JP4261532B2 - Logical disk management method and virtualization apparatus - Google Patents

Description

  The present invention relates to a logical disk management method and a virtualization apparatus that manage a logical disk that is recognized as a single disk area from a host computer and that is configured using a storage area of at least one disk drive.

  Generally, a disk array device includes a plurality of disk drives, for example, hard disk drives (HDD), and an array controller connected to these HDDs. The array controller manages a plurality of HDDs by a commonly known RAID (Redundant Arrays of Independent Disks, Redundant Arrays of Inexpensive Disks) method. In other words, in response to a data read / write request from a host (host computer), the array controller executes data read / write in a distributed manner by moving a plurality of HDDs connected to the controller in parallel. As a result, in the disk array device, it is possible to increase the speed of data access requested from the host. In the disk array device, the reliability can be improved by the redundant configuration of the disks.

  In the conventional disk array device, the physical arrangement of logical disk areas recognized by the host is fixed. For this reason, in the conventional disk array device, the relationship between the block address of the logical disk and the array configuration corresponding thereto does not basically change. Similarly, the relationship between the block address of the logical disk and the corresponding block address of the HDD does not basically change.

  On the other hand, in the disk array device, as a result of starting operation of the disk array device, the access load to the logical disk may differ from the original plan. In addition, the access load may change with time. In such a case, in the conventional disk array device, even if a bottleneck or a hot spot occurs in the array or HDD constituting the logical disk, it is not easy to eliminate it. This is because the correspondence between the logical disk and the array and the correspondence between the logical disk and the HDD are fixed. In order to eliminate the occurrence of bottlenecks and hot spots, it is necessary to back up the data stored in the logical disk to, for example, a tape and then recreate the logical disk. In addition, it is necessary to restore data from the tape to the re-created logical disk. Note that a hot spot means a state in which an access load is concentrated in a certain area of the HDD.

  Recently, a plurality of hosts often share a disk array device. In such a case, an access load may change due to an increase in the number of hosts connected to the disk array device, and a bottleneck or a hot spot may occur.

  However, in the conventional disk array device, the physical arrangement of logical disk areas is fixed. For this reason, once the operation is started, it is not easy to cope with a change in access load.

For this reason, for example, Patent Document 1 uses the value indicating the performance (I / O performance) of input / output (I / O) processing of each HDD (physical disk) to optimize the I / O performance of each physical disk. A technique for relocating a logical disk (hereinafter referred to as prior art) is disclosed. In this prior art, the busy rate of each HDD is brought close to the optimum busy rate.
JP 2003-5920 (paragraphs 0012, 0050 to 0065, FIGS. 6 and 7)

  In the above-described prior art, the access load on the entire logical disk can be reduced by rearranging the logical disks. However, in the prior art, the logical disk is rearranged in units of logical disks. For this reason, when a bottleneck or hot spot occurs in an array or HDD constituting a logical disk, it is not easy to eliminate the occurrence of the bottleneck or hot spot.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a logical disk management method and a virtualization apparatus capable of improving access performance.

  According to one aspect of the present invention, there is provided a logical disk management method that manages a logical disk that is configured using a storage area of at least one disk drive and that is recognized as a single disk volume by a host computer. Is done. This method is an array composed of a set of slices defined by defining a storage area of the at least one disk drive as one physical array area, and the physical array area is constant. A logical disk is formed by combining a step of forming an array in which the divided area is defined as the slice and a plurality of slices included in the array. On the basis of the step of configuring, the step of acquiring the statistical information relating to the access processing for the slice for each slice constituting the logical disk and storing it in the storage means, and the statistical information acquired for each slice Detecting a slice having a read access load higher than a certain level and detecting the detected slice A second area for storing a copy of the data in the first area in the array in which the first and second disks are arranged in a different array from the array; and from the host computer, the first of the logical disk When it is requested to read the data of the slice included in the area, the data from the corresponding slice of either the first area in the array or the second area in the other array And when the host computer is requested to write data to a slice included in the first area of the logical disk, the first area in the array and the other area Writing the same data to each corresponding slice of the second region in the array.

  According to the present invention, the physical arrangement of the logical disk area is virtually managed, and the physical arrangement of the logical disk area can be partially flexibly changed during operation of the apparatus. The performance can be improved.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of a computer system provided with a disk array device according to an embodiment of the present invention. This computer system includes a disk array device 10 and a host (host computer) 20. The host 20 is connected to the disk array device 10 by a host interface HI such as a small computer system interface (SCSI) or a fiber channel. The host 20 uses the disk array device 10 as an external storage device.

  The disk array device 10 includes at least one array (physical array), for example, four arrays 11a (#a), 11b (#b), 11c (#c), and 11d (#d) as shown in FIG. Has a group of arrays. However, in FIG. 1, the group of arrays is omitted for the sake of drawing. Each array is configured by defining a storage area of at least one HDD (hard disk drive) as a physical area (array area) of the array. In this embodiment, each array is configured by defining a storage area of a plurality of HDDs as a physical array area of the array.

  The disk array device 10 also has at least one array controller, for example, duplicated array controllers 12-1 and 12-2. The disk array device 10 further includes a semiconductor disk device 131 and an HDD group including HDDs (hard disk drives) 132A (#A), 132B (#B), 132C (#C), and 132D (#D).

  The semiconductor disk device 131 is a storage device generally called a silicon disk drive, such as a battery backup type RAM disk device. That is, the semiconductor disk device 131 is a storage device configured using a plurality of memory elements such as a dynamic RAM (DRAM). The semiconductor disk device 131 is designed so that the same access method (interface) as that for the HDD can be applied to access the device 131 from the host. Since the semiconductor disk device 131 is configured using a memory element, it is generally very expensive and has a small capacity compared to an HDD, but can be accessed at a very high speed.

