JP2009252106A - Disk array control method and disk array device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a disk array device maintaining an optimum load state. <P>SOLUTION: In the disk array device which manages addresses for a plurality of pools 111 each composed of storage areas of a plurality of hard disk drives (HDD) 11, and performs access according to a request from a host device, the load state of the pools 111-113 is periodically monitored by use of timers 12b and 13b. When certain data in a pool is accessed with high frequency, the data is relocated from the pool to the other pool to distribute the load. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a disk array control method and a disk array apparatus.

FIG. 7 is a diagram showing an outline of a conventional disk array device.
In this disk array device, a disk drive group is accommodated in the disk unit 1, a logical disk (LD) 2 is formed in a storage area of the disk drive with respect to a host device, and a pool 3 is formed by an arbitrary number of logical disks 2. In addition, the controller 5 manages access from the host device 4.

  As shown in FIG. 7, a conventional disk array device includes a redundant interface path to the host device 4 for load distribution. In this case, the load distribution of each path is effective, but when access to a specific pool 3 of the disk unit 1 is concentrated, the disk drive forming the storage area of the pool is overloaded, and the response of the disk unit 1 Performance decreases. Therefore, there is a problem that the input / output performance as the disk array device is lowered.

  On the other hand, the disk device disclosed in Patent Document 1 below counts the number of accesses to the disk drive group from the host device, and stripes corresponding to the pool are generated based on the number of accesses within a predetermined period. It is determined whether the load is high or low, and for the stripe determined to be high, data can be moved to a disk drive prepared in advance for the data update processing after the determined time point. In addition, a stripe area of the same capacity is secured as a copy destination, data is directly written to the copy destination area, data is not written in the write process, and data that has not been updated is later processed by internal processing. To the copy destination area.

The storage device disclosed in the following Patent Document 2 measures the load applied to the volume corresponding to the pool for each access from the higher-level device, and based on the result, the contents of the volume whose load exceeds a predetermined value are stored in the spare volume. The copy source volume and the copy destination spare volume are associated with one virtual volume and provided to the host device, and a write request (read modify write) accompanied by data read processing from the host device to the virtual volume is received. When given, data is read from one volume in the virtual volume, data is written to all the volumes and spare volumes, and when a read request for the virtual volume is given alone, it is read by read modify write. From other volumes or spare volumes Read the data you want to respond.
JP 2007-293486 A JP 2007-328734 A

  In Patent Document 1 and Patent Document 2, load distribution at the time of accessing a stripe and volume corresponding to a virtual logical disk is possible, but the capacity and load of each stripe and volume cannot be made uniform. Therefore, there is a problem that the optimum load state cannot be maintained.

  An object of the present invention is to provide a disk array device that can solve the above-described problems and maintain an optimum load state.

In order to achieve the above object, a disk array control method according to the first aspect of the present invention includes:
In a disk array control method for managing addresses for a plurality of pools each configured with a storage area of a physical disk and performing access according to a request from a host device,
A monitoring process for periodically monitoring the load status in each pool;
And load distribution processing for moving data stored in a specific pool to another pool based on the monitoring result of the load status.

In order to achieve the above object, a disk array device according to a second aspect of the present invention provides:
In a disk array device that manages addresses for a plurality of pools each configured with a storage area of a physical disk and performs access according to a request from a host device,
Monitoring means for periodically monitoring the load status in each pool;
Load balancing means for moving data stored in a specific pool to another pool based on the monitoring result of the load status is provided.

  According to the present invention, a disk array device capable of maintaining an optimum load state can be realized.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[First Embodiment]
FIG. 1 is a configuration diagram showing a disk array device according to an embodiment of the present invention.
The disk array device 10 integrates a plurality of hard disk drives (hereinafter referred to as HDDs) 11 which are physical drives into RAID (Redundant Array of Inexpensive Disks), and stores a storage space 100 composed of storage areas of the HDD 11 group as a host device 20. It is a device to provide.

  The disk array device 10 includes two controllers 12 and 13, a disk unit 14, and a memory 15. A plurality of HDDs 11 are accommodated in the disk unit 14, and a storage space 100 is formed by storage areas of the plurality of HDDs 11.

  The controller 12 includes a disk control unit 12a and a timer 12b. The disk control unit 12a periodically monitors the state of the disk unit 14 using the timer 12b, manages the address of the storage space 100 of the disk unit 14 using the memory 15, and accesses the storage space 100 of the disk unit 14. And load distribution in the storage space 100 is performed.

