JP2013206009A - Drive control device, storage system, and drive control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively use disks which stop revolving, while preventing occurrence of I/O delay and reduction of energy saving effect.SOLUTION: When revolution of disks of HDD 300 are stopped, a disk array device 100 writes data in a memory of SSD 400 each time a host computer 500 requires writing data. If a situation of writing data in the memory of SSD 400 satisfies a predetermined condition, the disk array device 100 starts revolution of the disks of HDD 300, and transfers the data from the memory of SSD 400 to the disks of HDD 300.

Description

  The present invention relates to a drive control device, a storage system, and a drive control method. The present invention particularly relates to a disk array device with an energy saving mode.

  2. Description of the Related Art A disk array apparatus having an energy saving function that stops the rotation of a disk of one of the duplex disk arrays is known (see, for example, Patent Documents 1 and 2).

JP 2010-61291 A JP 2009-80603 A

  In the conventional disk array device, the disk array in which the rotation of the disk is stopped cannot be used effectively, or even if it can be used, the disk rotation starts every time there is an access request. Input / output) delays occurred, and energy saving effects were reduced.

  An object of the present invention is to effectively use a disk that stops rotating while avoiding, for example, the occurrence of an I / O delay and a reduction in energy saving effect.

A drive control device according to one aspect of the present invention includes:
A drive control device that controls a disk drive that stores data in a disk and a memory drive that stores data in a memory,
A host interface for receiving input from the host computer;
When the disk drive has stopped rotating, the memory drive writes the data to the memory each time a request to write data to the disk is input from the host computer to the host interface unit. A memory access control unit for controlling
A memory write status determination unit that determines whether a data write status to the memory under the control of the memory access control unit satisfies a predetermined condition in a state where the disk drive stops rotating the disk When,
A disk rotation control unit for controlling the disk drive to start the rotation of the disk, when the memory writing state determination unit determines that the writing state satisfies the condition;
When the disk drive starts rotating the disk under the control of the disk rotation control unit, the memory drive is controlled to transfer the data written in the memory and the data transferred from the memory drive is transferred to the memory drive. A disk access control unit for controlling the disk drive to write to the disk.

In one example,
The drive control device further includes:
The memory is different from the memory, and includes a cache for storing the data every time a request to write data to the disk is input from the host computer to the host interface unit,
When the disk drive stops the rotation of the disk, the memory access control unit stores the data every time a request to write data to the disk is input from the host computer to the host interface unit. Controlling the memory drive to transfer from the cache to the memory drive and write the data to the memory;
When the disk drive has started rotating the disk, the disk access control unit receives the data every time a request to write data to the disk is input from the host computer to the host interface unit. The disk drive is controlled to transfer data from the cache to the disk drive and write the data to the disk.

In one example,
The memory write status determination unit sets the amount of data written to the memory under the control of the memory access control unit and not yet written into the disk under the control of the disk access control unit to a predetermined amount. If it has reached, it is determined that the write status satisfies the condition.

In one example,
The memory write status determination unit determines that the write status satisfies the condition when the frequency at which data is written to the memory reaches a predetermined frequency under the control of the memory access control unit.

In one example,
The drive control device further includes:
An address mapping unit for storing a correspondence relationship between a plurality of logical units obtained by dividing the storage area of the disk and addresses of the memory individually assigned to the respective logical units;
When the disk drive has stopped rotating the disk, the memory access control unit receives a write request for designating any one of the plurality of logical units as a data write request to the disk. Each time the host computer inputs the host interface unit, the memory drive corresponding to the logical unit is read from the address mapping unit and the memory drive is controlled to write the data to the address.

In one example,
The memory write status determination unit is configured to determine whether the data write status to the memory under the control of the memory access control unit is in advance for each logical unit in a state where the disk drive has stopped rotating the disk. Judgment is made as to whether or not the conditions specified in the above are satisfied.

In one example,
The drive control device further includes:
When the disk drive stops rotating the disk, when a data read request from the disk is input from the host computer to the host interface unit, it is determined whether the data is written in the memory. A memory writing presence / absence determining unit for determining,
The disk rotation control unit determines the disk drive to start rotation of the disk when the memory write presence / absence determination unit determines that the data requested to be read from the host computer is not written to the memory. Control
The disk access control unit controls the disk drive to read data requested to be read from the host computer from the disk when the disk drive starts rotating the disk under the control of the disk rotation control unit. ,
The host interface unit outputs data read from the disk to the host computer under the control of the disk access control unit;
The memory access control unit controls the memory drive to write data read from the disk into the memory under the control of the disk access control unit.

A storage system according to an aspect of the present invention includes:
The drive control device;
The disk drive;
And the memory drive.

A drive control method according to one aspect of the present invention includes:
A drive control method for controlling a disk drive for storing data in a disk and a memory drive for storing data in a memory,
When a drive control device including a host interface unit that receives an input from a host computer stops the rotation of the disk, a data write request to the disk is input from the host computer to the host interface unit. Each time the memory drive is controlled to write the data to the memory,
The drive control device determines whether or not the state of writing data to the memory satisfies a predetermined condition in a state where the disk drive stops rotating the disk,
If the drive control device determines that the write status satisfies the condition, the drive control device controls the disk drive to start rotation of the disk;
When the disk drive starts to rotate the disk, the drive control device controls the memory drive to transfer data written in the memory and transfers data transferred from the memory drive to the disk. The disk drive is controlled to write.

  According to one aspect of the present invention, when the rotation of the disk is stopped, the data is written to the memory every time there is a write request, and the disk rotation is performed when the condition of data writing to the memory satisfies a certain condition. Since the data is transferred from the memory to the disk by starting the process, it is possible to effectively use the disk that stops rotating while avoiding the occurrence of I / O delay and the reduction of the energy saving effect.

