JP2009187406A - Disk array device, data switch back method, and data switch back program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a disk array device capable of suppressing a deterioration in availability without having an excessive effect on an I/O. <P>SOLUTION: The disk array device has a plurality of disk enclosures 14 for housing a data disk so as not to include a plurality of data disks, a first disk 15 which is normally a spare disk, and when either of the data disks fails, restores data recorded in the failed data disk to inherit a function of the failed data disk, a second disk 15 restored as a spare disk by exchanging the failed data disk, and a control part 12 for copying data restored in the first disk to the second disk if either the first disk or a data disk is included in the same disk enclosure. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a disk array device configured by RAID, and more particularly to processing after data is restored in the disk array device.

  In a disk array device composed of a nonvolatile medium such as a magnetic disk device or an optical disk device, in addition to a plurality of data disks that configure RAID and store user data, as an alternative when a data disk fails Usually, a spare disk to be used is mounted.

  Many disk array apparatuses have a plurality of disk enclosures that are physical aggregates of a plurality of disk apparatuses including data disks and spare disks. The supply of power to the disk device or the operation of the cooling device is controlled for each disk enclosure. In such a case, if a RAID is configured so that a plurality of data disks straddle a plurality of disk enclosures, the availability is improved, and further, an energy saving effect is obtained by controlling power supply and cooling devices in units of disk enclosures. You can also.

  When the data disk fails and the data of the failed data disk is restored, the spare disk takes over the function of the failed data disk and is included in the RAID configuration. When the failed data disk is removed and a new disk device is installed instead, the new disk device replaced and replaced becomes a spare disk.

  At that time, the recorded data is copied from the spare disk that is operating as a data disk after the data is restored to the new disk device, the copy source disk is returned to the original spare disk, and the new disk device is restored. May function as a data disk. This is called data switchback. In general, the disk array device can set whether or not to switch back this data when the data is restored. The setting for performing data switching back is referred to as “data switching back mode”, and the setting for not performing data switching back is referred to as “no data switching back mode”.

  The following patent literatures are available as technologies related to the disk array device. Japanese Patent Application Laid-Open No. 2004-228561 discloses a technique of setting a RAID group from a disk having a large remaining capacity to reduce the remaining capacity unevenness. Patent Document 2 discloses a technique for reducing the data restoration time.

  Patent Document 3 discloses a technique in which a plurality of identical logical volumes are arranged in different physical disk units in a RAID device. Patent Document 4 discloses a technique for restoring data of a failed disk device to a spare disk device using a disk array, and restoring the spare disk device data to the replaced disk device when the failed disk device is replaced. .

JP 2006-146680 A JP 2007-087039 A Japanese Patent Laid-Open No. 10-133826 JP-A-11-184463

  When the disk array device is in data switchback mode, the position of the spare disk is fixed, but every time a data disk failure occurs, the entire data for one disk device is copied, resulting in a large amount of data transmission Will do. For this reason, there is a risk of excessively affecting the I / O connecting the disk device.

  On the other hand, when the disk array device is in the no data switchback mode, a large amount of data transmission associated with such copying does not occur, so there is no excessive influence on I / O. However, in this mode, when a failed data disk is replaced, the new disk device replaced and installed there becomes a spare disk as it is, so that the position of the spare disk changes each time a data disk failure occurs. become.

  Therefore, when a RAID is configured such that a plurality of data disks straddle a plurality of disk enclosures as described above, the position of the spare disk changes due to a failure of the data disk, so that the plurality of data disks are the same. It may be included in the disk enclosure. In this case, for example, there is a risk that a plurality of data disks included in the same disk enclosure fail simultaneously due to a trouble such as a power supply system of the disk enclosure. Therefore, availability is reduced.

  On the other hand, among the above-mentioned patent documents, Patent Document 4 discloses a technique for switching back the above-mentioned data. However, there is no description of a configuration that solves the above-described problem that occurs when data is switched back. Of course, Patent Documents 1 to 3 do not describe a configuration for solving the problem.

  An object of the present invention is to provide a disk array device, a data switchback method, and a data switchback program that do not exert an excessive influence on I / O when a data disk failure occurs and can suppress a decrease in availability. It is to provide.

  In order to achieve the above object, a disk array device according to the present invention is a disk array device including a plurality of data disks constituting a RAID, and usually includes a plurality of disk enclosures that contain a plurality of data disks. A spare disk, and when one of the data disks fails, the data recorded on the failed data disk is restored, and the first disk takes over the function of the failed data disk, and the failed data disk If the second disk installed as a spare disk and any one of the first disk and the data disk are included in the same disk enclosure, the data restored to the first disk And a controller for copying to the second disk.