  The HDDs 132A and 132B are low-cost but inexpensive and large-capacity HDDs, and are used to construct an array. The HDDs 132C and 132D are high-performance and low-capacity HDDs that have higher performance than the HDDs 132A and 132B, and are used to construct an array. The HDDs 132A, 132B, 132C, and 132D are connected to the array controllers 12-1 and 12-2 through the storage interface SI such as SCSI or fiber channel together with the semiconductor disk device 131. In response to a data read / write request from the host 20, the array controllers 12-1 and 12-2 operate a plurality of HDDs constituting the array in parallel to distribute and execute data read / write. The array controllers 12-1 and 12-2 are synchronized so as to always be in the same state by communicating with each other.

  The array controllers 12-1 and 12-2 have virtualization units 120-1 and 120-2, respectively. The virtualization units 120-1 and 120-2 combine at least one arbitrary slice in an arbitrary array to constitute at least one logical disk recognized by the host 20. The slice will be described later. The virtualization unit 120-1 includes a logical disk configuration unit 121 and a memory 122. The logical disk configuration unit 121 includes an array / slice definition unit 121a, a logical disk definition unit 121b, a slice migration unit 121c, a data read / write unit 121d, a statistical information acquisition unit 121e, and a detection unit 121f. A part of the storage area of the memory 122 is used as a map table area for storing the map table 122a. Another part of the storage area of the memory 122 is used as a statistical information area (statistical information storage means) 122b for storing I / O statistical information, which will be described later, acquired by the statistical information acquisition unit 121e. For example, a RAM is used as the memory 122. Although omitted in FIG. 1, the virtualization unit 120-2 has the same configuration as the virtualization unit 120-1.

  The logical disk configuration unit 121 is realized by a processor (not shown) in the array controller 12-1 reading and executing a specific software program installed in the controller 12-1. This program can be stored in advance in a computer-readable storage medium and distributed. Further, this program may be downloaded (distributed) via a network.

  The array / slice definition unit 121a defines an array and a slice. The array / slice definition by the array / slice definition unit 121a will be described with reference to FIGS. 2 (a) and 2 (b).

  The array / slice definition unit 121a defines at least one group, for example, a plurality of groups. Each defined group includes at least one HDD, eg, a plurality of HDDs. The array / slice definition unit 121a defines an array for each defined group. Here, each array is defined (managed) as an array in which storage areas of a plurality of HDDs in a group corresponding to the array are used as their own physical areas (array areas) by a RAID method.

  First, FIG. 2A shows the array 11a (#a). This array 11a is composed of a set 21a of four HDDs in the HDD group including the HDDs 132A, 132B, 132C, and 132D, and is an array managed at the RAID 1 + 0 level. On the other hand, FIG. 2B shows the array 11b (#b). This array 11b is composed of a set 21b of five HDDs in the HDD group, and is an array managed at the RAID5 level. Here, in order to simplify the explanation, it is assumed that there is no HDD common to the two groups constituting the arrays 11a and 11b. In this case, the capacities of physical storage areas (array areas) of the arrays 11a and 11b are equal to the total capacity of four HDDs and the total capacity of five HDDs, respectively.

  The array / slice definition unit 121a divides the storage area of each array including the arrays 11a and 11b with a certain fixed capacity (for example, 1 GB) in the disk array device 10. The array / slice definition unit 121a defines each divided area having a certain capacity as a slice. That is, the array / slice definition unit 121a divides the storage area of each array into a plurality of slices having a certain capacity. Thereby, every slice of any array in the disk array device 10 has the same capacity. This point is important for the movement of the slice by the slice moving unit 121c described later. Here, each slice included in each array is assigned a number (slice number) as an ID (identification information) of the slice in ascending order of addresses in the array. That is, the slice number assigned to each slice in the array also represents the position (physical position) of the slice in the array.

  The logical disk definition unit 121b defines a logical disk that is recognized by the host 20 as a single disk (disk volume). The definition of the logical disk by the logical disk definition unit 121b will be described with reference to FIG. The logical disk definition unit 121b connects (combines) a plurality of arbitrary slices included in at least one arbitrary array. The logical disk definition unit 121b defines a logical disk in which an area in which a plurality of arbitrary slices are connected (combined) is managed as a logical storage area. In FIG. 3, four arrays 11a, 11b, 11c and 11d are shown. In the example of FIG. 3, a slice group including a slice # a0 in the array 11a, a slice # c0 in the array 11c, a slice # a1 in the array 11a, and a slice # d0 in the array 11d is combined ( A logical disk 31-0 (# 0) is defined by concatenation. Similarly, a logical disk 31-1 is formed by combining slice groups including the slice # a2 in the array 11a, the slice # b0 in the array 11b, the slice # b1 in the array 11b, and the slice # c1 in the array 11c. (# 1) is defined.

  As described above, the storage area of the logical disk is physically discontinuously allocated in units of slices. The capacity of the logical disk is 1 slice capacity × slice number. A logical disk is a unit that is finally recognized as a single disk volume from the host 20. From the host 20, this logical disk is recognized as if it were one HDD. Slice numbers are assigned to each slice assigned to a logical disk in ascending order of addresses (logical addresses) in the logical disk. That is, each slice allocated to a logical disk is managed by two slice numbers, a slice number indicating a position (logical position) in the logical disk and a slice number indicating a position (physical position) in the corresponding array. .

