load area For a stripe determined as a high load area or a low load area by the load determination means, and for a data update process after the determination, data is prepared for a disk drive prepared in advance. A copy destination securing procedure for securing a stripe size area having the same capacity as the disk drive to be moved as a copy destination so that the data can be migrated, and data to be written for the write processing for the stripe. In contrast, a copy procedure for writing directly in block units, which are write units in the stripe, When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing procedure for performing the processing to be performed, and after a predetermined period has elapsed since the determination of the load determination means for a block that has not been updated and data has not been written in the write processing for the stripe A data movement procedure for copying a block to the copy destination area as an internal process, and a free area release for releasing the area where the stripe originally existed as a free data area after completing the copy process of all the blocks of the stripe And a procedure.

(Function)
FIG. 1 is a diagram illustrating the principle of the present invention. The operation of the present invention will be described below with reference to FIG.

(a): Actions of (1) and (4) The disk device has the following actions under the control of the disk controller.

  The access count collecting means counts the number of data read / write accesses in a stripe unit, which is a data read / write unit, when a data read / write access is made from the host to the disk drive group. The count data is stored in the storage unit.

  The load determination means compares the number of data accesses within a predetermined period with a threshold set in advance for each stripe, and determines that the area is a high load area if the number of accesses within a predetermined period exceeds the threshold. If the access count does not exceed the threshold value, it is determined as a low load area.

  The copy destination securing unit can move data to a disk drive prepared in advance with respect to the data update process for the stripe determined as the high load region or the low load region by the load determination unit. As described above, an area having a stripe size having the same capacity as the disk drive to be moved is secured as a copy destination.

  The copy means directly writes data to be written to the reserved copy destination area in units of blocks, which are units of writing in the stripe, for the write processing for the stripe.

  When the host access processing means accesses the data written in the copy destination area, the host access processing means simulates the state as if the data existed in the area where the disk drive should have been originally written. To make the host recognize automatically.

  The data moving means performs internal processing after a predetermined period has elapsed after the determination by the load determining means for a block that has not been updated in the writing process for the stripe and has not been updated. A block is copied to the copy destination area.

  The free area releasing means releases the area where the stripe originally exists as a free data area after the copy processing of all the blocks of the stripe is completed.

  In this way, the number of accesses from the host is counted, a high load or a low load is determined based on the number of counts, and data is moved according to the determination result, so that the load on the data recorded in the disk drive is large. Data can be prevented from concentrating locally.

(b): Action (2) The above (2) has the following action in addition to the action (1).

  The high-load copy destination securing unit can move a stripe determined as a high-load area by the load determination unit to a previously prepared high-performance disk area for data update processing after the time when the stripe is determined as a high-load area. In addition, a stripe size area having the same capacity as the stripe to be moved is secured as a copy destination in the high performance disk area to be moved.

  The high load copy means directly writes the data to be written to the copy destination area in the secured high-performance disk area in units of blocks, which are the write units in the stripe, for the write process on the stripe.

  In this way, it is possible to prevent the data with a large load from being concentrated locally with respect to the data recorded in the disk drive.

(c): Action of (3) The above (3) has the following action in addition to the action of (1).

  The low-load copy destination securing unit can move a stripe determined as a low-load area by the load determination unit to a previously prepared low-performance disk area for data update processing after the time when it is determined as a low-load area. In addition, a stripe size area having the same capacity as the stripe to be moved is secured in the low performance disk area to be moved.

  The low-load copy means directly writes data to be written to the copy destination area in the reserved low-performance disk area in units of blocks, which are write units in the stripe, for the write process on the stripe.

  In this way, it is possible to prevent the data with a large load from being concentrated locally with respect to the data recorded in the disk drive.

  (1): Counts the number of accesses from the host, determines whether the load is high or low based on the number of counts, and moves the data according to the determination result, so the load on the data recorded on the disk drive is large There is an effect that data can be prevented from concentrating locally.