1 is a block diagram showing a configuration of a disk array device according to a first embodiment. 1 is a block diagram showing a configuration of a disk array device according to a comparative example of Embodiment 1. FIG. FIG. 3 is a diagram showing a data flow when the disk array device according to the first embodiment processes a write request. FIG. 3 is a diagram showing a data flow when the disk array device according to the first embodiment processes a write request. FIG. 3 is a block diagram showing a configuration of a disk array device according to a second embodiment. FIG. 10 is a diagram showing a data flow when the disk array device according to the second embodiment processes a read request. FIG. 10 is a diagram showing a data flow when the disk array device according to the second embodiment processes a read request. FIG. 10 is a diagram showing a data flow when the disk array device according to the second embodiment processes a read request. FIG. 10 is a diagram showing a data flow when the disk array device according to the second embodiment processes a read request.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
In this embodiment, a plurality of HDDs (Hard, Disks, Drives) are mounted, and a RAID (Redundant, Arrays, of, Independent, Disks) function with a plurality of HDDs is provided, and energy saving is realized by stopping the mounted HDDs. In a disk array device having a power saving function, an SSD (Solid / State / Drive) is provided between a cache memory in the disk array controller and a disk control unit. When the HDD installed in the disk array device has stopped the spindle motor from being rotated (spin down) by the power save function, and there is an access request from the host computer, the HDD is in the Ready (spin up) state. By absorbing the access delay until the time is reached by Read / Write (read / write) to the SSD, an energy saving mode for automatically controlling the spin-up / down of the HDD is realized while suppressing a decrease in access performance. In addition, this energy saving mode makes it possible to apply the power saving function that was mainly available only for backup storage to online storage that is normally used by the host computer, thereby improving the energy saving effect.

  FIG. 1 is a block diagram showing a configuration of a disk array device 100 according to the present embodiment. FIG. 2 is a block diagram showing a configuration of a disk array device 200 according to a comparative example of the present embodiment.

  First, the configuration of the disk array device 200 according to the comparative example shown in FIG. 2 will be described.

  In FIG. 2, the disk array device 200 includes n (n is a natural number) RAID groups 1 to n and two controllers 201.

  The RAID groups 1 to n are disk arrays, each of which is composed of a plurality of HDDs 300.

  The controller 201 includes a host interface unit 202 (host I / F unit), a network interface unit 203 (network I / F unit), a disk control unit 204, a disk array control unit 205, and a cache memory 206, respectively.

  The host interface unit 202 is connected to the host computer 500. The network interface unit 203 is connected to a management LAN (Local / Area / Network). The disk control unit 204 is connected to the HDDs 300 of the RAID groups 1 to n. When the disk array control unit 205 receives a request from the host computer 500 via the host interface unit 202, the disk array control unit 205 writes data to the HDDs 300 of the RAID groups 1 to n via the disk control unit 204 or The data is read from the n HDDs 300. In addition, when receiving an instruction from the LAN input via the network interface unit 203, the disk array control unit 205 performs setting, status notification, and the like of each unit of the disk array device 200. The cache memory 206 is a volatile memory that temporarily stores data written to the HDDs 300 of the RAID groups 1 to n and data read from the HDDs 300 of the RAID groups 1 to n.

  In the first method for realizing the power saving function of the disk array device 200, the disk array control unit 205 is connected to the start / stop of SCSI (Small / Computer / System / Interface) via the host interface unit 202 from the host computer 500. • Receive a Unit command. Alternatively, the disk array control unit 205 receives an instruction from the LAN via the network interface unit 203. Then, the disk array control unit 205 turns on / off the spindle motor of the HDD 300 in accordance with a SCSI command or instruction.

  In the second method for realizing the power saving function of the disk array device 200, the disk array control unit 205 automatically performs the operation of the HDD 300 when the access from the host computer 500 has not passed a preset time. Stop the spindle motor rotation and enter standby mode.

  In the first method, when access to the HDD 300 is resumed, the disk array control unit 205 receives the SCSI START / STOP / Unit command again and turns on the spindle motor of the HDD 300. In the second method, the disk array control unit 205 automatically starts rotation of the spindle motor of the HDD 300 upon receiving a write (read) or read (read) command during the standby mode, and returns to the idle mode. .

  In either method, the host computer 500 cannot perform actual disk access until the spindle motor reaches the specified rotation speed and enters the Ready state. Therefore, when access to the HDD 300 is resumed, a host access I / O delay occurs. Therefore, these methods can be used only when it is known in advance that there is no disk access for a long time, or when a disk access time zone can be scheduled, such as for backup purposes. Further, in these methods, if there is a write or read request when the spindle motor is off, the spindle motor must be rotated every time, so the energy saving effect is limited.

  Next, the configuration of the disk array device 100 according to the present embodiment shown in FIG. 1 will be described.

  In FIG. 1, a disk array device 100 is an example of a storage system, and includes n (n is a natural number of 1 or more) RAID groups 1 to n and two controllers 101.

  The RAID groups 1 to n are disk arrays, each of which is composed of a plurality of HDDs 300, as in the example shown in FIG. Each RAID level (for example, RAID1, RAID5, RAID6) of the RAID groups 1 to n is arbitrary. RAID groups with different RAID levels may be mixed. Further, instead of the RAID groups 1 to n, a disk array by a technique other than RAID may be used, or one or a plurality of single disk drives that are not disk arrays may be used.

  In each RAID group, a plurality of arbitrary disk capacities can be set as LUs (Logical Units) (logical units). For example, if the RAID group 1 is a disk array composed of five HDDs 300, a specific address range of each of the five HDDs 300 is LU1, another address range of each of the five HDDs 300 is LU2, and so on. A plurality of LUs can be set in the form. Each LU is recognized by the host computer 500 as a separate physical disk.