  The map table 122a is storage means for storing information (map information) for associating the logical disk with the array. An example of the data structure of the map table 122a is shown in FIG. In the example of FIG. 4, slice information is stored in the row direction of the map table 122a in order from the slice assigned to the smaller address in the logical disk. In the present embodiment, the information of each slice included in the logical disk includes information stored in the fields (items) 41 to 48. The field 41 is used to store a logical disk number as an ID (identification information) of a logical disk to which the corresponding slice is allocated. The field 42 is used to store the slice number in the logical disk of the corresponding slice. The field 43 is used to store the array number as the ID of the array to which the corresponding slice belongs, and the field 44 is used to store the slice number of the slice in the array. The field 45 is used to store a copy flag indicating whether the data of the corresponding slice is being copied to another slice. The field 46 is used to store an array number indicating a copy destination array of the corresponding slice. The field 47 is used to store the slice number of the copy destination slice in the copy destination array of the data of the corresponding slice. Used to store the slice number of the destination slice. The field 48 is used to store size information indicating the size of data that has already been copied in the corresponding slice. The map table 122a does not store position information relating the position of each slice in the array and the position of the slice in the corresponding HDD. The reason is that the position of each slice in the array in the corresponding HDD is uniquely calculated from the slice number of the slice (slice number indicating the position in the array) and the size of the slice. Of course, the position information can be stored in the map table 122a.

  The slice moving unit 121c moves data of an arbitrary slice in the logical disk. This movement is performed as follows. The slice moving unit 121c copies the data of an arbitrary slice (first slice) in an arbitrary logical disk to another slice (second slice) that has not been entered (assigned) in the logical disk. To do. Next, the slice moving unit 121c swaps both slices. That is, the slice moving unit 121c switches the former slice (first slice) to a state where it is not entered in the logical disk (an unused state). Further, the slice moving unit 121c switches the latter slice (second slice) to the state entered (allocated) in the logical disk.

  As described above, in this embodiment, the logical disk can be easily reconfigured by simply replacing the slices entered (allocated) in the logical disk. For this reason, even after the operation is started, it is possible to easily cope with a change in the access load without stopping the logical disk (that is, online) and to improve the access performance.

  Details of the slice movement by the slice movement unit 121c will be described with reference to the map table 122a of FIG. In this example, it is assumed that the slice with slice number 3 included in the logical disk with logical disk number 0 is the target of movement. The slice with slice number 3 corresponds to the slice with slice number 10 included in the array with array number 2. The data of the slice with the slice number 3 included in the logical disk with the logical disk number 0 is included in the array with the array number 1 from the slice with the slice number 10 included in the array with the array number 2. It is assumed that the data is copied to the slice with slice number 5. The position at which the data copy is completed in the slice of slice number 5 is indicated by the size information stored in the field 48.

  When the copying of all data stored in the slice with slice number 3 is completed, the slice moving unit 121c switches the copy source and copy destination slices. That is, the slice moving unit 121c includes the slice with the slice number 3 included in the logical disk with the logical disk number 0 from the slice with the slice number 10 included in the array with the array number 2 in the array with the array number 1. Change to the slice with slice number 5. As a result, the physical allocation of the slice of slice number 3 included in the logical disk of logical disk number 0 is transferred from the slice of slice number 10 included in the array of array number 2 to the array of array number 1. It has been moved (changed) to the slice with the included slice number 5. When the copy is completed, the copy flag is cleared ("0" is cleared), and the array number indicating the copy destination array and the slice number indicating the copy destination slice in the copy destination array are also cleared ("0" clear). Is done.

Next, the procedure of the slice movement start process and the completion process by the slice moving unit 121c will be described with reference to the flowchart of FIG.
First, slice movement start processing will be described. The slice moving unit 121c temporarily stops the I / O processing (data reading / writing) by the array controller 12-1 to the logical disk that is the target of slice movement (step S11). Here, a row of the map table 122a corresponding to a slice to be moved (copied) is referred to as a row X of the map table 122a. When executing the step S11, the slice moving unit 121c proceeds to the process of step S12. In step S12, the slice moving unit 121c sets the number of the array to which the slice at the movement (copy) destination belongs and the number of the slice at the movement (copy) destination in the fields 46 and 47 of the row X of the map table 122a, respectively. To do.

  Next, the slice moving unit 121c sets zero as the copy completion size in the field 48 of the row X of the map table 122a (step S13). In step S13, the slice moving unit 121c sets a copy flag in the field 45 in the row X of the map table 122a. Next, the slice moving unit 121c saves the map information including the contents of the map table 122a at that time, that is, the information of the row X updated in steps S12 and S13 (step S14). In step S14, the map information is saved in a management information area, which will be described later, secured in each HDD in the disk array device 10. Then, the slice mover 121c resumes the I / O processing by the array controller 12-1 for the logical disk that has been the target of slice copy (move) (step S15).

  Next, slice movement (copy) completion processing will be described. When the slice copy (move) is completed, the slice mover 121c temporarily stops the I / O processing by the array controller 12-1 for the logical disk that is the target of the slice copy (step S21). Next, the slice moving unit 121c changes the contents of the fields 43 and 44 in the row X of the map table 122a to the array number of the array to which the copy destination slice belongs and the slice number of the copy destination slice, respectively (step S22). ).

  Next, the slice moving unit 121c clears the array number of the array to which the copy-destination slice belongs and the slice number of the copy-destination slice set in the fields 46 and 47 of the row X of the map table 122a, respectively (Step S1). S23). In step S23, the slice moving unit 121c also clears the copy flag set in the field 49 of the row X of the map table 122a. Next, the slice moving unit 121c saves the map information including the contents of the map table 122a at that time, that is, the information of the row X updated in steps S22 and S23 (step S24). In step S24, the map information is saved in the management information area secured in each HDD in the disk array device 10. Then, the slice moving unit 121c restarts the I / O processing by the array controller 12-1 for the logical disk that is the target of slice copy (step S25).

