JP4985391B2 - Disk array device, physical disk recovery method, and physical disk recovery program - Google Patents

Description

  The present invention relates to suppression of power consumption of a storage system, and particularly relates to suppression of power consumption of a storage system in which a plurality of physical disks are connected, such as a RAID configuration.

  In recent years, a storage system configured by a disk array in which a plurality of physical disks are combined is combined with an increasing number of physical disks, and the capacity of each physical disk is also increasing. For this reason, the cost, especially power consumption, of operating these storage systems has become a problem.

  As a technique for suppressing power consumption in such a disk array, MAID (Massive Arrays of Inactive Disks) is known. This is based on the locality of data access that 80% of the data access occurs with respect to 20% of the physical area, and the rotation of most of the physical disks constituting the disk array is normally stopped ( Hereinafter, the physical disk is rotated again only when an access to the physical disk in the stopped state occurs.

  At that time, there are two methods of rotating the physical disk in the stopped state again (hereinafter referred to as recovery), that is, an I / O detection method and a command method. In the I / O detection method, when there is no data read / write data access request (hereinafter referred to as I / O request) to a physical disk for a certain period of time, the physical disk is stopped, and then there is an I / O request. It is a method of returning in the event of a failure. The command method is a method in which the physical disk is recovered when a return command is received from the outside.

  There are the following patent documents concerning the suppression of the power consumption of the disk array. Patent Document 1 discloses a disk array in which a physical disk is stopped when there is no I / O request for a predetermined time. Patent Document 2 discloses a disk array in which one or more physical disks are left in a stopped state when there is no I / O request for a certain period of time.

  Patent Document 3 discloses a disk device in which buffer memories are compared in units of blocks and only one block in which the same data is recorded is left and the rest is released. Patent Document 4 discloses a disk array that controls supply of power to a plurality of physical disks constituting the disk array in accordance with write access. Patent Document 5 discloses a disk array that corresponds to an I / O request for a stopped physical disk with a cache memory and a spare disk device.

JP 2000-293314 A JP 2004-252570 A JP 2006-286083 A JP 2006-302300 A JP 10-083614 A

  However, when the physical disk is restored by the I / O detection method, the physical disk resumes after the disk array receives the I / O request, so that the physical disk can be restored and respond to the I / O request. It may take time to become. At that time, depending on the application, a timeout (I / O error) may occur if the response to the I / O request does not exceed a certain time.

  In order to restore a physical disk by the command method, it is necessary to issue a restoration command explicitly from the side of the device that issues the I / O request. In order to do so, it is necessary to modify an existing backup script or the like. Therefore, particularly in the case of a disk array including a large number of physical disks, the labor involved in the modification becomes complicated.

  In the above-mentioned Patent Document 5, as a response to the I / O request in such a case, the stopped physical disk is started while recording the requested data in the spare disk device, and the startup is completed. The data is written to the physical disk. However, none of Patent Documents 1 to 5 describes a configuration for bringing a stopped physical disk into a stopped state for a longer time without causing a timeout.

  An object of the present invention is to provide a disk array device, a physical disk recovery method, and a physical disk recovery program that do not require an explicitly issued return command and can maintain a stopped state for a longer time without causing a timeout. There is.

In order to achieve the above object, a disk array device according to the present invention includes a plurality of physical disks that can be individually stopped or operated, and an I / O receiving unit that receives a read command and a write command from a host device. A buffer for temporarily storing data to be written to the physical disk by a write command, and when a stopped physical disk receives a write command, the physical disk is controlled to start operation and the data to be written to the buffer is stored. return a response indicating that holding is successful the data to the upper device, to flush the buffer from the physical disk becomes in the running state with a control means for writing the data stored in the buffer to the physical disk, control When a read command is received from the host device, the physical disk is stopped Restore data from operation and stored in the other physical disks are data and parity without commanding the operation start, characterized <br/> to return the recovered data to the host device.

In order to achieve the above object, a physical disk recovery method according to the present invention is applied to a disk array apparatus including a buffer for temporarily storing data and a plurality of physical disks that can be stopped or operated individually. A physical disk recovery method for operating a stopped physical disk, wherein the disk array device receives a write command from a host device, and a buffering step for storing data to be written to the physical disk in a buffer Start the operation of the stopped physical disk and return to the host device a response indicating that the data to be written to the physical disk has been successfully retained, and flush the buffer after the physical disk becomes active Flash process and data stored in buffer to physical disk A write step of burn them, and a read instruction reception step of the disk array device receives a read command from the host device, without starting running physical disk stopped, data stored in other physical disks running and It has a data restoration process which restores data from parity, and a data return process which returns restored data to a higher rank apparatus .