  The HDD 300 is a disk drive that stores data on a disk (specifically, a magnetic disk). Instead of the HDD 300, another type of disk drive (for example, a disk drive that stores data on an optical disk) may be used.

  The controller 101 is a drive control device that controls the HDDs 300 of the RAID groups 1 to n and the built-in SSD 400. The SSD 400 may be installed outside the controller 101.

  The SSD 400 is a memory drive that stores data in a memory (specifically, a flash memory). Note that other types of memory drives may be used instead of the SSD 400. Also, a plurality of SSD 400 or other types of memory drives may be used.

  In this embodiment, two controllers 101 having the same function are mounted on the disk array device 100 to have a redundant configuration. However, only one controller 101 may be mounted, or three or more controllers 101 are mounted. You may make it do.

  The controller 101 includes a host interface unit 102, a network interface unit 103, a disk control unit 104, a disk array control unit 105, and a cache memory 106 (cache).

  The host interface unit 102 is connected to the host computer 500. The host interface unit 102 receives input from the host computer 500 and outputs to the host computer 500. For this purpose, the host interface unit 102 performs a SCSI channel connection with the host computer 500, for example.

  The network interface unit 103 is connected to a management LAN. The network interface unit 103 receives a signal (a signal for performing setting or status notification of each unit of the disk array device 100) from a LAN (actually, a management terminal connected to the LAN), Send signals to the LAN.

  The disk control unit 104 is connected to the HDDs 300 of the RAID groups 1 to n. The disk control unit 104 controls disk I / O for the HDD 300. For this purpose, for example, the disk control unit 104 connects the HDD 300 to a fiber channel, a SCSI channel, a SAS (Serial Attached SCSI) channel, or the like.

  The disk array control unit 105 accepts an input of energy saving mode designation in RAID group units (or LU units) via the host interface unit 102 or the network interface unit 103. If the access from the host computer 500 does not elapse for a preset time for the RAID group designated for the energy saving mode or the LU belonging to the RAID group, the disk array control unit 105 automatically The rotation of the spindle motor of the HDD 300 of the RAID group is stopped and the standby mode is entered. Alternatively, the disk array control unit 105 receives a SCSI START / STOP / Unit command for instructing to stop the operation of a specific RAID group from the host computer 500 via the host interface unit 102 or a network from the LAN. When the same instruction is received via the interface unit 103, the rotation of the spindle motor of the HDD 300 of the RAID group is stopped and the standby mode is entered.

  In the present embodiment, access from the host computer 500 to the RAID group or an LU belonging to the RAID group until the HDD 300 of the RAID group in the standby mode is spun up and enters the idle mode (Ready state). It is received by the SSD 400 provided between the cache memory 106 and the disk control unit 104 that controls the HDDs 300 of the RAID groups 1 to n. By adopting such a system, it is possible to minimize the delay of the disk I / O.

  The disk array control unit 105 includes, as components for implementing the above-described method, a disk access control unit 107, a memory access control unit 108, an address mapping unit 109, a memory write status determination unit 110, and a disk rotation control unit 111. Is provided. The operation of each unit of the disk array control unit 105 will be described later.

  The cache memory 106 is a volatile memory that temporarily stores data written to the HDDs 300 of the RAID groups 1 to n and data read from the HDDs 300 of the RAID groups 1 to n. Note that a non-volatile memory may be used as the cache memory 106.

  Although not shown, in the present embodiment, each of the controllers 101 is a computer and includes various hardware devices. Hardware devices are connected by cables and signal lines.

  The controller 101 includes, as a processing device, for example, a CPU (Central Processing Unit) that executes a program. The CPU is connected to various interface units and storage devices via a bus and controls them. The controller 101 includes a host interface unit 102, a network interface unit 103, a disk control unit 104, and the like as interface units. In addition to the cache memory 106 and the SSD 400, the controller 101 includes a ROM (Read / Only / Memory), a RAM (Random / Access / Memory), and the like as storage devices. A program group and a file group are stored in the storage device. The programs in the program group are executed by the CPU. The program group includes a program for executing the function of each unit constituting the disk array control unit 105. The program is read and executed by the CPU. The file group includes data, information, signal values, and the like that are described as “˜data”, “˜information”, “˜ID (identifier)”, “˜flag”, “˜result” in the description of the present embodiment. Variable values and parameters are included as items of “˜file”, “˜database”, and “˜table”. The “˜file”, “˜database”, and “˜table” are stored in the storage device. Data, information, signal values, variable values, and parameters stored in the storage device are read into the memory by the CPU via a read / write circuit, and extracted, searched, referenced, compared, operated, calculated, controlled, output, printed, Used for CPU processing (operation) such as display. During CPU processing such as extraction, search, reference, comparison, calculation, calculation, control, output, printing, and display, data, information, signal values, variable values, and parameters are temporarily stored in the memory.

  The arrows in the drawings used in the description of this embodiment mainly indicate input / output of data and signals. Data and signals are recorded in a storage device. Data and signals are transmitted by a bus, a signal line, a cable, or other transmission media.

  In the description of the present embodiment, what is described as “to part” may be “to circuit”, “to device”, “to device”, and “to step”, “to process”, “to”. ~ Procedure "," ~ process ". That is, what is described as “˜unit” may be realized by firmware stored in the ROM. Alternatively, what is described as “˜unit” may be realized only by software, or only by hardware such as an element, a device, a board, and wiring. Alternatively, what is described as “to part” may be realized by a combination of software and hardware, or a combination of software, hardware and firmware. Firmware and software are stored in the storage device as programs. The program is read by the CPU and executed by the CPU. That is, the program causes the computer to function as “to part” described in the description of the present embodiment. Or a program makes a computer perform the procedure and method of "-part" described by description of this Embodiment.

  In the following, the operation of each part of the disk array device 100 according to the present embodiment shown in FIG. 1 (drive control method according to the present embodiment, program processing procedure according to the present embodiment) will be described.