  (2): Conventionally, when a disk drive group configuration is changed due to the construction of a new system or the enhancement or change of the system, the redesign is performed after examining the operation method and performance requirements for the existing system. It is necessary to make changes to the configuration based on the design after stopping the host system. For this reason, there is an impact on customer operations, costs due to system shutdown, and human costs for configuration design and changes, and if the system is not actually operated, the validity of the set configuration can be verified. could not.

  However, according to the present invention, it is possible to prevent the data with a large load from being concentrated locally with respect to the data recorded in the disk drive, so that all the conventional problems as described above are eliminated. is there.

§1: Description of overview An overview of a disk device will be described below with reference to FIG.

  (1): The disk device includes a disk drive group 6 having a plurality of disk drives, a controller 2 that is connected to the host 1 and is operated, and controls the disk drive group 6. A host interface control unit (host I / F control unit) 3 that performs interface control, a storage unit 5 that stores information, and a disk control unit 4 that controls data read / write with respect to the disk drive group 6 ( Disk control program).

  When the host 1 accesses the disk drive group 6 from the host 1 for data read / write access, the disk controller 4 (disk control program) stores data in stripe units, which are data read / write units. The access count collecting means for counting the number of read / write accesses and storing the count data in the storage unit 3, and comparing the number of data accesses within a predetermined period with a preset threshold value for each stripe, If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area. If the number of accesses does not exceed the threshold, the load determination means determines that the load is low, and the load determination means Or, for the stripe determined to be a low-load area, the data update processing is prepared in advance after the determination. A copy destination securing unit that secures an area having a stripe size of the same capacity as the disk drive to be moved as a copy destination so that data can be moved to the disk drive, Copy means for directly writing to the reserved copy destination area in units of blocks, which are write units in the stripe, and when access from the host to the data written in the copy destination area occurs, the disk drive In addition, the host access processing means for performing a process for making the host recognize the state as if the data was originally in the area where the data should have been originally written, and the data is not written by the write process for the stripe. , Negative for blocks not updated After a predetermined period has elapsed since the determination by the determination means, the data movement means for copying the block to the copy destination area as an internal process, and after the copy processing for all the blocks of the stripe is completed And an empty area releasing means for releasing the area where the stripe originally exists as an empty data area.

  (2): In the disk device of (1), the disk control unit 4 (disk control program) further determines that the stripe determined as the high load area by the load determination means is the high load area. After that, with regard to data update processing, a high-speed disk area with the same capacity as the moving stripe is secured as a copy destination in the high-performance disk area to be moved so that it can be moved to a high-performance disk area prepared in advance. For the write processing for the load copy destination and the write processing for the stripe, the data to be written is directly written to the copy destination area in the reserved high-performance disk area in units of blocks which are the write units in the stripe. High-load copy means.

  (3): In the disk device of (1), the disk control unit 4 (disk control program) further determines when the stripe determined as the low load area by the load determination means is the low load area. Thereafter, with respect to the data update processing, a low-size disk area having the same capacity as the moving stripe is secured as a copy destination in the low-performance disk area to be moved so that it can be moved to a low-performance disk area prepared in advance. For the load copy destination securing means and the write processing for the stripe, the data to be written is directly written to the copy destination area in the secured low-performance disk area in units of blocks that are the write unit in the stripe. And a low-load copy means.

§2: Description of configuration of disk device FIG. 2 is a configuration diagram of the disk device. The configuration of the disk device will be described below with reference to FIG.

  The disk device (magnetic disk device) shown in FIG. 2 is composed of a controller 2 and a disk drive group 6 connected to the controller 2 and is operated by being connected to the host 1.

  The controller 2 includes a host interface control unit (hereinafter referred to as “host I / F control unit”) for performing interface control with a plurality of hosts 1-1, 1-2, and 1-3. 3, a disk control unit 4 that performs various disk controls including data write and read controls on the disk drive group 6, and a storage unit 5 that stores various types of information (or data).