  In order to achieve the above object, a data switch-back method according to the present invention includes a plurality of data disks constituting a RAID, a plurality of disk enclosures accommodated so as not to include a plurality of data disks, and normally a spare disk. A disk array apparatus having a first disk, and a method for switching back data, wherein data recorded on a failed data disk when one of the data disks fails A data restoration process that restores the disk and takes over the function of the failed data disk; a spare disk installation process that uses the second disk replaced from the failed data disk as a spare disk; and the first disk and the data disk To determine whether any of the above are included in the same disk enclosure If the determination process determines that either the first disk or the data disk is included in the same disk enclosure, the data restored to the first disk is copied to the second disk. And a data copying process.

  In order to achieve the above object, a data switch-back program according to the present invention is a plurality of data disks constituting a RAID, a plurality of disk enclosures accommodated so as not to include a plurality of data disks, and normally a spare disk. A computer that controls a disk array device having a first disk restores the data recorded on the failed data disk to the first disk when one of the data disks fails, and the failed data disk Data restoration processing that takes over the above functions, spare disk installation processing in which the second disk replaced from the failed data disk is a spare disk, and either the first disk or the data disk is in the same disk enclosure A judgment process for judging whether or not it is included, and a judgment process If it is determined that one of the first disk and the data disk is included in the same disk enclosure, a data copying process for copying the data restored to the first disk to the second disk; Is executed.

  The present invention determines whether or not a plurality of data disks are included in the same disk enclosure, and does not cause an unnecessarily large amount of data transmission while maintaining availability. As a result, an unprecedented superior disk array device, data switching method, and data that do not have an excessive impact on I / O when a data disk failure occurs and can suppress a decrease in availability can be suppressed. A switchback program can be provided.

  FIG. 1 is a functional block diagram showing a configuration of a disk array 10 according to an embodiment of the present invention. The disk array 10 includes an I / O reception unit 11, a control unit 12, a buffer 13, a management table 16, and a plurality of disk enclosures 14. Each disk enclosure 14 includes a plurality of physical drives 15.

  The disk array 10 is connected to the host computer 20 and reads / writes data according to I / O requests such as a read command (read), a write command (write), and a stop command from the host computer 20. At that time, any interface such as Fiber Channel (FC) or Ultra SCSI can be applied to the connection between the disk array 10 and the host computer 20.

  The I / O accepting unit 11 accepts an I / O request from the host computer 20, causes the control unit 12 to perform operations such as reading and writing data to each disk enclosure 14 based on the I / O request, and results of execution (I / O result) is returned to the host computer 20. The buffer 13 holds an unprocessed I / O request and an I / O result. Further, the control unit 12 detects the presence or absence of a failure in the physical drive 15, and when a failure is detected, performs a data restoration operation described later.

  In the disk array 10, a plurality of physical drives 15 form RAID1 or RAID5. In this case, each physical drive 15 is included in a separate disk enclosure 14.

  The control unit 12 stores the RAID configuration configured by each physical drive 15 in the management table 16. This RAID configuration includes information on which disk enclosure 14 each physical drive 15 constituting the RAID is included. Information about whether or not the RAID configuration is in the data switchback mode is also included in the RAID configuration.

  When the data restoration operation performed based on the RAID standard and the data switchback operation of the present invention are performed, the RAID configuration is changed. In this case, the control unit 12 causes the management table 16 to change. Rewrite data stored in.

  FIG. 2 is a flowchart showing the data restoration operation performed by the control unit 12 when a data disk fails in the disk array 10 of FIG. If the control unit 12 is controlled by a computer, the flowchart of FIG. 2 can be configured as a program executed by the computer.

  When the operation is started and the control unit 12 detects a data disk failure (step S101), the control unit 12 first refers to the RAID configuration of the management table 16 to determine whether there is a spare disk corresponding to the failed data disk. Is detected (step S102). If there is no spare disk, the process is terminated as it is and no data restoration operation is performed. If there is a spare disk, the process proceeds to step S103.

  If there is a spare disk in step S102, the control unit 12 restores the data of the failed data disk on the spare disk with the data stored in the other data disk that has not failed (step S103). This operation is performed based on the RAID standard. Then, the control unit 12 refers to the management table 16 and determines whether or not the RAID configuration is in the data switchback mode (step S104). If it is the data switchback mode, the process proceeds to step S106, and the control unit 12 executes the data switchback.