In order to achieve the above object, a physical disk recovery program according to the present invention is applied to a disk array device including a buffer for temporarily storing data and a plurality of physical disks that can be stopped or operated individually. The disk array device includes a buffer for temporarily storing data and a plurality of physical disks that can be individually stopped or operated, and writing to a computer included in the disk array device from a host device Processing to store data to be written to the physical disk in response to the command, and processing to return to the host device a response indicating that the stopped physical disk has been started and that the data to be written to the physical disk has been successfully held The process of flushing the buffer after the physical disk is in operation, A process of writing data stored in § physical disk, in response to the read command from the host device, stored physical disks stopped without starting operation, in addition to the physical disks running A process for restoring data from the data and parity and a process for returning the restored data to the host device are executed.

  In the present invention, as described above, the data to be written to the physical disk is stored in the buffer, and after the physical disk is in the operating state, the buffer is flushed and the data is written, so the rotation of the physical disk is resumed. There is no need to wait until it is done, and a reply can be sent to the host device. As a result, an unprecedented superior disk array device, physical disk recovery method, and physical disk recovery that can maintain a stopped state for a longer period of time without requiring an explicit return instruction and without causing a timeout. A program can be provided.

[First Embodiment]
FIG. 1 is a conceptual diagram showing a configuration of a storage system 10 according to the first embodiment of the present invention. The storage system 10 includes a business server 100 that is a host computer, and a disk array 200 having a RAID 6 (2 + PQ) configuration that reads and writes data in response to an I / O request from the business server 100.

  The business server 100 and the disk array 200 may be physically separate devices, or may be devices integrated into one. As a connection method between the business server 100 and the disk array 200, any interface such as Fiber Channel (FC), Ultra SCSI, or the like can be applied.

  The business server 100 includes an I / O request unit 101. The disk array 200 includes an I / O reception unit 201, a management unit 202, a control unit 203, a buffer 204, and a disk enclosure 210. The disk enclosure 210 includes physical disks 211 to 214 and a cooling device 215.

  The I / O request unit 101 issues I / O requests such as a read command (read), a write command (write), and a stop command to the disk array 200. The I / O accepting unit 201 accepts an I / O request from the I / O request unit 101 and returns a result of executing the I / O request (hereinafter referred to as an I / O result) to the I / O request unit 101. The buffer 204 holds an unprocessed I / O request and an I / O result.

  The physical disks 211 to 214 are assigned drive symbols D01 to D04, respectively. These physical disks 211 to 214 constitute a RAID 6 logical disk, and follow the I / O request from the I / O request unit 101. Operate. The physical disks 211 to 214 can be individually stopped or operated under the control of the control means 203 described later. The cooling device 215 cools heat released from the physical disks 211 to 214 inside the disk enclosure 210.

  The control unit 203 controls the operation / stop of the physical disks 211 to 214 and the cooling device 215. The management unit 202 stores and manages the state of the disk enclosure 210, RAID configuration information, the state of the logical disk, the state of the physical disk, and the condition for stopping and starting the physical disk.

  FIG. 2 is a state transition diagram for explaining the operation state of the disk enclosure 210 disclosed in FIG. The disk enclosure 210 has three operation states: full operation S1, power saving S2, and returning S3. In the full operation S1, all of the physical disks 211 to 214 are rotating. Hereinafter, the state in which each physical disk is rotating is referred to as operation.

  When a stop command is received from the full operation S1, the disk enclosure 210 shifts to the power saving S2. In the power saving S2, the physical disks 211 and 212 (D01 and D02) continue to operate, and the physical disks 213 and 214 (D03 and D04) stop. The management unit 202 stores in advance a condition for stopping the physical disks 213 and 214 (D03 and D04) in the power saving S2.

  When an I / O command (details will be described later) is received from the power saving S2, the disk enclosure 210 shifts to S3 during restoration. In the returning S3, the physical disks 211 and 212 (D01 and D02) continue to rotate, and the physical disks 213 and 214 (D03 and D04) accelerate the rotation speed from the stopped state.