  Further, the storage unit 5 is provided with a control information setting area, a definition information storage area, a table information storage area, and other information storage areas, and each stores necessary information (or data). It is like that.

  The disk drive group 6 includes a plurality of disk drives D # 0 to D # n. Each of the plurality of disk drives D # 0 to D # n includes a plurality of disk drives therein to constitute a RAID device (RAID0 to RAID5). Hereinafter, the disk drives constituting these RAID devices will be described simply as “disk drives D # 0 to D # n”.

§3: Detailed description of information (or data) in the storage unit 5 FIG. 3 shows host-drive correspondence information and performance-specific disk drive setting information, FIG. A shows host-drive correspondence information, and FIG. B shows performance-specific information. This is disk drive setting information. Hereinafter, the initial setting information (or data) in the storage unit 5 will be described in detail with reference to FIGS.

(1): Host-drive correspondence information FIG. 3A shows host-drive correspondence information. This information includes the hosts 1-1, 1-2, 1-3 shown in FIG. 2 and the disks used by the hosts 1-1, 1-2, 1-3 before the start of normal disk device operation. Regarding the relationship with the drive number, the connection relationship is set in the control information setting area in the storage unit 5 in advance by manual operation by an operator or the like when the disk device is initially installed (setting information in the storage unit 5). To remember).

  In this case, the host-drive correspondence information associates the host name with the set disk drive name. In this example, host name = host 1-1 corresponds to setting disk drive name = D # 0, D # 1, D # 3, and host name = host 1-2 corresponds to setting disk drive name = D. # 2 and D # 4 correspond to each other, and host name = host 1-3 corresponds to the set disk drive name = D # 5 and D # 6.

  If such information is stored in the storage unit 5, the correspondence between the host and the disk drive becomes clear by control during operation. In this case, the host-drive correspondence information is stored in the control information setting area in the storage unit 5 of FIG.

(2): Performance-specific disk drive setting information FIG. 3B shows performance-specific disk drive setting information.

  This information is information necessary for performing control for distinguishing between a high performance disk drive and a low performance disk drive. Accordingly, the setting information of the performance-specific disk drive is to divide the high-performance disk drive and the low-performance disk drive and store the corresponding disk drives.

  In this example, D # 0, D # 3, and D # 4 are registered in the high-performance disk drive, and D # 2 and D # 5 are registered in the low-performance disk drive. In this case, the setting information of the performance-specific disk drive is set in the definition information storage area in the storage unit 5. This information is information set in advance by an operator or the like.

(3): Threshold Setting Example FIG. 4 is an example of threshold setting. This information is information in which a disk drive name and a threshold value are paired, and is set for each disk drive. This threshold value is a threshold value for a predetermined period (for example, 30 days, half a year, etc.).

  For example, disk drive name = D # 0 threshold = SH1. Also, disk drive name = D # 1 threshold = SH2. Also, disk drive name = D # 2 threshold = SH3. Also, disk drive name = D # 3 threshold = SH4. In this case, the threshold setting information is set in the definition information storage area in the storage unit 5 of FIG. This information is information set in advance by an operator or the like.

§4: Explanation of processing during operation, etc. FIG. 5 is an explanatory diagram during disk drive operation (part 1), FIG. 6 is an example of an access count table, FIG. 7 is an explanatory diagram during disk drive operation (part 2), FIG. Is an example of an access count table, FIG. 9 is a block-unit data use state table, and FIG. 10 is a data movement state table. Hereinafter, the processing of the disk device will be described with reference to FIGS.

(1): Processing 1
In this example, the data recorded in the disk drive group 6 can be prevented from concentrating locally on the data having a large load. As shown in FIG. 5A, the disk drive group 6 is provided with a large number of disk drives, such as disk drives D # 0, D # 1, and D # 2.

  Of these disk drives, this information is predefined by dividing it into high performance disk drives and low performance disk drives. That is, it is stored in the control information setting area of the storage unit 5 as setting information of the performance-specific disk drive (see FIG. 3B).