The controller 13 is redundantly provided with respect to the controller 12, and includes a disk control unit 13a and a timer 13b. The disk control unit 13a of the controller 13 functions in the same manner independently of the disk control unit 12a of the controller 12.
The controller 12 and the controller 13 may each be composed of a CPU or the like.

  The controller 12 is connected to a plurality of higher-level devices 20 through interface paths. The controller 13 is connected to a plurality of higher-level devices 20 through interface paths independently of the controller 12. That is, redundant interface paths are provided so that the controllers 12 and 13 can receive requests from the host device 20, respectively.

FIG. 2 is a diagram for explaining the storage space 100 of the disk unit 14.
A virtual logical disk 110 in which a plurality of pools 111, 112, 113 are integrated is set in the storage space 100 of the disk unit 14. In each of the pools 111, 112, and 113, a logical disk (LD) 114 is arranged. The capacity of the logical disk 114 is the sum of the storage areas of a plurality of HDDs 11, for example.

Next, the operation of the disk array device 10 will be described.
FIG. 3 is a flowchart for explaining a disk array control method performed by the disk array apparatus 10.

  The disk control units 12a and 13a of the controllers 12 and 13 virtually manage the pools 111 to 113 of the disk unit 14 as one logical disk, and the logical addresses for the physical addresses of the individual pools 111 to 113 To manage. Further, the disk control units 12a and 13a integrate the logical addresses of the individual pools 111 to 113 and manage them as logical addresses of one virtual logical disk. Thus, the correlation between an arbitrary logical address of the virtual logical disk and each Pool (LD) in the virtual logical disk can be maintained. Access to each of the pools 111 to 113 in the virtual logical disk is performed using this logical address.

  In step S101, the disk controllers 12a and 13a have an instruction from the host device 20, and when an instruction to stop control is given (step S101: YES), the disk controller 12a and 13a stop the disk array control method and end the processing.

  On the other hand, the timers 12b and 13b measure the load confirmation time for periodically confirming the load state and notify the disk control units 12a and 13a. If the disk control units 12a and 13a are not instructed to stop the control (step S101: NO) and are notified by the timers 12b and 13b that the load confirmation time has elapsed (step S102: YES), the timers 12b and 13a 13b is reset and information indicating the load state is acquired from the memory 15 (step S103).

  The information indicating the load status includes, for example, current information indicating whether the current load status of each of the pools 111 to 113 is high or low, use information indicating the amount of data written in each of the pools 111 to 113, and each pool There is the number of accesses per predetermined time of data written in 111 to 113, that is, statistical information indicating the access frequency. The disk controllers 12a and 13a share and refer to information indicating these load states.

  The disk controllers 12a and 13a obtain the load amount of each of the pools 111 to 113 based on the statistical information, and start the load distribution process when the load amount exceeds a predetermined value (step S104).

FIG. 4 is an explanatory diagram of the load distribution process.
In the load distribution process in step S104, for example, when data with high access frequency is concentrated in a specific pool, data with high access frequency is relocated to another pool. In addition, a part of data is rearranged in another pool from a pool in which a large amount of data is written.

  Since the pool is in a high load state during the load distribution process, information indicating that the load of the pool is high is stored in the memory 15 as current information at the start of the load distribution process. After the load distribution process is completed, information indicating that the load is low is indicated on the memory 15 as current information.

  When a load request or a read request is received from the host device 20 after the load distribution process is started or when the timer 12b or 13b is not informed that the load confirmation time has elapsed (step S102: NO) (step S102) S105: YES), the disk control units 12a and 13a refer to the information indicating the load state on the memory 15 (step S106), and access the disk unit 14 in response to the request from the host device 20 (step S107). ).

  When the request from the host device 20 is writing, data is written into the virtual logical disk 110 of the disk unit 14. At this time, the disk control units 12a and 13a set the pool in which the data is actually written as the lowest load state pool based on the load information stored in the memory 15. If the request from the host device 20 is a read request, the disk control units 12a and 13a read the data from the pool in the virtual logical disk 110 where the data is actually written, and the host device 20 Forward to.

  When the access in step S107 is completed, the disk control units 12a and 13a update the statistical information on the memory 15. That is, the information indicating the access frequency of the pool in which data has been written or read is rewritten (step S108).

  When step S108 is completed, or when there is no write or read request from the host device 20 (step S105: NO), the process returns to step S101, and a series of processes is repeated.

  As described above, in the disk array device 10 of the present embodiment, the load state of the pools 111 to 113 is monitored for each period set by the timers 12b and 13b, and data is transferred between pools according to the result. It is moved. As a result, data with high access frequency does not concentrate in a specific pool among the pools 111 to 113. Also, the usage rate of each pool is averaged. Therefore, the load state of each pool 111-113 can be equalized.