  If the rotational speeds of the physical disks 213 and 214 (D03 and D04) become the speed at which data can be read and written, all of the physical disks 211 to 214 are in an operating state, and the disk enclosure 210 is in an all operating S1 state. At that time, the buffer 204 is flushed, and the data stored in the buffer 204 is written to the physical disks 211 to 214 (details will be described later).

  FIG. 3 is an activity diagram showing processing when the I / O request unit 101 issues a read command when the disk enclosure 210 is power saving S2 in the storage system 10 shown in FIG. When the I / O request unit 101 issues a read command (step S301) and the I / O reception unit 201 receives the read command (step S302), the I / O reception unit 201 determines the state of the disk enclosure 210. (Step S303). As described above, the disk enclosure 210 is power saving S2.

  In response to the determination that the disk enclosure 210 is in power saving S2 in step S303, the control unit 203 issues an instruction to transition to S3 during restoration to the management unit 202 (step S304). In response to this instruction, the management unit 202 shifts to S3 during restoration, and returns information on the physical disk to be started to the control unit 203 (step S305). In the example shown in FIG. 2, the physical disks 213 and 214 (D03 and D04) that are stopped at the power saving S2 are the targets of the operation start.

  The control unit 203 instructs the disk enclosure 210 to start the operation of the physical disks 213 and 214 (D03 and D04) to be started based on the information received from the management unit 202 in step S305 (step S306). ). The disk enclosure 210 starts operating the instructed physical disk (step S307) and returns the operation result to the control means 203 (step S308). If the operation is successful, the disk enclosure 210 becomes the full operation S1 in which all of the physical disks 211 to 214 (D01 to 04) are operating.

  Subsequently, the I / O accepting unit 201 issues a read instruction to the operating physical disk group (step S309). In response to the read instruction, the disk enclosure 210 reads data from the operating physical disk group (step S310).

  The I / O accepting unit 201 restores the original data on the stopped physical disk from the data and parity received from the operating physical disk group (step S311), and returns them to the I / O request unit 101. (Step S312). Finally, the I / O request unit 101 receives the requested data (step S313). The method for restoring the original data in step S311 is well known in the RAID technology and will not be described in particular.

  In the process of the read command, steps S304 to S308 can be omitted to keep the power saving S2 longer. In this case, as indicated by the thick line S351 in FIG. 3, the process jumps directly from step S303 to step S309.

  FIG. 4 is an activity diagram showing processing when the I / O request unit 101 issues a write command when the disk enclosure 210 is power saving S2 in the storage system 10 shown in FIG. Steps up to step S408 are substantially the same as those in the case of the read command described in FIG. 3, and therefore, reference numerals are also given by adding +100 to each step in FIG.

  However, when the disk enclosure 210 is instructed to start operation of the physical disks 213 and 214 (D03 and D04) in step S406, if there is data to be written stored in the buffer 204, the control unit 203 In parallel with the processing subsequent to the start of operation of the physical disk in step S407, the buffer 204 is flushed after the physical disk is in an operating state, and the data stored in the buffer 204 is written to the physical disk. (Step S451).

  The I / O accepting unit 201 that has received the operation result instructs the buffer 204 to write data (step S409). The buffer 204 receives the data to be written (step S410), retains the data, and returns a response indicating that the data has been successfully retained to the I / O accepting unit 201 (step S411). The I / O reception unit 201 returns the processing result to the I / O request unit 101 (step S412). Finally, the I / O request unit 101 receives the processing result (step S413).

  FIG. 5 is an activity diagram showing processing when the I / O request unit 101 issues a stop command in the storage system 10 shown in FIG. When the I / O request unit 101 issues a stop command (step S501), the I / O reception unit 201 receives the stop command (step S502) and determines the state of the disk enclosure 210 (step S503). In this example, since the disk enclosure 210 is already in power saving S2, no new processing is performed.

  The I / O reception unit 201 returns a result indicating that the disk enclosure 210 is already in power saving S2 to the I / O request unit 101 (step S504), and the I / O request unit 101 receives the result of the processing (step S504). Step S505).

  The stop command may be issued explicitly from the I / O request unit 101, or the I / O reception unit 201 detects that there is no I / O request for a predetermined time and performs a step spontaneously. You may make it perform the process of S503.