  In this embodiment, the above-described slice copy (move) can be performed online while the logical disk to which the slice is assigned is operating. For this purpose, data write processing in accordance with the data write request sent from the host 20 to the disk array device 10 needs to be executed by the data read / write unit 121d according to the procedure shown in the flowchart of FIG. Hereinafter, data write processing when a data write request for the slice is sent from the host 20 to the disk array device 10 during copying of the slice will be described with reference to FIG. Here, the row of the map table 122a corresponding to the slice in the logical disk to be written is referred to as the row Y of the map table 122a.

  First, the data read / write unit 121d determines whether a copy flag is set (standing) in the field 45 of the row Y of the map table 122a (step S31). In this example, the copy flag is set. In this way, if the copy flag is set, that is, if the slice to be written is being copied as a copy source, the data read / write unit 121d copies the data in the area to be written in the slice. It is determined whether or not is completed (step S32). The determination in step S32 is made based on the size information (copy completion size) set in the field 48 of row Y of the map table 122a.

  First, it is assumed that the data copy of the area to be written has been completed (step S32). In this case, the data read / write unit 121d writes the data to the corresponding areas of both the copy (move) source slice (write target slice) and the copy (move) destination slice (step S33). Considering that copying to a slice is interrupted for some reason during the process, for the area where copying has been completed, as described above, not only the copy destination but also the copy source (original) slice is written as a double. It is good to do the light.

  Next, it is assumed that the slice to be written is not being copied (step S31), or even if the slice to be written is being copied, copying of data in the area to be written has not been completed (step S32). ). In this case, the data read / write unit 121d writes data only to the area to be written in the slice to be written (copy source slice) (step S34).

  Next, storage of the map table 122a will be described. The map table 122a is an important table that associates a logical disk with physical allocation of slices that constitute the logical disk. Therefore, the loss of information (map information) stored in the map table 122a leads to the loss of data. For this reason, losing information in the map table 122a is not allowed even if both the array controllers 12-1 and 12-2 fail or a power failure occurs. Therefore, in the present embodiment, a storage method that has sufficient redundancy for failure / replacement of the array controller or HDD and that does not easily disappear is applied. Furthermore, in the present embodiment, the procedure for preventing the information in the map table 122a from being lost is also adopted in the flowchart shown in FIG. In other words, in this embodiment, the I / O processing requested from the host 20 employs a procedure that is resumed after the saving of the information (map information) in the map table 122a updated with the movement of the slice is completed. Yes.

  As described above, the HDD group 70-n including the HDDs 132A to 132D is connected to the array controllers 12-1 and 12-2 of the disk array device 10. However, hereinafter, for convenience of explanation, it is assumed that only the HDDs 132A to 132D are connected to the array controllers 12-1 and 12-2 as HDDs. In the present embodiment, as described below, the information of the map table 122a is stored using these HDDs 132A to 132D.

  First, a part of the storage area of the HDDs 132 </ b> A to 132 </ b> D is secured as the management information area 71. The management information area 71 is a special area used for storing information (management information) for disk array management by the array controllers 12-1 and 12-2. The management information area 71 is excluded from use as a slice. That is, the management information area 71 cannot be used as an area (user volume) from which the user can read / write data.

  In steps S14 and S24 in FIG. 5, the updated information (map information) in the map table 122a is stored in multiple in the management information areas 71 of the HDDs 132A to 132D, as indicated by arrows 72 in FIG. . Thereby, map information is multiplexed. Further, when the map information is loaded into the map table area of the memory 122 as the map table 122a, the map information is read out from the entire management information area 71 in the HDDs 132A to 132D as indicated by the arrow 73 in FIG. Done to the subject. Here, the read four pieces of map information are compared, and correct map information is determined by, for example, majority vote. The determined correct map information is loaded into the map table area of the memory 122 as the map table 122a. In this way, it is possible to withstand HDD failures and array controller failures.

  The statistical information collection unit 121e in FIG. 1 collects statistical information (hereinafter referred to as I / O statistical information) regarding I / O processing (access processing) for the slice for each slice. It is assumed that collection of I / O statistical information for each slice by the statistical information collection unit 121e is periodically performed. The collected I / O statistical information for each slice is stored in the statistical information area 122b of the memory 122 in the array controller 12-1. This I / O statistical information is, for example, the number of writes per unit time, the number of reads, the transfer size, or the I / O processing time. In general, this type of I / O statistical information is collected in units of logical disks and HDDs as described in Patent Document 1. However, it should be noted that in this embodiment, I / O statistical information in units of slices is used to determine the load adjustment in order to adjust the access load to the array and HDD by moving the slice. Of course, by using (for example, adding) the value indicated by the statistical information for each slice, the statistical value of I / O processing at the logical disk or array level can also be calculated.

  In the present embodiment, I / O statistical information acquired for each slice described above is used. The detection unit 121f, for example, periodically checks the I / O statistical information for each slice to detect a slice whose statistical value indicated by the I / O statistical information exceeds a certain value (threshold value). The slice moving unit 121c automatically moves a slice according to a predefined policy for a slice whose statistical value indicated by the I / O statistical information exceeds a threshold value. Thereby, for example, when the access load on an array exceeds a certain ratio (N%) of the capacity of the array, the slice moving unit 121c automatically sets a certain number of slices as the same number of slices in the array with the lowest load. Can also be changed. In addition to this, the allocation of slices is reviewed at regular intervals, and slices of arrays to which the RAID 1 + 0 level is applied are used for slices with a high access load, and RAID 5 levels are applied to slices with a low access load. It is also possible to alternate slices to use slices.