  3 and 4 are diagrams showing a data flow when the disk array device 100 processes a write request.

  FIG. 3 shows the data flow when there is a write request specifying the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group to which the energy saving mode is specified is in a spin-up state (in operation). Show.

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a write request (Write request) from the host computer 500. The write request is, for example, a SCSI Write command, and includes a parameter specifying a RAID group or LU identifier (for example, a number).
(2) When a write request is input from the host computer 500 to the host interface unit 102, the disk array control unit 105 sends the write requested data, that is, write data (write data) via the host interface unit 102. From the host computer 500. The disk array control unit 105 stores the received write data in the cache memory 106.
(3) The disk array control unit 105 transmits a write completion report (I / O completion report) to the host computer 500 via the host interface unit 102.
(4) The disk access control unit 107 of the disk array control unit 105 transfers the write data stored in the cache memory 106 to the HDD 300 of the RAID group or LU specified by the write request via the disk control unit 104. The HDD 300 writes the write data transferred by the disk access control unit 107 to the disk. This operation is a so-called write back method, and data written to the cache memory 106 is written to the HDD 300 by destage processing asynchronously with the I / O of the host computer 500.

  As described above, in this embodiment, the cache memory 106 stores the data every time a data write request to the disk of the HDD 300 is input from the host computer 500 to the host interface unit 102.

  In the present embodiment, the disk access control unit 107 receives a data write request to the disk of the HDD 300 from the host computer 500 to the host interface unit 102 when the HDD 300 starts rotating the disk. Then, the HDD 300 is controlled to transfer the data from the cache memory 106 to the HDD 300 and write the data to the disk.

  FIG. 4 shows the data flow when there is a write request specifying the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group to which the energy saving mode is specified is in the spin-down state (operation stop). Show.

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a write request (Write request) from the host computer 500. As described above, the write request is, for example, a SCSI Write command, and includes a parameter specifying a RAID group and an LU identifier (for example, a number).
(2) When a write request is input from the host computer 500 to the host interface unit 102, the disk array control unit 105 receives write data (Write data) from the host computer 500 via the host interface unit 102. The disk array control unit 105 stores the received write data in the cache memory 106.
(3) The disk array control unit 105 transmits a write completion report to the host computer 500 via the host interface unit 102. The operation so far is the same as in the case of FIG.
(4) The memory access control unit 108 of the disk array control unit 105 transfers the write data to the SSD 400 at a predetermined timing (for example, almost simultaneously with the write data being stored in the cache memory 106). At this time, the memory access control unit 108 reads the address of the memory of the SSD 400 corresponding to the RAID group or LU designated by the write request from the address mapping unit 109 and designates the address to the SSD 400. Here, it is assumed that the address mapping unit 109 stores a correspondence relationship between each LU of the RAID groups 1 to n or RAID groups 1 to n and the memory address of the SSD 400 in advance. The SSD 400 writes the write data transferred by the memory access control unit 108 to the memory address designated by the memory access control unit 108. As described above, the data written to the cache memory 106 is written to the SSD 400 by destage processing asynchronously with the I / O of the host computer 500.
(5) The memory write status determination unit 110 of the disk array control unit 105 determines whether the data write status to the memory of the SSD 400 satisfies a predetermined condition. For example, the memory write status determination unit 110 writes data in the memory of the SSD 400 when the amount of data that has been written to the disk of the HDD 300 and has not yet been written to the disk of the HDD 300 reaches a predetermined amount, or When the frequency of writing reaches a predetermined frequency, it is determined that the state of data writing to the memory of the SSD 400 satisfies the above condition. The disk rotation control unit 111 of the disk array control unit 105 rotates the spindle motor of the disk with respect to the HDD 300 when the memory write state determination unit 110 determines that the data write state to the memory of the SSD 400 satisfies the above condition. Request to resume. When the HDD 300 is in the spin-up state (resumed operation), the disk access control unit 107 of the disk array control unit 105 handles write data that has not yet been written to the HDD 300 from the memory of the SSD 400 via the disk control unit 104. The data is transferred to the HDD 300 of the RAID group or LU. The HDD 300 writes the write data transferred by the disk access control unit 107 to the disk. As described above, the data written to the SSD 400 is finally written to the HDD 300 by destage processing after the HDD 300 is spun up.

  As described above, in the present embodiment, the address mapping unit 109 includes a plurality of LUs obtained by dividing the storage area of the disk of the HDD 300 and the addresses (storage areas) of the memory of the SSD 400 that are individually assigned to the respective LUs. Memorize the correspondence. For example, it is assumed that 20 GB (gigabyte) LU1 and 40 GB LU2 are set in RAID group 1, and 100 GB LU3 is set in RAID group 2. In this case, 10 GB (upper limit is 20 GB) from the start address of the SSD 400 memory is LU1, the next address is 5 GB (upper limit is 40 GB) is LU2, and the address of the SSD 400 memory is allocated to each LU. Can do. For example, the address range of the SSD 400 memory allocated to LU1 is an area for storing write data of a write request in which LU1 is designated. When the capacity allocated to each LU is smaller than the capacity of the LU (capacity of the HDD 300), data that is not referred to by the LRU (Least, Recently Used) or LFU (Least, Frequently Used) system, etc. (low usage frequency) Data) is the replacement target.

  The correspondence relationship between the LU and the memory address of the SSD 400 may be statically set (that is, fixed) in advance, may be dynamically set, may not be set at all, You may set the correspondence of a RAID group and the address of the memory of SSD400, without distinguishing. If the correspondence relationship is dynamically set, for example, the address mapping unit 109 monitors the memory usage state (usage amount, usage frequency, etc. of each LU) of the SSD 400, and assigns each LU according to the usage state. The amount is adjusted (for example, a larger number of addresses are allocated to an LU having a large memory usage or an LU having a high memory usage frequency). If no correspondence is set, the address mapping unit 109 is unnecessary.