  Then, as shown in FIG. 5A, performance information for data recorded in the disk drive group 6 is collected in units of stripe size. In this case, each time reading is performed on the stripe, information as the number of reads (number of reads) is stored in the access count table (see FIG. 6) in the table information storage area in the storage unit 5. to add. Further, whenever writing occurs to the stripe, information as the number of writes (number of writes) is additionally updated in the access count table (see FIG. 6) in the table information storage area in the storage unit 5. To do.

  That is, in FIG. 5B, the solid line indicates data read (read), and the dotted line indicates data write (write). When the host 1 accesses the disk drive group 6 to read / write data, the disk control unit 4 (disk control program) reads / writes data in stripe units, which are data read / write units. The access count is counted, and the count data is updated in the access count table (see FIG. 6) stored in the storage unit 5.

  In the example of the access count table in FIG. 6, there are stripe No (strip number), read count, write count, and total (read count + total write count, that is, read / write access count) items. Data is written to each of these items. In this case, the number of reads or the number of writes is written for each stripe No, and the total (the total number of reads + the number of writes) is written in the total item.

(2): Processing 2
As described above, the disk control unit 4 stores the total number of read / write accesses in the storage unit 5 while determining the number of data accesses within a predetermined period (30 days, six months, etc.). Compared with the threshold value preset for each stripe (see FIG. 4), if the number of accesses within a predetermined period (for example, 30 days) exceeds the threshold, it is determined as a high load area, and the number of accesses exceeds the threshold. If not, it is determined as a low load region.

  As described above, a high load area or a low load area is determined in units of stripes. For example, in the example shown in FIG. 8, if the threshold value of stripe No. = 4, the total value of stripe No = # 6 = 6 or the total value of stripe No = # 10 = 5 is the threshold value (see FIG. 8). 4), the load is determined to be high. Further, the total value of stripe No = # 1 = 0 and the total value = 0 of stripe No = # 7 are determined to have a low load because the total value is lower than the threshold value (see FIG. 4).

  In this case, the above description is an example, and it is natural that different determination results may be obtained if the threshold value is set differently.

(3): Processing 3
As described above, for the stripe determined as the “high load area”, the data update process is performed in the following sequence after the time when it is determined as the high load area.

(a): Same capacity as the stripe moving in the high-performance disk drive to be moved so that data can be moved with respect to the high-performance disk drive prepared in advance (D # 0 in FIG. 7A) An area of the stripe size is secured as a copy destination. (For example, stripes # 0 and # 2 in the high performance disk drive D # 0 shown in FIG. 7A are secured as copy destinations.)
(b): Regarding the write (update) processing for the stripe, the data to be written is directly written in a block unit (data write unit in the stripe) to a copy destination area prepared in advance. In this case, data determined to be highly loaded is moved (copied) to a high performance disk drive (for example, D # 0).

  In the example of FIG. 8, for example, the data of the stripe # 6 in the disk drive D # 1 is moved (copied) to the stripe # 2 in the copy destination disk drive D # 0. Further, the data of the stripe # 10 in the disk drive D # 2 is moved (copied) to the stripe # 0 in the copy destination disk drive D # 0.

  (c): When data is written from the host 1 to the data written in the copy destination area, it is as if the data exists in the area that should have been originally written under the control of the disk control unit 4. Processing is performed so that the host can artificially recognize such a state.

  (d): For a block that has not been updated, the copy destination is used as internal processing of the disk control unit 4 after a specified period (for example, 24 hours) has elapsed since it was determined as a high load area as described above. Copy the block to the area.

  (e): After the copy processing of all the blocks in the stripe is completed, the area where the stripe originally exists is released as an empty data area.

(4): Process 4
As described above, with respect to the stripe determined to be the “low load area”, the following sequence processing is performed regarding the data update processing after the time point determined to be the low load area.