  Next, with respect to a method for eliminating a decrease in read access performance (read performance) of a logical disk by moving slices using I / O statistical information periodically collected by the statistical information collection unit 121e, This will be described with reference to FIG.

  FIG. 8 shows a logical disk 141 composed of a plurality of slices. The logical disk 141 includes areas 141a (#m) and 141b (#n). The areas 141a (#m) and 141b (#n) of the logical disk 141 are a plurality of physically continuous areas constituting the areas 142a (#m) and 142b (#n) of the array 142-0 (# 0), respectively. Assume that slices are combined.

  Assume that the host 20 requests access to a slice included in the area 141a (#m) or 141b (#n) of the logical disk 141. In this case, physically, the corresponding slice of the area 142a (#m) or 142b (#n) of the array 142-0 (# 0) is accessed.

  It is assumed that the number of reads per unit time of each of a plurality of slices constituting the area 141b (#n) of the logical disk 141 exceeds a predetermined threshold value. On the other hand, it is assumed that the number of reads per unit time of each of a plurality of slices constituting the area 141a (#m) of the logical disk 141 does not exceed the threshold value. In other words, the read access load (read load) in the area 141b (#n) of the logical disk 141 is high, and the read load in the area 141a (#m) of the logical disk 141 is low. In this case, regarding read access to the logical disk 141, the area 142b (#n) of the array 142-0 (# 0) corresponding to the area 141b (#n) of the logical disk 141 becomes a bottleneck. As a result, the read access performance of the logical disk 141 is degraded.

  The detection unit 121f has a high read load on the logical disk 141 based on the number of reads per unit time indicated by the I / O statistical information for each slice acquired by the statistical information acquisition unit 121e and stored in the statistical information area 122b. An area where slices are arranged can be detected as an area with a high read load. In general, a plurality of slices are continuous in a heavily loaded region. This tendency becomes higher as the capacity of the slice is smaller. Therefore, in the following description, it is assumed that the high load area is composed of a plurality of consecutive slices, but the present invention is not limited to this. Here, the detection unit 121f detects the area 141b (#n) of the logical disk 141 as a high read load area.

  Then, the array / slice definition unit 121a defines a new array 142-1 (# 1) shown in FIG. In accordance with this definition, the slice moving unit 121c copies (mirrors) the area 142b (#n) of the array 142-0 (# 0) to the array 142-1 (# 1) as indicated by the arrow 144 in FIG. ) Area 143b (#n) to be assigned. The slice included in the area 143b (#n) of the array 142-1 is a duplicate of the slice included in the area 142b (#n) of the array 142-0 (# 0). The area 142b (#n) of the array 142-0 (# 0) corresponds to the area 141b (#n) of the logical disk 141 as described above.

  In this state, it is assumed that the host 20 requests the disk array device 10 to write data to the slice included in the area 141b (#n) of the logical disk 141. In this case, as shown by an arrow 145 in FIG. 8, the data read / write unit 121d performs a region 142b (#n) of the array 142-0 (# 0) and a region 143b (# 1) of the array 142-1 (# 1). Write the same data to n). That is, the data read / write unit 121d writes data to the corresponding slice included in the area 142b (#n) of the array 142-0 (# 0). At the same time, the data read / write unit 121d writes the same data (mirror writing) to the corresponding slice included in the area 143b (#n) of the array 142-1 (# 1).

  On the other hand, when the host 20 requests data read from a slice included in the area 141b (#n) of the logical disk 141, the data read / write unit 121d reads data as follows. That is, the data read / write unit 121d, as shown by the arrow 146-0 or 146-1 in FIG. 8, corresponds to the corresponding slice and the slice included in the area 142b (#n) of the array 142-0 (# 0). Data is read from any one of the corresponding slices included in the region 143b (#n) of the array 142-1 (# 1). Here, the data read / write unit 121d is configured such that the read access is distributed to the area 142b (#n) of the array 142-0 (# 0) and the area 143b (#n) of the array 142-1 (# 1). The data is read from the area 142b (#n) or the area 143b (#n). For example, the data read / write unit 121d receives the area 142b (#n) of the array 142-0 (# 0) and the array every time the host 20 requests data read from the area 141b (#n) of the logical disk 141. Data is alternately read from the area 143b (#n) of the 142-1 (# 1).

  As described above, in this embodiment, the region 143b (#n) that is a duplicate of the region 142b (#n) including the slice with a high read load in the array 142-0 (# 0) is the array 142-0 (# 0) is assigned to another array 142-1 (# 1). Thereby, read access to the area 142b (#n) can be distributed to the area 143b (#n). This distribution of read access eliminates the bottleneck of read access in the area 142b (#n) in the array 142-0 (# 0) and improves the read performance of the area 141b (#n) in the logical disk 141.

  Next, the frequency of read accesses to the slices included in the area 141b (#n) of the logical disk 141 has decreased, and the read load on the area 141b (#n) has decreased. It is assumed that it was detected by a proper inspection. In this case, the slice moving unit 121c releases the area (replication area) 142b (#n) in the array 142-0 (# 0). That is, the slice moving unit 121c restores the area arrangement in the array corresponding to the area 141b (#n) of the logical disk 141 to the original state. Thereby, it is possible to improve the read access performance of the logical disk by effectively using the limited capacity of the physical disk.

  As described above, in this embodiment, the physical arrangement of data is performed so as to follow the gradually changing access characteristic to the logical disk 141 while operating the disk array device 10 and always to exhibit the best performance. By regularly reviewing, it is possible to always maintain the best access performance.