[Second Embodiment]
FIG. 5 is a configuration diagram showing a disk array device 30 according to the second embodiment of the present invention.
In this disk array device 30, unused HDDs 16 are arranged in the virtual logical disk 110 of the disk unit 14, and unused HDDs 16 are used when the capacity of the pools 111 to 113 is insufficient. Other configurations are the same as those of the disk array device 10 of the first embodiment.

  In this case, the disk control units 12a and 13a of the controllers 12 and 13 manage the address of the storage space 100 of the disk unit 14 using the memory 15, and periodically check the status of the disk unit 14 using the timers 12b and 13b. , And accesses the storage space 100 of the disk unit 14 and performs load distribution and capacity expansion in the storage space 100.

Next, the operation of the disk array device 30 will be described.
FIG. 6 is a flowchart for explaining a disk array control method performed by the disk array device 30.

  The disk control units 12a and 13a of the controllers 12 and 13 respectively manage the pools 111 to 113 of the disk unit 14 virtually as one logical disk as in the first embodiment, and the individual pools 111 to 113 are managed. A physical address is managed in association with a logical address. The disk control units 12a and 13a integrate the logical addresses of the individual pools 111 to 113 and manage them as logical addresses of one virtual logical disk. Thereby, the correlation between an arbitrary logical address of the virtual logical disk and each of the pools 111 to 113 in the virtual logical disk can be maintained. Access to each of the pools 111 to 113 in the virtual logical disk is performed using this logical address.

  In step S201, the disk controllers 12a and 13a have an instruction from the host device 20, and when an instruction to stop control is given (step S201: YES), the disk controller 12a stops and stops the disk array control method.

  On the other hand, the timers 12b and 13b measure the capacity confirmation time for periodically confirming the capacity of the pools 111 to 113, and the load confirmation time for periodically confirming the load state of the pools 111 to 113. Keep time.

  The disk controllers 12a and 13a are not given an instruction to stop control (step S201: NO), and the timers 12b and 13b indicate that the capacity check time for checking the capacity of the pools 111 to 113 has passed by the timers 12b and 13b. When notified from 13b (step S202: YES), the timers 12b and 13b are reset, and a pool with insufficient capacity is detected from the pools 111 to 113 (step S203).

In the process of detecting a pool with insufficient capacity, the usage state of each of the pools 111 to 113, that is, the number of written data is checked and stored in the memory 15. Then, a pool whose usage state exceeds a predetermined value is detected as a pool with insufficient capacity.
The disk controllers 12a and 13a that have detected the insufficient capacity pool allocate the unused HDD 16 in the virtual logical disk 110 to the insufficient capacity pool and make it usable as a part of the insufficient capacity pool. That is, capacity expansion is performed (step S204).

  During the period of capacity expansion, the pool is in a high load state. At the start of capacity expansion, information indicating that the load of the pool is high is displayed on the memory 15 as current information, After the capacity expansion process is completed, information indicating that the load is low is displayed on the memory 15 as current information.

  Further, when there is no unused HDD 16 in the virtual logical disk 110, the disk control units 12a and 13a allocate a new unused HDD to the virtual logical disk 110 (step S205).

  If a pool with insufficient capacity is not detected, the processes in steps S204 and S205 are not performed.

  The disk control units 12a and 13a are not informed by the timers 12b and 13b that the capacity confirmation time has elapsed from the timers 12b and 13b (step S202: NO), and the load confirmation times of the pools 111 to 113 have elapsed. When notified of this (step S206: YES), the timers 12b and 13b are reset, and information indicating the load state is acquired from the memory 15 (step S207).

  The information indicating the load status includes, for example, current information indicating whether the current load status of each of the pools 111 to 113 is high or low, use information indicating the amount of data written in each of the pools 111 to 113, and each pool There is the number of accesses per predetermined time of data written in 111 to 113, that is, statistical information indicating the access frequency. The disk controllers 12a and 13a refer to information indicating these load states in common.

The disk controllers 12a and 13a start load distribution processing for each of the pools 111 to 113 based on the statistical information (step S208).
In the load distribution process in step S208, as in the first embodiment, when data with high access frequency is concentrated in a specific pool, a part of data with high access frequency is transferred to another pool. Rearrange. In addition, a part of data is rearranged in another pool from a pool in which a large amount of data is written.

  Since the pool is in a high load state during the load distribution process, information indicating that the load of the pool is high is stored in the memory 15 as current information at the start of the load distribution process. After the load distribution processing is completed, information indicating that the load is low is indicated on the memory 15 as current information.