  Further, in the present embodiment, the memory access control unit 108, when the HDD 300 stops the rotation of the disk, every time a data write request to the disk of the HDD 300 is input from the host computer 500 to the host interface unit 102. In addition, the SSD 400 is controlled to transfer the data from the cache memory 106 to the SSD 400 and write the data to the memory. For example, when the HDD 300 has stopped rotating the disk, the memory access control unit 108 receives a write request for designating one of a plurality of LUs as a data write request to the disk of the HDD 300. Is input to the host interface unit 102, the address of the memory of the SSD 400 corresponding to the LU is read from the address mapping unit 109. Then, the memory access control unit 108 controls the SSD 400 to write the data at the address.

  In the present embodiment, the memory write status determination unit 110 determines in advance the data write status to the memory of the SSD 400 under the control of the memory access control unit 108 in a state where the HDD 300 stops rotating the disk. Judgment is made as to whether the conditions are satisfied. For example, the memory write status determination unit 110 writes data that has been written to the address range of the memory of the SSD 400 corresponding to a certain LU and has not yet been written to the disk of the HDD 300 while the HDD 300 has stopped rotating the disk. Is equal to or greater than a threshold (for example, a ratio [%] to an allocated amount or an absolute capacity [GB]) determined for the LU, it is determined that the above condition is satisfied for the LU. Alternatively, the memory write status determination unit 110 may be configured such that the frequency at which data is written to the address range of the memory of the SSD 400 corresponding to a certain LU in a state where the HDD 300 stops rotating the disk, If it is equal to or greater than a threshold value determined for the LU (for example, the number of times per unit time [times / unit time]), it is determined that the above condition is satisfied for the LU. The memory write status determination unit 110 performs the same determination for each LU. As a result, it is possible to optimize the timing of spinning up the HDD 300 in order to transfer write data from the SSD 400 to the HDD 300 (delaying the spin-up of the HDD 300 as much as possible saves energy, but if it is delayed too much, a new The memory capacity of the SSD 400 for writing the write data is lost.

  Further, in the present embodiment, the disk rotation control unit 111 causes the HDD 300 to start the disk rotation when the memory writing state determination unit 110 determines that the data writing state to the memory of the SSD 400 satisfies the above condition. To control. It should be noted that the disk rotation control unit 111 is configured to start the disk rotation when there is no access to the RAID group to which the energy saving mode is designated or the RAID group to which the LU to which the energy saving mode is assigned belongs for a certain period of time. The HDD 300 is controlled.

  In the present embodiment, the disk access control unit 107 controls the SSD 400 to transfer the data written in the memory and the SSD 400 when the HDD 300 starts to rotate the disk under the control of the disk rotation control unit 111. The HDD 300 is controlled to write data transferred from the disk to the disk.

  As described above, in the present embodiment, when the rotation of the disk of the HDD 300 is stopped, data is written to the memory of the SSD 400 every time there is a write request, and the state of data writing to the memory of the SSD 400 is constant. If the condition is satisfied, the disk of the HDD 300 starts to rotate, and data is transferred from the memory of the SSD 400 to the disk of the HDD 300. Therefore, according to the present embodiment, while avoiding the occurrence of I / O delay and the reduction of the energy saving effect, the disk (the RAID group or LU HDD 300 in which the energy saving mode is designated) that stops the rotation is effectively used. It becomes possible.

  In the present embodiment, when the HDD 300 is spun down, I / O access delay can be minimized by writing write data from the host computer 500 to the SSD 400 from the cache memory 106.

  Also, by having the function of setting the HDD 300 targeted for the energy-saving mode in the disk array device 100 for each RAID group or LU, it is possible to perform spin-down / spin-up of the HDD 300 that has been intentionally controlled manually or externally. It can be automatically operated in the apparatus. A RAID group that requires constant access can be excluded from the energy-saving mode and operated while always in the spin-up state. The operation in the energy saving mode can be automated by referring to the I / O statistical information for the RAID group or LU.

Embodiment 2. FIG.
In the present embodiment, differences from the first embodiment will be mainly described.

  FIG. 5 is a block diagram showing the configuration of the disk array device 100 according to the present embodiment.

  In FIG. 5, the disk array control unit 105 of the controller 201 further includes a memory write presence / absence determination unit 112. Other configurations are the same as those of the disk array device 100 according to the first embodiment shown in FIG.

  Hereinafter, the operation of each part of the disk array device 100 according to the present embodiment shown in FIG. 5 (drive control method according to the present embodiment, program processing procedure according to the present embodiment) will be described.

  6 to 9 are diagrams showing the flow of data when the disk array device 100 processes a read request.

  FIG. 6 shows the data requested to be read when there is a read request specifying the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group to which the energy saving mode is specified is in the spin-up state (in operation). Shows the flow of data when is stored in the cache memory 106 (at the time of a cache hit).

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a read request (Read request) from the host computer 500. The read request is, for example, a SCSI Read command, and includes a parameter that specifies an identifier (for example, a number) of a RAID group or LU.
(2) When a read request is input from the host computer 500 to the host interface unit 102, the disk array control unit 105 reads the data requested to be read, that is, read data (Read data) from the cache memory 106, and The data is transmitted to the host computer 500 via the interface unit 102. As described above, when the data requested from the host computer 500 exists on the cache memory 106, the data is read from the cache memory 106.
(3) The disk array control unit 105 transmits a read completion report (I / O end report) to the host computer 500 via the host interface unit 102.

  Even when the HDD 300 of the RAID group to which the energy saving mode is designated is in the spin-down state (operation stop), the data flow at the time of a cache hit is the same as that shown in FIG.