  Next, a method for eliminating a decrease in write access performance (write performance) of a logical disk applied in this embodiment will be described with reference to FIG. FIG. 9 shows a logical disk 151 composed of a plurality of slices. The logical disk 151 includes areas 151a (#m) and 151b (#n). The areas 151a (#m) and 151b (#n) of the logical disk 151 are configured by combining a plurality of physically continuous slices that constitute the areas 152a (#m) and 152b (#n) of the array 152, respectively. It shall be.

  In the example of FIG. 9, it is assumed that the number of writes per unit time of each of a plurality of slices constituting the area 151b (#n) of the logical disk 151 exceeds a predetermined threshold value. On the other hand, it is assumed that the number of writes per unit time of each of a plurality of slices constituting the area 151a (#m) of the logical disk 151 does not exceed the threshold value. In this case, the detection unit 121f acquires the area of the logical disk 151 based on the number of writes per unit time indicated by the I / O statistical information for each slice acquired by the statistical information acquisition unit 121e and stored in the statistical information area 122b. 151b (#n) is detected as an area having a high write access load (write load). Similarly, the detection unit 121f detects the area 151a (#m) of the logical disk 151 as an area with a low write load.

  Then, the array / slice definition unit 121a defines an area 153b (#n) corresponding to the area 151b (#n) of the logical disk 151 in the storage area of the semiconductor disk device 131, as indicated by an arrow 154b in FIG. To do. In accordance with this definition, the slice moving unit 121c re-slices the slice constituting the area 151b (#n) of the logical disk 151 from the area 152b (#n) of the array 152 to the area 153b (#n) of the semiconductor disk device 131. Deploy. This semiconductor disk device 131 has an extremely high access speed compared to the HDDs that make up the array 152. Therefore, the above-described rearrangement improves the write performance of the area 151b (#n) in the logical disk 151.

  The semiconductor disk device 131 is very expensive compared to the HDD. For this reason, it is disadvantageous in terms of cost performance to allocate all slices constituting the logical disk 151 to the semiconductor disk device 131. However, in the present embodiment, only the slices constituting the area 151 b having a high write load in the logical disk 151 are allocated to the semiconductor disk device 131. As a result, it is possible to effectively utilize the small storage area of the expensive semiconductor disk device 131.

  Next, the detection unit 121f detects that the frequency of write access to the slices configuring the area 151b (#n) of the logical disk 151 is low and the write load of the area 151b (#n) is low. Shall be. In this case, the slice moving unit 121c rearranges the slices included in the area 151b (#n) of the logical disk 151 from the semiconductor disk device 131 to an array constituted by HDDs, for example, the original array 152. As a result, the limited capacity of the high-cost semiconductor disk device 131 can be used more effectively, and the write access performance of the logical disk can be improved.

  In the above example, the method for eliminating the decrease in the write access performance (write performance) of the logical disk has been described. However, it is also possible to detect areas with high and low read loads (read access loads) from the logical disk 151 and rearrange slices in units of areas as in the above example. In this case, it is possible to effectively reduce the read access performance (read performance) of the logical disk by effectively using the limited capacity of the high-cost semiconductor disk device 131. Alternatively, areas with high and low access frequencies (access frequencies) may be detected from the logical disk 151, and slices may be rearranged in units of areas as in the above example. In this case, the limited capacity of the high-cost semiconductor disk device 131 can be effectively used to eliminate the decrease in the access performance of the logical disk.

  As described above, in this embodiment, the cost performance of the disk array device 10 can be improved by optimizing the cost of the physical area for storing data in accordance with the access characteristics to the logical disk. it can.

  As described above, the disk array device 10 includes the HDDs 132A (#A) and 132B (#B), and the HDDs 132C and 132D having different types from the HDDs 132A (#A) and 132B (#B). Yes. A method for improving the logical disk access performance using HDDs of different types, which is applied in this embodiment, will be described with reference to FIG.

  FIG. 10 shows a logical disk 161 composed of a plurality of slices. The logical disk 161 includes areas 161a (#m) and 161b (#n). The area 161b (#n) of the logical disk 161 is composed of a plurality of logically continuous slices whose access frequency is higher than a threshold value. On the other hand, the area 161a (#m) of the logical disk 161 is composed of a plurality of logically continuous slices whose access frequency is lower than the threshold value. In this case, the detection unit 121f detects the area 161b (#n) of the logical disk 161 as a frequently accessed area.

  FIG. 10 also shows a plurality of arrays, for example two arrays 162 and 163. As indicated by an arrow 164, the array 162 is configured using storage areas of HDDs 132A (#A) and 132B (#B) with low performance but low cost and large capacity. On the other hand, the array 163 is configured using storage areas of HDDs 132C (#C) and 132D (#D) having high performance but high price and small capacity, as indicated by an arrow 165. Thus, the array 162 is configured with emphasis on capacity and cost, and the array 163 is configured with emphasis on performance.

  The slice moving unit 121c arranges the slices included in the area 161a (#m) where the access frequency of the logical disk 161 is low, for example, in the area 162a of the array 162 as indicated by an arrow 166 in FIG. Further, the slice moving unit 121c arranges the slices included in the area 161b (#n) with high access frequency of the logical disk 161 in, for example, the area 163b of the array 163 as indicated by an arrow 167 in FIG. If the access frequency of the area 161a (#m) or 161b (#n) of the logical disk 161 changes after this arrangement, the slice mover 121c is included in the area 161a (#m) or 161b (#n). Change the array where the existing slice should be placed.

  As described above, in this embodiment, the arrays 162 and 163 having different characteristics (types), that is, the array 162 focusing on capacity and cost and the array 163 focusing on performance are prepared, and the access performance (access frequency) in the logical disk 161 is prepared. For each of the different areas, the array to which the slices constituting the area are to be assigned is switched (use differently). Thereby, in this embodiment, the cost performance as the disk array apparatus 10 can be improved.