  As described above, the disk array 200 according to the first exemplary embodiment of the present invention does not have a return command issued explicitly from the server side. However, it is possible to operate the stopped physical disk again, judging from the read command (read) and the write command (write).

  Nevertheless, the disk array 200 buffers the I / O request and returns the I / O result to the I / O request unit 101 without waiting for the completion of the recovery of the physical disk when returning from the power saving S2. , Timeouts never occur.

  Further, as shown by a broken line S351 in FIG. 3, if steps S304 to S308 are omitted in the process of the read command, the read command does not return from the power saving S2, and the power saving S2 is performed only for the write command. Therefore, it is possible to maintain the power saving S2 for a longer time.

[Second Embodiment]
FIG. 6 is a conceptual diagram showing a configuration of the storage system 20 having two RAID1 configurations according to the second embodiment of the present invention. Since the storage system 20 has substantially the same configuration as the storage system 10 according to the first embodiment of the present invention, only the differences will be described here, and the same parts are denoted by the same reference numerals. The description will be omitted.

  The storage system 20 includes a first disk enclosure 610 (DE01) and a second disk enclosure 620 (DE02). The disk enclosure 610 (DE01) includes physical disks 611 (D01) and 612 (D02), and a cooling device 613. The disk enclosure 620 (DE02) includes physical disks 621 (D03) and 622 (D04), and a cooling device 623.

  The physical disks 611 (D01) and 621 (D03) constitute a RAID1 logical disk LD01 across the disk enclosures. The physical disks 612 (D02) and 622 (D04) constitute another RAID1 logical disk LD02 across the disk enclosures. The logical disks LD01 and LD02 can be accessed from the business server 100 via the I / O request unit 101.

  FIG. 7 is an activity diagram showing processing when the I / O request unit 101 issues a stop command in the storage system 20 shown in FIG. In FIG. 7 and the following description, the disk enclosure may be abbreviated as DE and the logical disk as LD. When the I / O request unit 101 issues a stop command (step S701), the I / O reception unit 201 receives the stop command (step S702), and determines the state of the disk enclosures 610 and 620 (step S703). In this example, it determines with it being an operation state.

  The control unit 203 instructs the management unit 202 to transition to the power saving state of the logical disk (step S704). In response to the instruction, the management unit 202 shifts the logical disk to the power saving state, and concentrates the physical disks in a specific disk enclosure from information such as RAID configuration information, logical disk status, physical disk status, and location. Information on the physical disk to be stopped is systematically returned so that it can be stopped (step S705). For example, in the power saving state, the management unit 202 holds in advance a condition such as always stopping the physical disk on the DE02 side, and determines a stop target physical disk based on this condition.

  Subsequently, the control unit 203 instructs the physical disk to stop (step S706), and the disk enclosures 610 and 620 stop the corresponding physical disk (step S707). Further, when the control unit 203 inquires of the management unit 202 about the disk enclosure in which all the physical disks are stopped (step S708), the management unit 202 determines that the disk enclosure in which all the physical disks are stopped (stop target DE). ) Information is returned (step S709).

  If there is a disk enclosure corresponding to the stop target DE, the control means 203 instructs the disk enclosure to stop the cooling device 613 or 623 (step S710), and the received disk enclosure (stop target DE) is cooled. The device 613 or 623 is stopped (step S711).

  The control unit 203 returns the result of the above stop process to the I / O accepting unit 201 (step S712), and the I / O accepting unit 201 further returns the result of the stop process to the I / O request unit 101. (Step S713), the I / O request unit 101 receives the result of the processing (Step S714).

  The processing when the I / O request unit 101 issues a read command (read) and a write command (write) in the storage system 20 has been described in the storage system 10 according to the first embodiment of the present invention. Since it can be the same as the processing, the description is omitted.

  By adopting this configuration, it is possible to manage the status of configuration information, physical disks, etc., and determine which physical disk should be stopped based on predetermined conditions. Furthermore, it is possible to stop the cooling device for the disk enclosure in which all the physical disks are stopped. Thereby, the power saving effect can be further enhanced.

  The processes shown in the activity diagrams of FIGS. 3 to 5 and 7 can be realized as a program executed by a computer included in the disk array 200.

  Although the present invention has been described with the specific embodiments shown in the drawings, the present invention is not limited to the embodiments shown in the drawings, and is known so far as long as the effects of the present invention are achieved. It goes without saying that any configuration can be adopted.