  FIG. 7 shows the data requested to be read when there is a read request specifying the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group to which the energy saving mode is specified is in the spin-up state (in operation). Is stored in the cache memory 106 but is written in the memory of the SSD 400 (at the time of a cache miss / SSD hit), the data flow is shown.

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a read request (Read request) from the host computer 500. As described above, the read request is, for example, a SCSI Read command, and includes a parameter specifying a RAID group or an LU identifier (for example, a number).
(2) The memory access control unit 108 of the disk array control unit 105 transmits a read request to the SSD 400. At this time, the memory access control unit 108 reads the address of the memory of the SSD 400 corresponding to the RAID group or LU designated by the read request from the host computer 500 from the address mapping unit 109, and designates the address to the SSD 400. . As described above, when the data requested from the host computer 500 does not exist in the cache memory 106, a read request is made to the SSD 400. At this time, the memory writing presence / absence determining unit 112 of the disk array control unit 105 determines whether read data (Read data) is written in the memory of the SSD 400. Here, the memory writing presence / absence determining unit 112 determines that the read data is written in the memory of the SSD 400.
(3) The SSD 400 reads the read data from the memory address designated by the memory access control unit 108 in response to the read request transmitted by the memory access control unit 108 of the disk array control unit 105, and the disk array control unit 105 Send to. The disk array control unit 105 receives the read data transmitted from the SSD 400 and stores it in the cache memory 106. As described above, the data requested from the host computer 500 does not exist on the cache memory 106, but when it exists on the memory of the SSD 400, the data is read from the memory of the SSD 400.
(4) The SSD 400 transmits a read completion report to the disk array control unit 105.
(5) The disk array control unit 105 reads the read data received from the SSD 400 from the cache memory 106 and transmits it to the host computer 500 via the host interface unit 102.
(6) The disk array control unit 105 transmits a read completion report to the host computer 500 via the host interface unit 102.

  Even when the HDD 300 of the RAID group to which the energy saving mode is designated is in the spin-down state (operation stop), the data flow at the time of a cache miss / SSD hit is the same as that shown in FIG.

  As described above, in the present embodiment, the memory writing presence / absence determining unit 112 determines whether the HDD 300 has a disk in the HDD 300 both when the disk 300 starts rotating and when the disk rotation stops. When a data read request from the host computer 500 is input to the host interface unit 102, it is determined whether or not the data is written in the memory of the SSD 400.

  In the present embodiment, the memory access control unit 108 determines that the data requested to be read from the host computer 500 by the memory write presence / absence determination unit 112 is written in the memory of the SSD 400. The SSD 400 is controlled so as to read the data requested to be read from the memory.

  In the present embodiment, the host interface unit 202 outputs data read from the memory of the SSD 400 to the host computer 500 under the control of the memory access control unit 108.

  Further, in the present embodiment, the cache memory 106 reads from the memory of the SSD 400 if it does not store the data when a request to read data from the disk of the HDD 300 is input from the host computer 500 to the host interface unit 102. The read data is stored. As a result, the frequency of cache misses can be reduced.

  FIG. 8 shows the data requested to be read when there is a read request designating the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group designated with the energy saving mode is in a spin-up state (in operation). Shows the flow of data when is not stored in the cache memory 106 and is not written in the memory of the SSD 400 (at the time of cache miss / SSD miss).

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a read request (Read request) from the host computer 500. As described above, the read request is, for example, a SCSI Read command, and includes a parameter specifying a RAID group or an LU identifier (for example, a number).
(2) The memory access control unit 108 of the disk array control unit 105 transmits a read request to the SSD 400. At this time, the memory access control unit 108 reads the address of the memory of the SSD 400 corresponding to the RAID group or LU designated by the read request from the host computer 500 from the address mapping unit 109, and designates the address to the SSD 400. . As described above, when the data requested from the host computer 500 does not exist in the cache memory 106, a read request is made to the SSD 400. At this time, the memory writing presence / absence determining unit 112 of the disk array control unit 105 determines whether read data (Read data) is written in the memory of the SSD 400. Here, the memory writing presence / absence determining unit 112 determines that the read data is not written in the memory of the SSD 400.
(3) The disk access control unit 107 of the disk array control unit 105 reads a read request from the host computer 500 when the memory write presence / absence determination unit 112 determines that read data is not written to the memory of the SSD 400. The data is transmitted to the HDD 300 of the RAID group or LU specified in the request. As described above, when the data requested from the host computer 500 does not exist in the cache memory 106 or the memory of the SSD 400, a read request is issued to the HDD 300 constituting the specified RAID group or LU. Done.
(4) The HDD 300 reads the read data from the disk and transmits it to the disk array control unit 105 in response to the read request transmitted by the disk access control unit 107 of the disk array control unit 105. The disk array control unit 105 receives the read data transmitted from the HDD 300 and stores it in the cache memory 106. As described above, when the data requested from the host computer 500 does not exist in the cache memory 106 or the memory of the SSD 400, a read request is issued to the HDD 300 constituting the specified RAID group or LU. Done.
(5) The HDD 300 transmits a read completion report to the disk array control unit 105.
(6) The disk array control unit 105 reads the read data received from the HDD 300 from the cache memory 106 and transmits it to the host computer 500 via the host interface unit 102.
(7) The disk array control unit 105 transmits a read completion report to the host computer 500 via the host interface unit 102.
(8) The memory access control unit 108 of the disk array control unit 105 transmits the read completion report to the host computer 500 at a predetermined timing (for example, almost simultaneously with the read data being stored in the cache memory 106 or the read completion report). Immediately after that, the read data is transferred to the SSD 400. At this time, the memory access control unit 108 specifies the address of the memory of the SSD 400 corresponding to the RAID group or LU specified by the write request to the SSD 400. The SSD 400 writes the read data transferred by the memory access control unit 108 to the memory address designated by the memory access control unit 108. As described above, data written from the HDD 300 to the cache memory 106 is copied to the SSD 400 by destage processing asynchronously with the I / O of the host computer 500.