[Modification]
Next, a modification of the above embodiment will be described. In the above embodiment, when a logical disk is configured, slices that configure the logical disk are allocated to the array. However, when the host requests the disk array apparatus to access the slice in the logical disk for the first time, the slice may be allocated in the storage area of the array.

  In the modified example of the above embodiment, when a slice in a logical disk is first used, that is, when the slice changes from an unused slice to a used slice, the array construction method allocates the slice in the storage area of the array. Applies. Assuming that this array construction method is applied to the disk array apparatus 10 shown in FIG. 1, the array construction method will be described with reference to FIG.

  FIG. 11 shows a logical disk 171 and an array 172 (# 0). The logical disk 171 includes slices 171a, 171b, 171c, 171d, 171e, 171f, and 171g. In this modified example, when the logical disk 171 is generated (defined), all of the slices constituting the logical disk 171 (that is, unused slices including the slices 171a to 171g) are in the array 172 (# 0). It is never assigned. Thereafter, the first access from the host 20 to the slice 171a occurs at time t1, and the first access from the host 20 to the slices 171d, 171e, and 171f occurs at time t2 after time t1.

  The array / slice definition unit 121a actually allocates an area of the array 172 to the slice 171a at time t1 when the first access to the slice 171a occurs, as indicated by an arrow 173a in FIG. Thereafter, the slice 171a is in a state where the allocation to the array 172 is completed, and becomes a used slice from an unused slice. Similarly, the array / slice definition unit 121a indicates that the slices 171d, 171e, and 171f are indicated by arrows 173d, 173e, and 173f in FIG. 11 at the time t2 when the first access to the slices 171d, 171e, and 171f occurs. The area of the array 172 is actually allocated. Thereafter, the slices 171d, 171e, and 171f are assigned to the array 172, and are used from unused slices.

  As described above, the array / slice definition unit 121a manages the physical real area of the array 172 in order from the slice in which the access has occurred among the slices constituting the logical disk 171. The disk array device 10 to which this management method is applied is optimal for a system in which the disk capacity actually used gradually increases due to an increase in the number of users, databases, or contents while continuing operation. The reason is that, at the time of system construction, a logical disk having a capacity predicted to be finally required can be generated regardless of the actual capacity of the array. Here, out of all slices of the logical disk, only the actually used slices are arranged in the array. For this reason, when the disk capacity to be used gradually increases, it becomes possible to add arrays in accordance with the increased capacity.

  Thereby, in the modification of the said embodiment, the initial investment at the time of system construction can be restrained few. Also, since the array area is not consumed for the unused area of the logical disk, the use efficiency of the physical disk capacity is increased. Further, in the modified example, as a result of the lack of physical disk capacity after the system operation is started, an array is added and the real area of the added array is allocated to the newly used slice of the logical disk. . Here, the logical disk itself is generated (defined) in advance with a finally required capacity. Therefore, even if an array is added and an actual area of the array is allocated as described above, it is not necessary to review the configuration recognized by the host computer, such as the capacity of the logical disk, and the operation of the system is facilitated.

  In FIG. 1, only the host 20 is shown as a host that uses the disk array device 10. However, by connecting a plurality of hosts including the host 20 and the disk array device 10, the plurality of hosts can share the disk array device 10.

  In addition, this invention is not limited to the said embodiment or its modification example as it is, A component can be deform | transformed and embodied in the range which does not deviate from the summary in an implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment or the modification thereof. For example, you may delete a some component from all the components shown by embodiment or its modification.

1 is a block diagram showing a configuration of a computer system provided with a disk array device according to an embodiment of the present invention. The figure for demonstrating the definition of the array applied to the embodiment, and a slice. The figure for demonstrating the definition of the logical disk applied in the embodiment. The figure which shows the data structure example of the map table 122a in FIG. 6 is a flowchart showing a procedure of slice movement start processing and completion processing in the embodiment. 6 is a flowchart showing a procedure of data write processing in the embodiment. The figure for demonstrating the preservation | save method of the map table 122a applied in the embodiment. The figure for demonstrating the method to eliminate the fall of the read performance of the logical disk applied in the embodiment. The figure for demonstrating the method of canceling the fall of the write performance of the logical disk applied in the embodiment. The figure for demonstrating the method of improving the cost performance as a disk array apparatus applied by the embodiment. The figure for demonstrating the array construction method applied in the embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Disk array apparatus, 11a-11d, 142-0, 142-1, 152, 162, 163, 172 ... Array, 12-1, 12-2 ... Array controller, 20 ... Host, 31-0, 31-1 , 141, 151, 161, 171 ... logical disk, 71 ... management information area, 120-1, 120-2 ... virtualization unit (virtualization device), 121a ... array / slice definition unit, 121b ... logical disk definition unit, 121c ... Slice moving unit 121d ... Data write unit 121e ... Statistical information collecting unit 121f ... Detecting unit 122 ... Memory 122a ... Map table 122b ... Statistical information area 131 ... Semiconductor disk device 132A-132D ... HDD (Hard disk drive).