  The present invention can be applied to a storage system configured by a disk array that combines physical disks. In particular, it is suitable for a large-scale storage system and an application (for example, a data center, archive, etc.) that needs to suppress power consumption.

1 is a conceptual diagram showing a configuration of a storage system according to a first embodiment of the present invention. FIG. 2 is a state transition diagram for explaining an operation state of the disk enclosure disclosed in FIG. 1. FIG. 3 is an activity diagram showing processing when the I / O requesting unit issues a read command when the disk enclosure is power saving S2 in the storage system shown in FIG. FIG. 3 is an activity diagram showing processing when the I / O requesting unit issues a write command when the disk enclosure is in power saving S2 in the storage system shown in FIG. FIG. 2 is an activity diagram showing processing when an I / O request unit issues a stop command in the storage system shown in FIG. 1. It is a conceptual diagram which shows the structure of the storage system which has 2 RAID1 structure based on the 2nd Embodiment of this invention. FIG. 7 is an activity diagram showing processing when an I / O request unit issues a stop command in the storage system shown in FIG. 6.

Explanation of symbols

10, 20 Storage system 100 Business server (host device)
101 I / O request means 200 Disk array (disk array device)
201 I / O acceptance means 202 Management means 203 Control means 204 Buffer 210, 610, 620 Disk enclosure 211, 212, 213, 214, 611, 612, 621, 622 Physical disk 215, 613, 623 Cooling device S1 Full operation S2 Saving Power S3 returning

Claims (6)

Multiple physical disks that can be stopped or run individually, and
I / O accepting means for accepting a read command and a write command from the host device;
A buffer for temporarily storing data to be written to the physical disk by the write command;
When the stopped physical disk receives the write command, the host device controls the start of operation of the physical disk, stores the data to be written in the buffer, and returns a response indicating that the data has been successfully held. Control means for flushing the buffer after the physical disk is in operation and writing the data stored in the buffer to the physical disk ,
The control means receives data from data and parity stored in another physical disk operating without commanding the start of operation of the stopped physical disk when the read command is received from the host device. The disk array device is characterized in that the restored data is returned to the host device. A management unit that determines the physical disk to be stopped and started among the plurality of physical disks based on a condition held in advance;
2. The disk array device according to claim 1 , wherein the control unit stops or starts the operation of the physical disk determined by the management unit. The plurality of physical disks form a plurality of disk enclosures each including at least one or more physical disks and a cooling device;
If the control means is a disk enclosure all physical disks included in the disk enclosure is stopped, characterized by stopping the cooling device included in the disk enclosure, the disk array according to claim 1 apparatus. A physical disk recovery method for operating a stopped physical disk in a disk array device including a buffer for temporarily storing data and a plurality of physical disks that can be stopped or operated individually. And
A write command receiving step in which the disk array device receives a write command from a host device;
A buffering step of storing in the buffer data to be written to the physical disk;
An operation start step of starting the operation of the stopped physical disk and returning a response indicating that the data to be written to the physical disk has been successfully held to the host device;
A flushing step of flushing the buffer after the physical disk is in operation;
A writing step of writing the data stored in the buffer to the physical disk ;
A read command receiving step in which the disk array device receives a read command from the host device;
A data restoration step of restoring data from data and parity stored in another physical disk that is operating without starting the operation of the physical disk that is stopped;
A physical disk recovery method comprising: a data return step of returning the restored data to the host device . The disk enclosure is a plurality of disk enclosures each including at least one physical disk and a cooling device;
5. The physical unit according to claim 4 , further comprising: a cooling stop step of stopping a cooling device included in the disk enclosure if there is a disk enclosure in which all of the physical disks included in the disk enclosure are stopped. Disk recovery method. In a disk array device including a buffer for temporarily storing data and a plurality of physical disks that can be individually stopped or operated, a computer included in the disk array device includes:
Processing to store in the buffer data to be written to the physical disk in response to a write command from a host device;
A process of starting the operation of the stopped physical disk and returning a response indicating that the data to be written to the physical disk has been successfully held to the higher-level device;
A process of flushing the buffer after the physical disk is in operation;
A process of writing the data stored in the buffer to the physical disk ;
In response to a read command from the host device, the operation of restoring the data from the data and parity stored in the other physical disks that are operating without starting the operation of the stopped physical disk;
A physical disk recovery program that executes a process of returning the restored data to the host device .

Images