  As described above, in the present embodiment, the disk access control unit 107 determines that the data requested to be read from the host computer 500 by the memory write presence / absence determination unit 112 is not written to the memory of the SSD 400. The HDD 300 is controlled so that the data requested to be read from the host computer 500 is read from the disk.

  In the present embodiment, the host interface unit 202 outputs data read from the disk of the HDD 300 to the host computer 500 under the control of the disk access control unit 107.

  Further, in the present embodiment, the cache memory 106 starts from the disk of the HDD 300 if it does not store the data when a request to read data from the disk of the HDD 300 is input from the host computer 500 to the host interface unit 102. The read data is stored. As a result, the frequency of cache misses can be reduced.

  In this embodiment, the memory access control unit 108 controls the SSD 400 to write data read from the disk of the HDD 300 to the memory under the control of the disk access control unit 107. As a result, the frequency of SSD misses can be reduced.

  FIG. 9 shows the data requested to be read when there is a read request specifying the RAID group or an LU belonging to the RAID group when the HDD 300 of the RAID group to which the energy saving mode is specified is in a spin-down state (operation stop). Shows the flow of data when is not stored in the cache memory 106 and is not written in the memory of the SSD 400 (at the time of cache miss / SSD miss).

In this case, each unit of the disk array device 100 operates as follows.
(1) The host interface unit 102 receives an input of a read request (Read request) from the host computer 500. As described above, the read request is, for example, a SCSI Read command, and includes a parameter specifying a RAID group or an LU identifier (for example, a number).
(2) The memory access control unit 108 of the disk array control unit 105 transmits a read request to the SSD 400. At this time, the memory access control unit 108 reads the address of the memory of the SSD 400 corresponding to the RAID group or LU designated by the read request from the host computer 500 from the address mapping unit 109, and designates the address to the SSD 400. . As described above, when the data requested from the host computer 500 does not exist in the cache memory 106, a read request is made to the SSD 400. At this time, the memory writing presence / absence determining unit 112 of the disk array control unit 105 determines whether read data (Read data) is written in the memory of the SSD 400. Here, the memory writing presence / absence determining unit 112 determines that the read data is not written in the memory of the SSD 400.
(3) The disk rotation control unit 111 of the disk array control unit 105 rotates the spindle motor of the disk with respect to the HDD 300 when the memory writing presence / absence determination unit 112 determines that the read data is not written in the memory of the SSD 400. Request to resume. As described above, when the data requested from the host computer 500 does not exist in the cache memory 106 or the memory of the SSD 400, and the HDD 300 is spun down, the specified RAID group or LU is stored. A spin-up request is made to the HDD 300 that is configured before a read request.
(4) When the HDD 300 enters the spin-up state (resumption of operation), the HDD 300 reports to that effect to the disk array control unit 105.
(5) When the disk access control unit 107 of the disk array control unit 105 confirms that the HDD 300 is in the spin-up state, the read request is sent to the HDD 300 of the RAID group or LU specified by the read request from the host computer 500. Send to. Thus, when the data requested from the host computer 500 does not exist in the cache memory 106 or the memory of the SSD 400 and the HDD 300 is spun down, the HDD 300 is in the idle mode (Ready state). After waiting for this, a read request is made to the HDD 300 constituting the designated RAID group or LU.
(6) The HDD 300 reads the read data from the disk in response to the read request transmitted by the disk access control unit 107 of the disk array control unit 105 and transmits it to the disk array control unit 105. The disk array control unit 105 receives the read data transmitted from the HDD 300 and stores it in the cache memory 106. As described above, when the data requested from the host computer 500 does not exist in the cache memory 106 or the memory of the SSD 400, a read request is issued to the HDD 300 constituting the specified RAID group or LU. Done.
(7) The HDD 300 transmits a read completion report to the disk array control unit 105.
(8) The disk array control unit 105 reads the read data received from the HDD 300 from the cache memory 106 and transmits it to the host computer 500 via the host interface unit 102.
(9) The disk array control unit 105 transmits a read completion report to the host computer 500 via the host interface unit 102.
(10) The memory access control unit 108 of the disk array control unit 105 transmits the read completion report to the host computer 500 at a predetermined timing (for example, almost simultaneously with the read data being stored in the cache memory 106 or the read completion report). Immediately after that, the read data is transferred to the SSD 400. At this time, the memory access control unit 108 specifies the address of the memory of the SSD 400 corresponding to the RAID group or LU specified by the write request to the SSD 400. The SSD 400 writes the read data transferred by the memory access control unit 108 to the memory address designated by the memory access control unit 108. As described above, data written from the HDD 300 to the cache memory 106 is copied to the SSD 400 by destage processing asynchronously with the I / O of the host computer 500.

  As described above, in this embodiment, the disk rotation control unit 111 determines that the data requested to be read from the host computer 500 by the memory writing presence / absence determination unit 112 is not written to the memory of the SSD 400. The HDD 300 is controlled to start the disk rotation.

  In this embodiment, when the HDD 300 starts rotating the disk under the control of the disk rotation control unit 111, the disk access control unit 107 reads the HDD 300 so that the data requested to be read from the host computer 500 is read from the disk. Control.

  As described above, in the present embodiment, when the HDD 300 is spun down, the I / O access delay can be minimized by storing the read data in the SSD 400 by the read I / O before spin down. it can.