  If it is not the data switchback mode in step S104, the control unit 12 restores the data on the spare disk, and whether or not a plurality of data disks included in the same RAID exists in one disk enclosure. Is determined (step S105). If there is no disk enclosure in which a plurality of data disks exist, the process ends as it is.

  If there is a disk enclosure containing a plurality of data disks, the process proceeds to step S106, and after confirming that the failed data disk has been replaced and restored as a spare disk, the control unit 12 performs data switchback. The data switch-back here is a process of copying the recording data of the copy source data disk to the copy destination spare disk, and changing the spare disk to the data disk and the data disk to the spare disk. When the data switching back in step S106 is completed, the processing shown in the flowchart of FIG.

  Hereinafter, first to fourth examples of the data restoration process shown in FIG. 2 will be described. All data disks and spare disks shown in the following examples are the physical drives 15 in FIG. Similarly, all the disk enclosures shown in the following examples are the disk enclosures 14 in FIG. Description of the I / O reception unit 11, the control unit 12, the buffer 13, the management table 16, and the host computer 20 in FIG. 1 is omitted in the following examples.

  Each disk enclosure 14 is called by the symbols DE0, DE1, DE2,. Further, the physical drive 15 included in each disk enclosure is simply referred to as a data disk or a spare disk depending on whether the physical drive 15 is operating as a data disk or a spare disk. When the physical drive 15 is operating as a data disk, it is called by symbols D0, D1, D2,..., And when the physical drive 15 is operating as a spare disk, it is called by symbols S0, S1, S2,.

  Therefore, the same physical drive 15 may be a data disk or a spare disk depending on the operating state. Further, symbols such as “D0”, “D1”, and “S0” are also changed accordingly. For this reason, for example, when the data disk with the symbol “D0” is changed to the spare disk with the symbol “S0”, the data disk 201 (D0) becomes the spare disk 201 (S0). In this way, the same device is assigned the same reference number regardless of whether the disk is a data disk or a spare disk.

  FIG. 3 is a conceptual diagram showing a first example of the data restoration processing shown in FIG. As shown in FIG. 3A, in the disk array 10, the data disk 201 (D0) is mounted in the disk enclosure 211 (DE0), and the data disk 202 (D1) is mounted in the disk enclosure 212 (DE1). A data disk 201 (D0) and a data disk 202 (D1) mounted in different disk enclosures constitute RAID1. The spare disk 203 (S0) is mounted in the disk enclosure 212 (DE1).

  Now, it is assumed that the data disk 201 (D0) has failed. In this case, as shown in FIG. 3B, the data of the failed data disk 201 (D0) is restored on the spare disk 203 (S0) by the data stored in the data disk 202 (D1). Then, based on the alarm from the control unit 12, the maintenance person replaces the failed data disk 201 (D 0) with a new physical disk 204 and installs it.

  As shown in FIG. 3C, the spare disk 203 (S0) whose data has been restored on the disk enclosure 212 (DE1) operates as the data disk 203 (D0). Then, the new physical disk 204 exchanged on the disk enclosure 211 (DE0) operates as a spare disk 204 (S0). However, in this state, the data disk 203 (D0) and the data disk 202 (D1) are mounted in the same disk enclosure 212 (DE1).

  In this case, the control unit 12 switches back the recording data of the data disk 203 (D0) on the disk enclosure 212 (DE1) to the spare disk 204 (S0) on the disk enclosure 211 (DE0). As a result, as shown in FIG. 3D, the spare disk 204 (S0) from which data has been cut back becomes a new data disk 204 (D0), and the data disk 203 (D0) becomes the spare disk 203 (S0). Become. As a result, each data disk constituting RAID 1 is originally mounted in a separate disk enclosure.

  FIG. 4 is a conceptual diagram showing a second example of the data restoration processing shown in FIG. As shown in FIG. 4A, in the disk array 10, the data disk 221 (D0) is mounted in the disk enclosure 231 (DE0), and the data disk 222 (D1) is mounted in the disk enclosure 232 (DE1). A data disk 221 (D0) and a data disk 222 (D1) mounted in different disk enclosures constitute RAID1. The spare disk 223 (S0) is mounted in the disk enclosure 233 (DE2).

  Now, it is assumed that the data disk 221 (D0) has failed. In this case, as shown in FIG. 4B, the data of the failed data disk 221 (D0) is restored on the spare disk 223 (S0) by the data stored in the data disk 222 (D1). Then, based on the alarm from the control unit 12, the maintenance disk replaces the failed data disk 221 (D 0) with a new physical disk 224 and installs it.