(a): In a low-performance disk drive to be moved so that data can be moved to a low-performance disk drive prepared in advance (for example, the low-performance disk drive D # 2 shown in FIG. 7A) An area having a stripe size having the same capacity as the stripe to be moved is secured as a copy destination. (For example, stripes # 9 and # 11 in the low performance disk drive D # 2 shown in FIG. 7A are secured as copy destinations.)
(b): For writing (updating) processing on the stripe, data to be written is directly written in block units in a copy destination area prepared in advance.

  In the example of FIG. 8, for example, the data of the stripe # 1 in the disk drive D # 0 is moved (copied) to the stripe # 9 in the copy destination disk drive D # 2. Further, the data of the stripe # 7 in the disk drive D # 1 is moved (copied) to the stripe # 11 in the copy destination disk drive D # 2.

  (c): When data is written from the host 1 to the data written in the copy destination area, the disk controller 4 indicates that the data exists in the area that should have been originally written. Are processed so that the host 1 can recognize them.

  (d): As for the block that has not been updated, after the designated period such as 24 hours has elapsed since the determination as the low load area as described above, the disk control unit 4 performs block processing on the copy destination area. Make a copy of

  (e): After the copy processing of all the blocks in the stripe is completed, the area where the stripe originally exists is released as an empty data area.

(5): Process 6
With the above processing, the disk drive group 6 is arranged so that it can respond normally to the read / write (read / write) request from the host 1 to the relocated block or stripe. The management state is maintained until the logical volume in which the data unit is logically arranged is erased.

(6): Other explanation
(a): The high-performance disk area is an area defined in advance in the RAID device as a RAID configuration in which the maximum performance is prioritized over the cost by adopting RAID 0 + 1, a high-rotation disk, or the like.

  (b): The low-performance disk area is an area defined in advance in the RAID device as a RAID configuration in which cost (unit price per byte) is prioritized over performance by adopting RAID 5, a large capacity disk, or the like. That is.

(c): Block Unit Data Usage Status Table FIG. 9 is a block unit data usage status table. In the above processing, data is moved in units of stripes. In this case, data is moved in units of blocks (write units) in the stripes. The block unit data usage state table shown in FIG. 9 is used at this time and is stored in advance in the table information storage area of the storage unit 5.

  This block-unit data usage status table contains information about the disk drive name, stripe number, data block number, and usage status (for example, ○ indicates unused status, × indicates used status) information in sequence. Accordingly, the disk control unit 4 updates the data. Therefore, when data in units of blocks is used, the use state and the unused state are known, and processing such as data movement becomes easy.

(d): Example of data movement state table FIG. 10 is an example of a data movement state table. This example of data movement state table is used when performing data movement (data copying), and is stored in advance in the table information storage area of the storage unit 5.

  This data movement state table has items such as a movement source data block number, a movement flag, a movement destination data block number, and disk drive / stripe information, and data is registered respectively. In FIG. 10, DB # indicates a data block number, movement flag = 0 indicates no movement, and movement flag = 1 indicates movement.

  The information in the table is sequentially updated by the disk control unit 4 according to the usage state. Therefore, when data in units of blocks is used, the data movement state (data movement source, movement destination, whether moved, etc.) can be known, and processing such as data movement becomes easy.

§5: Explanation of processing of controller FIG. 11 is a flowchart of processing of the controller. Hereinafter, the processing of the controller will be described based on FIG. The processing shown below is processing performed by a disk control unit (disk control program) in the controller, and in FIG. 11, S1 to S9 indicate processing steps.

  First, the disk control unit (disk control program) 3 clears the access table (access count table in FIG. 6) in the storage unit 5 (initializes by clearing the stripe number, read count, write count, and total). (S1).

  Next, it is determined whether there is a read / write access from the host (S2). If there is no read / write access, the host waits until there is an access. In this way, if there is a read / write access from the host, the number of accesses is recorded in the access table (S3).