  Further, when the disk control units 12a and 13a are not informed that the load confirmation time has elapsed from the timers 12b and 13b (step S206: NO), when there is a write request or a read request from the host device 20 (step S206). S209: YES), referring to the information indicating the load state on the memory 15 (step S210), the disk unit 14 is accessed in response to the request from the host device 20 (step S211).

  When the request from the host device 20 is writing, data is written into the virtual logical disk of the disk unit 14. At this time, the disk control units 12a and 13a set the pool in which the data is actually written as the lowest load state pool based on the load information stored in the memory 15. If the request from the host device 20 is a read request, the disk controllers 12a and 13a read the data from the pool in the virtual logical disk where the data is actually written, and send it to the host device 20. Forward.

  When the access is completed, the disk controllers 12a and 13a update the statistical information on the memory 15. That is, the information indicating the access frequency of the pool in which data has been written or read is rewritten (step S212).

  When step S212 is completed, or when there is no write or read request from the host device 20 (step S209: NO), the process returns to step S201, and a series of processes is repeated.

  As described above, in the disk array device of this embodiment, capacity expansion processing is performed. Therefore, in addition to the advantages of the first embodiment, even if the capacity of each of the pools 111 to 113 is insufficient, it is unused. There is an advantage that the capacity of each of the pools 111 to 113 is supplemented by the HDD 16.

1 is a configuration diagram showing a disk array device according to a first embodiment of the present invention. FIG. It is a figure for demonstrating the storage space of a disc part. It is a flowchart for demonstrating the disk array control method which a disk array apparatus performs. It is explanatory drawing of a load distribution process. It is a block diagram which shows the disk array apparatus based on the 2nd Embodiment of this invention. It is a flowchart for demonstrating the disk array control method which a disk array apparatus performs. It is a figure which shows the outline | summary of the conventional disk array apparatus.

Explanation of symbols

10, 30 Disk array device 11 HDD
12, 13 Controller 12a, 13a Disk control unit 12b, 13b Timer 14 Disk unit 20 Host device 100 Storage space 110 Virtual logical disk 111-113 Pool

Claims (12)

In a disk array control method for managing addresses for a plurality of pools each configured with a storage area of a physical disk and performing access according to a request from a host device,
A monitoring process for periodically monitoring the load status in each pool;
Load distribution processing for moving data stored in a specific pool to another pool based on the monitoring result of the load status;
A disk array control method comprising: The monitoring process monitors the load status including the frequency of access in each pool,
2. The disk array control according to claim 1, wherein the load distribution processing moves a part of the frequently accessed data from the pool in which frequently accessed data is concentrated to the other pool. Method. The monitoring process monitors the load situation including the amount of data written to each pool,
3. The disk array control method according to claim 1, wherein the load distribution processing moves a part of data from a pool with a large amount of written data to the other pool. An unused physical disk is placed on a virtual logical disk composed of the plurality of pools,
The capacity expansion process of periodically monitoring the usage status of the storage area of each pool and allocating the storage area of the unused physical disk to a pool whose usage status exceeds a predetermined value. 4. The disk array control method according to any one of 1 to 3.   5. The disk array control method according to claim 4, wherein when the capacity of the virtual logical disk is insufficient, a new unused physical disk is arranged on the virtual logical disk.   6. The disk array control method according to claim 1, wherein when the request from the host device is data writing, the data is written to a pool with the lowest load. In a disk array device that manages addresses for a plurality of pools each configured with a storage area of a physical disk and performs access according to a request from a host device,
Monitoring means for periodically monitoring the load status in each pool;
A disk array device comprising load distribution means for moving data stored in a specific pool to another pool based on the monitoring result of the load status. The monitoring means monitors the load situation including the frequency of access in each pool,
8. The disk array device according to claim 7, wherein the load distribution unit moves a part of the frequently accessed data from a pool in which frequently accessed data is concentrated to the other pool. . The monitoring means monitors the load situation including the amount of data written to each pool,
9. The disk array device according to claim 7, wherein the load distribution processing moves part of data from a pool with a large amount of written data to the other pool. An unused physical disk is placed on a virtual logical disk composed of the plurality of pools,
10. The storage area of each pool is periodically monitored, and the storage area of the unused physical disk is allocated to a pool whose usage condition exceeds a predetermined value. 2. The disk array device according to claim 1.   11. The disk array device according to claim 10, wherein when the capacity of the virtual logical disk is insufficient, a new unused physical disk is arranged on the virtual logical disk. The disk array device according to any one of claims 7 to 11, wherein when the request from the host device is a data write, the data is written to a pool with the lowest load.