  As shown in FIG. 4C, the spare disk 223 (S0) whose data has been restored on the disk enclosure 233 (DE2) operates as the data disk 223 (D0). Then, the new physical disk 224 exchanged on the disk enclosure 231 (DE0) operates as a spare disk 224 (S0).

  In this state, the data disk 223 (D0) is mounted on the disk enclosure 233 (DE2), and the data disk 222 (D1) is mounted on the disk enclosure 232 (DE1), both of which are mounted on different disk enclosures. Therefore, unless the RAID configuration recorded in the management table 16 is in the data switchback mode, data switchback is not performed.

  5 to 6 are conceptual diagrams showing a third example of the data restoration processing shown in FIG. As shown in FIG. 5A, in the disk array 10, the data disk 241 (D0) is the disk enclosure 251 (DE0), the data disk 242 (D1) is the disk enclosure 252 (DE1), and the data disk 243 (D2). ) Is mounted in the disk enclosure 253 (DE2), the data disk 244 (D3) is mounted in the disk enclosure 254 (DE3), and the data disk 245 (D4) is mounted in the disk enclosure 255 (DE4). Data disks 241 (D0) to 245 (D4) mounted across the disk enclosures constitute RAID5. The spare disk 246 (S0) is mounted in the disk enclosure 255 (DE4).

  Now, it is assumed that the data disk 244 (D3) has failed. In this case, as shown in FIG. 5B, the data stored in the data disks 241 (D0) to 245 (D4) excluding the data disk 244 (D3) is used to store the data on the failed data disk 244 (D3). Is restored on the spare disk 246 (S0). Then, based on the alarm from the control unit 12, the data disk 244 (D3) in which the maintenance staff has failed is replaced with a new physical disk 247 and installed.

  As shown in FIG. 6C, the spare disk 246 (S0) whose data has been restored on the disk enclosure 255 (DE4) operates as the data disk 246 (D3). Then, the new physical disk 247 exchanged on the disk enclosure 254 (DE3) operates as a spare disk 247 (S0). However, in this state, the data disk 245 (D4) and the data disk 246 (D3) are mounted in the same disk enclosure 255 (DE4).

  In this case, the control unit 12 switches back the recording data of the data disk 246 (D3) on the disk enclosure 255 (DE4) to the spare disk 247 (S0) on the disk enclosure 254 (DE3). As a result, as shown in FIG. 5D, the spare disk 247 (S0) from which data has been cut back becomes a new data disk 247 (D3), and the data disk 246 (D3) becomes the spare disk 246 (S0). Become. As a result, each data disk constituting the RAID 5 is mounted across the disk enclosure as it was originally.

  7 to 8 are conceptual diagrams showing a fourth example of the data restoration processing shown in FIG. As shown in FIG. 7A, in the disk array 10, the data disk 261 (D0) is the disk enclosure 271 (DE0), the data disk 262 (D1) is the disk enclosure 272 (DE1), and the data disk 263 (D2). ) Is mounted on the disk enclosure 273 (DE2), the data disk 264 (D3) is mounted on the disk enclosure 274 (DE3), and the data disk 265 (D4) is mounted on the disk enclosure 275 (DE4). Data disks 261 (D0) to 265 (D4) mounted across the disk enclosures constitute RAID5. The spare disk 266 (S0) is mounted in the disk enclosure 276 (DE5).

  Now, it is assumed that the data disk 264 (D3) has failed. In this case, as shown in FIG. 7 (B), the data stored in the data disks 261 (D0) to 265 (D4) except for the data disk 264 (D3) is stored in the data of the failed data disk 264 (D3). Is restored on the spare disk 266 (S0). Then, based on the alarm from the control unit 12, the data disk 264 (D3) in which the maintenance staff has failed is replaced with a new physical disk 267 and installed.

  As shown in FIG. 8C, the spare disk 266 (S0) whose data has been restored on the disk enclosure 276 (DE5) operates as the data disk 266 (D3). Then, the new physical disk 267 exchanged on the disk enclosure 274 (DE3) operates as a spare disk 267 (S0).

  In this state, the data disks 261 (D0) to 263 (D2) and 265 (D4) and 266 (D3) constituting the RAID 5 are mounted in different disk enclosures. Data is not switched back unless in the return mode.

  As described above, the disk array 10 according to the present embodiment determines whether or not a plurality of data disks are included in the same disk enclosure unless the data switchback mode is set. The data is switched back only when it is included. Accordingly, it is possible to suppress a decrease in availability due to a plurality of data disks being included in the same disk enclosure.