  Then, it is determined whether or not the set time limit (for example, 30 days) has passed (S4). If the set time limit has not passed, the process proceeds to S2, and if the set time limit has passed, the access count is set. Compare with the threshold value (S5). Next, it is determined whether or not the access count exceeds the threshold (see FIG. 4) (S6). If the access count does not exceed the threshold, the process proceeds to S2. If the access count exceeds the threshold, the data The movement is performed (S7).

  Next, after the data is moved, a table data flag (see FIG. 10) is set (S8). Next, the moved stripe data is cleared (S9), and this process ends.

§6: Description of recording medium and program The disk control unit 4 in the controller 2 shown in FIG. 2 is composed of a program (disk control program), and performs processing of the controller 2. In this case, the program (disk control program) is taken into the controller 2 from the outside and stored in the storage unit 5, for example, and a processor (not shown) reads and executes the program, It is also possible to perform processing.

  In this case, for example, a program created by another device via the communication line and transmitted to the recording medium via the communication line is received by the controller 2 via the communication unit (not shown). Then, the program is stored in the storage unit 5. And it is also possible to perform the process of the controller 2 by the processor executing the program.

(Appendix)
The following configuration is appended to the above description.

(Appendix 1)
A disk drive group having a plurality of disk drives, and a controller that is operated while connected to a host and controls the disk drive group,
In the disk device, the controller includes a host interface control unit that performs interface control with the host, a storage unit that stores information, and a disk control unit that performs read / write control of data with respect to the disk drive group. ,
The disk controller is
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Means for collecting the number of accesses stored in the department;
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, load determination means for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing means for securing a stripe size area having the same capacity as the disk drive as a copy destination;
A copy unit that directly writes data to be written to the reserved copy destination area in units of blocks, which is a unit of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing means for performing processing
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data moving means for copying blocks to the
Free space releasing means for releasing the area where the stripe originally existed as a free data area after the copy processing of all the blocks of the stripe is completed;
A disk device comprising:

(Appendix 2)
For the stripe determined as the high load area by the load determination means, the data to be moved can be moved to the high-performance disk area prepared in advance for the data update process after the determination as the high load area. A high-load copy destination securing unit that secures a stripe size area of the same capacity as the moving stripe in the performance disk area,
A high-load copy unit that directly writes data to be written to a copy destination area in the secured high-performance disk area in units of blocks, which is a write unit in the stripe, with respect to a write process on the stripe;
The disc device according to appendix 1, characterized by comprising:

(Appendix 3)
For the stripe determined as the low load area by the load determination unit, the data update process is performed so that the stripe to be moved can be moved to a low performance disk area prepared in advance after the time when it is determined as the low load area. A low-load copy destination securing unit that secures a stripe size area having the same capacity as the moving stripe in the performance disk area,
A low-load copy unit that directly writes data to be written to a copy destination area in the reserved low-performance disk area in units of blocks that are write units in the stripe, with respect to a write process on the stripe;
The disc device according to appendix 1, characterized by comprising:

(Appendix 4)
A disk drive group having a plurality of disk drives; a controller that is connected to a host and is operated, and that controls the disk drive group; and a host interface control unit that performs interface control with the host in the controller; In a disk control method for a disk device, comprising a storage unit for storing information and a disk control unit for controlling read / write of data with respect to the disk drive group,
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Access frequency collection procedure to be stored in the
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, a load determination procedure for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing procedure to secure a stripe size area of the same capacity as the disk drive as the copy destination,
A copy procedure for directly writing data to be written to the reserved copy destination area in units of blocks, which are units of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing procedure for performing processing,
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data movement procedure to copy the block to
After completing the copy processing of all the blocks of the stripe, a free area releasing procedure for releasing the area where the stripe originally existed as a free data area,
A disk control method comprising:

(Appendix 5)
For the stripe determined as the high load area by the load determination means, the data to be moved can be moved to the high-performance disk area prepared in advance for the data update process after the determination as the high load area. A high-load copy destination securing procedure that secures a stripe size area with the same capacity as the moving stripe in the performance disk area,
A high-load copy procedure for directly writing data to be written to a copy destination area in the reserved high-performance disk area in units of blocks, which is a write unit in the stripe, with respect to a write process on the stripe;
The disk control method according to appendix 4, characterized by comprising:

(Appendix 6)
For the stripe determined as the low load area by the load determination unit, the data update process is performed so that the stripe to be moved can be moved to a low performance disk area prepared in advance after the time when it is determined as the low load area. A low-load copy destination securing procedure that secures a stripe size area with the same capacity as the moving stripe in the performance disk area,
A low-load copy procedure for directly writing data to be written to the copy destination area in the reserved low-performance disk area in units of blocks, which is a write unit in the stripe, with respect to the write process for the stripe;
The disk control method according to appendix 4, characterized by comprising:

(Appendix 7)
A disk drive group having a plurality of disk drives, and a controller that is operated while connected to a host and controls the disk drive group,
A disk device having a host interface control unit that performs interface control with a host, a storage unit that stores information, and a disk control unit that performs read / write control of data with respect to the disk drive group. ,
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Means for collecting the number of accesses stored in the department;
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, load determination means for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing means for securing a stripe size area having the same capacity as the disk drive as a copy destination;
A copy unit that directly writes data to be written to the reserved copy destination area in units of blocks, which is a unit of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing means for performing processing
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data moving means for copying blocks to the
A program for realizing the function of an empty area releasing means for releasing the area where the stripe originally existed as an empty data area after the copy processing of all the blocks of the stripe is completed, or a computer-readable recording medium on which the program is recorded .

(Appendix 8)
For the stripe determined as the high load area by the load determination means, the data to be moved can be moved to the high-performance disk area prepared in advance for the data update process after the determination as the high load area. A high-load copy destination securing unit that secures a stripe size area of the same capacity as the moving stripe in the performance disk area,
A function of high-load copy means for directly writing data to be written to a copy destination area in the reserved high-performance disk area in units of blocks, which is a write unit in the stripe, for write processing on the stripe The program according to appendix 7 for realizing, or a computer-readable recording medium on which the program is recorded.

(Appendix 9)
For the stripe determined as the low load area by the load determination unit, the data update process is performed so that the stripe to be moved can be moved to a low performance disk area prepared in advance after the time when it is determined as the low load area. A low-load copy destination securing unit that secures a stripe size area having the same capacity as the moving stripe in the performance disk area,
A function of a low-load copy unit that directly writes data to be written to a copy destination area in the reserved low-performance disk area in units of blocks, which is a write unit in the stripe, for a write process on the stripe. The program according to appendix 7 for realizing, or a computer-readable recording medium on which the program is recorded.

It is a principle explanatory view of the present invention. 1 is a configuration diagram of a disk device in an embodiment. FIG. FIG. 4A shows host-drive correspondence information and performance-specific disk drive setting information in the embodiment, FIG. 3A shows host-drive correspondence information, and FIG. B shows performance-specific disk drive setting information. It is an example of a threshold setting in an embodiment. It is explanatory drawing (the 1) at the time of disk drive operation | movement in embodiment, A figure is Example 1 and B figure is Example 2. FIG. It is an example of the access frequency table in an embodiment. FIG. 4 is an explanatory diagram (part 2) when the disk drive is operated in the embodiment, and FIG. It is an example of the access frequency table in an embodiment. It is a use status table of block unit data in an embodiment. It is an example of a data movement state table in an embodiment. It is a process flowchart of the controller in an embodiment. It is explanatory drawing of the prior art example 1. FIG.