Claims (2)

The storage area of at least one disk drive is divided into a predetermined capacity, and is composed of a combination of a plurality of the aforementioned slices in an array consisting of a set of slices having a certain capacity area, and is simply received from the host computer. An array controller connected to the host computer and the at least one disk drive holds a logical disk recognized as one disk volume in the map table in association with each slice constituting the logical disk. a logical disk management method for managing on the basis of the slice information,
The slice information held in the map table for each slice configuring the logical disk identifies the first array including the first slice allocated to the logical disk as the slice configuring the logical disk. First array identification information, first slice identification information for identifying the first slice in the first array, and data of the first slice being moved or copied to another slice Flag information indicating whether or not the data of the first slice is being moved or copied to another slice in an array different from the first array. Second array identification information for identifying the second array and second slice for identifying the second slice in the second array And a separate information,
The logical disk management method includes:
The statistical information collection means of the array controller acquires statistical information related to access processing for the slice for each slice constituting the logical disk, and the statistical information acquired for the slice is stored in the array controller. Storing in a storage means possessed by:
Based on the obtained statistics for each of the slices, the detecting means, wherein the array controller comprises the steps of: loading the access detecting high chair rice than a predetermined level,
The first array is detected as the detected slice by the first array identification information and the first slice identification information in the slice information held in the map table in association with the detected slice. If the first slice in the array controller is allocated , the slice moving means included in the array controller stores the data of the first slice in the logic to which the first slice is allocated. Moving or copying to a second slice in a second array that is not assigned to a disk and that is different from the first array that includes the first slice;
When executing the step of moving or copying, the slice moving unit temporarily stops data reading / writing on the logical disk by the data reading / writing unit included in the array controller, and then stores the data in the map table. In the slice information held in association with the detected slice, the second array for identifying the second array to which the second slice of the data movement or copy destination of the first slice belongs the flag indicating the an array identification information, and said second slice ID information for identifying the second slice in the second array, the data of the first slice is being moved or copied Information and then save the contents of the map table to the disk drive A step of resuming the data read / write to target the logical disk by the data read / write unit and thereafter,
When the moving or copying step is completed, the slice moving unit temporarily stops the data reading / writing on the logical disk by the data reading / writing unit, and then detected in the map table. In the slice information held in association with the slice, the first array identification information in the slice information indicates the second array, and the first slice identification information in the slice information is the first slice identification information. The second array identification information, the second slice identification information, and the flag information set in the slice information are updated to indicate the second slice in the second array. Clear, and then store the contents of the map table in the disk drive, and then read the data read / write A step of resuming the data read / write to target the logical disk by site means,
When the host computer requests to write data to a slice in the logical disk , the data read / write unit holds slice information associated with the requested slice in the map table If the flag information indicating that the migration or copying is in progress is not set, or if it is set and the migration or copying is not completed, the first array identification information in the slice information and the Data is written to the first slice in the first array identified by the first slice identification information, flag information indicating that the move or copy is in progress is set, and the move or copy is completed If, identified by the first array identification information and the first slice identification information in the slice information The second slice in the second array identified by the first of the first slice and the second array identification information and the second slice identification information in the slice information in the array A logical disk management method comprising: writing the same data. The storage area of at least one disk drive is divided into a predetermined capacity, and is composed of a combination of a plurality of the aforementioned slices in an array consisting of a set of slices having a certain capacity area, and is simply received from the host computer. In a virtualization apparatus for managing a logical disk recognized as one disk volume,
For each slice constituting the logical disk, first array identification information for identifying the first array including the first slice allocated as the slice constituting the logical disk, and the first array First slice identification information for identifying the first slice, flag information indicating whether the data of the first slice is being moved or copied to another slice, and data of the first slice Is moving or copying to another slice in an array different from the first array, second array identification information for identifying the other array to be moved or copied; while the slice information and a second slice identification information for identifying the another slice in the another array, held in association with the slice And up table,
For each slice constituting the logical disk, statistical information collection means for acquiring statistical information related to access processing for the slice;
Statistical information storage means for storing statistical information for each slice constituting the logical disk acquired by the statistical information collection means;
Based on the statistics for each slice constituting the logical disk that is stored in the statistical information storage means, and detecting means for loading the access detecting high chair rice than a predetermined level,
Data read / write means for performing data read / write on the logical disk requested by the host computer;
The first slice identification information and the first slice identification information in the slice information held in the map table in association with the slice detected by the detection means are used as the detected slice. If it is indicated that the first slice in one array is allocated, the data of the first slice is not allocated to the logical disk to which the first slice is allocated. a 2 slices, upon the from the first array including a first slice moved or copied to the second slice in another second array, the moving or copying of execution, the data read / Reading / writing data to / from the logical disk by the writing means is suspended, and then the map table is The in the is held in association with the detected slice away rice information, for identifying the second array the second slice of the moving or copying destination of the data of the first slice belongs said second array identification information, and said second slice ID information for identifying the second slice in the second array, the data of the first slice is being moved or copied the flag is set and the information, the contents of the map table stored in the disk drive Thereafter, thereafter the data read / write unit according to the slice moving means for resuming the data read / write to target logical disk shown And
When the movement or copying is completed, the slice moving unit temporarily stops data reading / writing on the logical disk by the data reading / writing unit, and then the detected data is detected in the map table . the absence Rice information is held in association with the slice, the first array identification information in the slice information indicates the second array, the first slice identification information in the slice information the first The second array identification information, the second slice identification information, and the flag information set in the slice information are updated to indicate the second slice in the second array. Clear, and then store the contents of the map table in the disk drive, and then read the data read / write means To restart the data read / write to target the logical disk by,
Said read / write means, said if it is required from the host computer writes data to the slices in the logical disk, the requested absence of rice information is held in association with the slice to the map table If the flag information indicating that the migration or copying is in progress is not set, or is set and the migration or copying is not completed, the first array identification information and the first If data is written to the first slice in the first array identified by one slice identification information, flag information indicating that the move or copy is in progress is set, and the move or copy is completed in the first identified by the first array identification information and the first slice identification information in the slice information The first slice and the same data to the second slice in the second array identified by said second array identification information and the second slice identification information in the slice information in the array of A virtualization apparatus characterized by writing

Images