  In addition, when a plurality of data disks are not included in the same disk enclosure, it can be considered that there is no reduction in availability, and therefore data is not switched back. As a result, an unnecessarily large amount of data transmission is not generated, so that an excessive influence on I / O can be suppressed.

  RAID includes RAID types such as RAID1, RAID3, RAID4, RAID5, RAID6, RAID10, RAID30, RAID50, triple mirror, DP, etc. The processing described in this embodiment is applicable to any RAID type. It is possible to apply. Further, even if the number of disk enclosures and physical disks increases, the processing described in this embodiment can be applied.

  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 used in a disk array apparatus including a plurality of data disks constituting a RAID.

It is a functional block diagram which shows the structure of the disk array which concerns on embodiment of this invention. 2 is a flowchart showing a data restoration operation performed by a control unit when a data disk fails in the disk array of FIG. It is a conceptual diagram which shows the 1st example of the data restoration process shown in FIG. It is a conceptual diagram which shows the 2nd example of the data restoration process shown in FIG. It is a conceptual diagram which shows the 3rd example of the data restoration process shown in FIG. It is a continuation of FIG. It is a conceptual diagram which shows the 4th example of the data restoration process shown in FIG. FIG. 7 is a continuation of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Disk array 11 I / O reception part 12 Control part 13 Buffer 14 Disk enclosure 15 Physical drive 16 Management table 20 Host computer

Claims (10)

A disk array device including a plurality of data disks constituting a RAID,
A plurality of disk enclosures for accommodating a plurality of the data disks; and
A first disk that is normally a spare disk and takes over the function of the failed data disk by restoring the data recorded on the failed data disk when any of the data disks fails;
A second disk in which the failed data disk is replaced and installed as a spare disk;
A controller that copies data restored to the first disk to the second disk if either of the first disk and the data disk are included in the same disk enclosure; A disk array device comprising:   The control unit copies the data restored to the first disk to the second disk, and then uses the second disk as a data disk that takes over the function of the failed data disk. 2. The disk array device according to claim 1, wherein the disk array device is returned to a spare disk.   The management unit that stores a RAID configuration constituted by the plurality of data disks and that can be rewritten by the control unit is provided in the control unit. Disk array device. The management table includes information about whether the RAID is in a data switchback mode, and
If the RAID is in the data switchback mode, the controller restores to the first disk even if neither the first disk nor the data disk is included in the same disk enclosure. 4. The disk array device according to claim 3, wherein the copied data is copied to the second disk.   The disk according to claim 2, wherein the RAID type of the RAID includes at least one of RAID1, RAID3, RAID4, RAID5, RAID6, RAID10, RAID30, RAID50, triple mirror, and DP. Array device. A disk array device having a plurality of data disks constituting a RAID, a plurality of disk enclosures for accommodating a plurality of data disks, and a first disk that is normally a spare disk, A method for performing switchback,
A data restoration step of restoring the data recorded on the failed data disk to the first disk when any of the data disks fails, and taking over the function of the failed data disk;
A spare disk installation step in which the second disk replaced from the failed data disk is a spare disk;
A determination step of determining whether one of the first disk and the data disk is included in the same disk enclosure;
If it is determined in the determination step that any one of the first disk and the data disk is included in the same disk enclosure, the data restored to the first disk is stored in the second disk. And a data copy method for copying to a disk.   A function handover step is provided, wherein after the data copying step, the second disk is a data disk that takes over the function of the failed data disk, and the first disk is returned to a spare disk. 6. The data switchback method according to 6. The determining step includes the step of confirming whether the RAID is in a data switchback mode;
If the RAID is in the data switchback mode, the data copying step is performed even if the first disk and the data disk are not included in the same disk enclosure. 8. The data switchback method according to claim 7, wherein the data restored in step (b) is copied to the second disk. A computer that controls a disk array device having a plurality of data disks constituting a RAID, a plurality of disk enclosures that contain the data disks so as not to include a plurality of data disks, and a first disk that is normally a spare disk,
A data restoration process that restores the data recorded on the failed data disk to the first disk and takes over the function of the failed data disk when any of the data disks fails;
A spare disk installation process in which the second disk replaced from the failed data disk is a spare disk;
Determination processing for determining whether one of the first disk and the data disk is included in the same disk enclosure;
If it is determined in the determination process that one of the first disk and the data disk is included in the same disk enclosure, the data restored to the first disk is stored in the second disk. A data switch-back program for executing a data copy process for copying to a disk.   The function takeover process for returning the first disk to a spare disk is executed after the data copying process, wherein the second disk is a data disk that takes over the function of the failed data disk. 9. The data switchback program according to 9.

Images