Explanation of symbols

1, 1-1, 1-2, 1-3 Host 2 Controller 3 Host interface control unit (host I / F control unit)
4 Disk control unit 5 Storage unit 6 Disk drive group

Claims (5)

A disk drive group having a plurality of disk drives, and a controller that is operated while connected to a host and controls the disk drive group,
In the disk device, the controller includes a host interface control unit that performs interface control with the host, a storage unit that stores information, and a disk control unit that performs read / write control of data with respect to the disk drive group. ,
The disk controller is
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Means for collecting the number of accesses stored in the department;
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, load determination means for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing means for securing a stripe size area having the same capacity as the disk drive as a copy destination;
A copy unit that directly writes data to be written to the reserved copy destination area in units of blocks, which are units of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing means for performing processing
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data moving means for copying blocks to the
Free space releasing means for releasing the area where the stripe originally existed as a free data area after the copy processing of all the blocks of the stripe is completed;
A disk device comprising: For the stripe determined as the high load area by the load determination means, the data to be moved can be moved to the high-performance disk area prepared in advance for the data update process after the determination as the high load area. A high-load copy destination securing unit that secures a stripe size area of the same capacity as the moving stripe in the performance disk area,
A high-load copy unit that directly writes data to be written to a copy destination area in the secured high-performance disk area in units of blocks, which is a write unit in the stripe, with respect to a write process on the stripe;
The disk apparatus according to claim 1, comprising: For the stripe determined as the low load area by the load determination unit, the data update process is performed so that the stripe to be moved can be moved to a low performance disk area prepared in advance after the time when it is determined as the low load area. A low-load copy destination securing unit that secures a stripe size area having the same capacity as the moving stripe in the performance disk area,
A low-load copy unit that directly writes data to be written to a copy destination area in the reserved low-performance disk area in units of blocks that are write units in the stripe, with respect to a write process on the stripe;
The disk apparatus according to claim 1, comprising: A disk drive group having a plurality of disk drives; a controller that is connected to a host and is operated, and that controls the disk drive group; and a host interface control unit that performs interface control with the host in the controller; In a disk control method for a disk device, comprising a storage unit for storing information and a disk control unit for controlling read / write of data with respect to the disk drive group,
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Access frequency collection procedure to be stored in the
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, a load determination procedure for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing procedure to secure a stripe size area of the same capacity as the disk drive as the copy destination,
A copy procedure for directly writing data to be written to the reserved copy destination area in units of blocks, which are units of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing procedure for performing processing,
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data movement procedure to copy the block to
After completing the copy processing of all the blocks of the stripe, a free area releasing procedure for releasing the area where the stripe originally existed as a free data area,
A disk control method comprising: A disk drive group having a plurality of disk drives, and a controller that is operated while connected to a host and controls the disk drive group,
A disk device having a host interface control unit that performs interface control with a host, a storage unit that stores information, and a disk control unit that performs read / write control of data with respect to the disk drive group. ,
When data read / write access is made from the host to the disk drive group, the number of data read / write accesses is counted in stripe units, which are data read / write units, and the count data is stored in the memory Means for collecting the number of accesses stored in the department;
The number of data accesses within a predetermined period is compared with a threshold set in advance for each stripe. If the number of accesses within a predetermined period exceeds the threshold, it is determined as a high load area, and the number of accesses is If the threshold is not exceeded, load determination means for determining a low load region;
The stripe determined as the high load area or the low load area by the load determination means is set as a movement target so that data can be moved with respect to a disk drive prepared in advance for the data update process after the determined time. A copy destination securing means for securing a stripe size area having the same capacity as the disk drive as a copy destination;
A copy unit that directly writes data to be written to the reserved copy destination area in units of blocks, which is a unit of writing in the stripe, with respect to the write processing for the stripe;
When access to the data written in the copy destination area occurs from the host, the host recognizes the state as if the data exists in the area where the data was originally written in the disk drive. Host access processing means for performing processing
For a block in which data is not written in the write processing for the stripe and has not been updated, after a predetermined period has elapsed since the determination by the load determination unit, the internal processing is performed in the copy destination area. Data moving means for copying blocks to the
A program for realizing the function of a free area releasing means for releasing the area where the stripe originally existed as a free data area after the copy processing of all the blocks of the stripe is completed.

Images