  The capacity of the volatile memory used for the cache memory 106 is generally several gigabytes. The capacity of the flash memory used for the SSD 400 is generally several tens to several hundreds gigabytes. The capacity of a magnetic disk used for the HDD 300 is generally several terabytes. As in the example illustrated in FIG. 2, when the SSD 400 is not provided between the cache memory 106 and the HDD 300, when trying to use the standby mode HDD 300 in the same manner as the normal (idle mode) HDD 300, cache misses frequently occur. Will occur. As a result, the number of times that the HDD 300 is spun up may be increased, or the HDD 300 may be spun up for a long time, and the energy saving effect may hardly be obtained. On the other hand, in the present embodiment, since the SSD 400 is provided between the cache memory 106 and the HDD 300, there is a high possibility that read data is stored in the SSD 400 even if a cache miss occurs. Therefore, according to the present embodiment, the energy saving effect can be enhanced.

  Conventionally, the energy saving function can be applied only to a RAID group that has been stopped for a long time, such as a RAID group for backup use, and the time period is also determined in advance. It is possible to apply the energy saving function to the RAID group, and reduction of power consumption can be expected.

  As mentioned above, although embodiment of this invention was described, you may implement in combination of 2 or more among these embodiment. Alternatively, one of these embodiments may be partially implemented. Alternatively, two or more of these embodiments may be partially combined. In addition, this invention is not limited to these embodiment, A various change is possible as needed.

  100 disk array device, 101 controller, 102 host interface unit, 103 network interface unit, 104 disk control unit, 105 disk array control unit, 106 cache memory, 107 disk access control unit, 108 memory access control unit, 109 address mapping unit, 110 memory write status determination unit, 111 disk rotation control unit, 112 memory write presence / absence determination unit, 200 disk array device, 201 controller, 202 host interface unit, 203 network interface unit, 204 disk control unit, 205 disk array control unit, 206 Cache memory, 300 HDD, 400 SSD, 500 Host computer.

Claims (9)

A drive control device that controls a disk drive that stores data in a disk and a memory drive that stores data in a memory,
A host interface for receiving input from the host computer;
When the disk drive has stopped rotating, the memory drive writes the data to the memory each time a request to write data to the disk is input from the host computer to the host interface unit. A memory access control unit for controlling
A memory write status determination unit that determines whether a data write status to the memory under the control of the memory access control unit satisfies a predetermined condition in a state where the disk drive stops rotating the disk When,
A disk rotation control unit for controlling the disk drive to start the rotation of the disk, when the memory writing state determination unit determines that the writing state satisfies the condition;
When the disk drive starts rotating the disk under the control of the disk rotation control unit, the memory drive is controlled to transfer the data written in the memory and the data transferred from the memory drive is transferred to the memory drive. A drive control apparatus comprising: a disk access control unit that controls the disk drive to write to a disk. The drive control device further includes:
The memory is different from the memory, and includes a cache for storing the data every time a request to write data to the disk is input from the host computer to the host interface unit,
When the disk drive stops the rotation of the disk, the memory access control unit stores the data every time a request to write data to the disk is input from the host computer to the host interface unit. Controlling the memory drive to transfer from the cache to the memory drive and write the data to the memory;
When the disk drive has started rotating the disk, the disk access control unit receives the data every time a request to write data to the disk is input from the host computer to the host interface unit. 2. The drive control apparatus according to claim 1, wherein the disk drive is controlled to transfer the data from a cache to the disk drive and write the data to the disk.   The memory write status determination unit sets the amount of data written to the memory under the control of the memory access control unit and not yet written into the disk under the control of the disk access control unit to a predetermined amount. 3. The drive control device according to claim 1, wherein when it reaches, the write status is determined to satisfy the condition.   The memory write status determination unit determines that the write status satisfies the condition when the frequency at which data is written to the memory reaches a predetermined frequency under the control of the memory access control unit. The drive control apparatus according to claim 1. The drive control device further includes:
An address mapping unit for storing a correspondence relationship between a plurality of logical units obtained by dividing the storage area of the disk and addresses of the memory individually assigned to the respective logical units;
When the disk drive has stopped rotating the disk, the memory access control unit receives a write request for designating any one of the plurality of logical units as a data write request to the disk. The memory drive is controlled so that the address of the memory corresponding to the logical unit is read from the address mapping unit and the data is written to the address each time the host computer inputs the host interface unit. The drive control device according to any one of claims 1 to 4.   The memory write status determination unit is configured to determine whether the data write status to the memory under the control of the memory access control unit is in advance for each logical unit in a state where the disk drive has stopped rotating the disk. 6. The drive control apparatus according to claim 5, wherein it is determined whether or not a condition defined in the above is satisfied. The drive control device further includes:
When the disk drive stops rotating the disk, when a data read request from the disk is input from the host computer to the host interface unit, it is determined whether the data is written in the memory. A memory writing presence / absence determining unit for determining,
The disk rotation control unit determines the disk drive to start rotation of the disk when the memory write presence / absence determination unit determines that the data requested to be read from the host computer is not written to the memory. Control
The disk access control unit controls the disk drive to read data requested to be read from the host computer from the disk when the disk drive starts rotating the disk under the control of the disk rotation control unit. ,
The host interface unit outputs data read from the disk to the host computer under the control of the disk access control unit;
7. The drive control according to claim 1, wherein the memory access control unit controls the memory drive to write data read from the disk into the memory under the control of the disk access control unit. apparatus. A drive control device according to any one of claims 1 to 7;
The disk drive;
A storage system comprising the memory drive. A drive control method for controlling a disk drive for storing data in a disk and a memory drive for storing data in a memory,
When a drive control device including a host interface unit that receives an input from a host computer stops the rotation of the disk, a data write request to the disk is input from the host computer to the host interface unit. Each time the memory drive is controlled to write the data to the memory,
The drive control device determines whether or not the state of writing data to the memory satisfies a predetermined condition in a state where the disk drive stops rotating the disk,
If the drive control device determines that the write status satisfies the condition, the drive control device controls the disk drive to start rotation of the disk;
When the disk drive starts to rotate the disk, the drive control device controls the memory drive to transfer data written in the memory and transfers data transferred from the memory drive to the disk. A drive control method comprising controlling the disk drive to write.

Images