JP2018169728A - Storage device, storage management method and storage management program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To dispose a block of a desired volume in the highest tier while optimizing data arrangement.SOLUTION: A storage device relating to this invention comprises: a storage unit in which performance requirement management information for managing setting of a lower limit value of a performance requirement in each of a plurality of volumes and block management information for, with respect to each data block of the plurality of volumes, managing a tier in a tier pool and a volume which it belongs to, in association with each other are stored; and a block rearrangement unit which refers to the performance requirement management information and the block management information to replace a first data block associated with the highest tier and a volume to which the lower limit value of the performance requirement is not set and a second data block associated with the lowest tier and a volume to which the lower limit value of the performance requirement is set, with each other.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a storage apparatus, a storage management method, and a storage management program, and more particularly to a storage apparatus, a storage management method, and a storage management program that perform optimal data placement.

  In Patent Document 1, the frequency of access to data stored in a storage is analyzed, and if the access frequency is high, data blocks are arranged on a pool using a faster and more expensive device (for example, SSD (Solid State Drive)). However, if the access frequency is low, it is possible to effectively store storage by arranging data blocks on a pool using a low-speed and inexpensive device (for example, Nearline Serial Attached SCSI (NL-SAS)). A technique to be used (data optimum arrangement technique) is disclosed.

  Patent Document 2 discloses a technology related to a storage management system that configures a pool using a plurality of storage devices having different performances and manages a storage subsystem that provides a plurality of virtual logical volumes from the pool. . The system according to Patent Document 2 is classified into a plurality of groups so that each of the plurality of virtual logical volumes satisfies the target response performance based on the information on the target response performance for each of the plurality of virtual logical volumes. Then, a new pool is assigned to each of the plurality of groups of the virtual logical volume. Further, information indicating the correspondence relationship between the pool and the virtual logical volume included in each group is generated, and the information is output.

  Patent Document 3 discloses a technique for replacing virtual pages based on access frequency.

International Publication No. 2011-117925 International Publication No. 2012/164718 Special table 2013-536478 gazette

  However, block rearrangement between hierarchies in the above-described technique is performed according to the analysis result of the access frequency in volume units. Therefore, there is a problem that a block of a desired volume cannot always be arranged in the highest hierarchy.

  The present invention has been made to solve such problems. A storage apparatus, a storage management method, and storage management for arranging a block of a desired volume in the highest hierarchy while performing optimal data arrangement The purpose is to provide a program.

The storage device according to the first aspect of the present invention includes:
Performance requirement management information for managing the setting of the lower limit value of the performance requirement for each of the multiple volumes,
For each data block of the plurality of volumes, block management information for managing the tier in the tier pool in association with the volume to which it belongs,
A storage unit for storing
Referring to the performance requirement management information and the block management information, the first data block associated with the highest tier and the volume for which the lower limit value of the performance requirement is not set, the lower tier and the lower limit value of the performance requirement are A block rearrangement unit that replaces the second data block associated with the set volume;
Is provided.

The storage management method according to the second aspect of the present invention includes:
Performance requirement management information for managing the setting of the lower limit value of the performance requirement in each of the plurality of volumes, and block management for managing the data blocks of the plurality of volumes in association with the tiers in the tier pool and the volumes to which the volumes belong. See information,
A first data block associated with a volume in which a lower limit value of the highest tier and the performance requirement is not set, and a second data block associated with a volume in which the lower tier and the lower limit value of the performance requirement are set; Replace.

A storage management program according to the third aspect of the present invention includes:
Performance requirement management information for managing the setting of the lower limit value of the performance requirement in each of the plurality of volumes, and block management for managing the data blocks of the plurality of volumes in association with the tiers in the tier pool and the volumes to which the volumes belong. Information reference processing, first data block associated with the highest tier and volume for which the lower limit value of the performance requirement is not set, and lower volume and associated with the volume for which the lower limit value of the performance requirement is set A process of replacing the second data block;
Is executed on the computer.

  According to the present invention, it is possible to provide a storage apparatus, a storage management method, and a storage management program for arranging a block of a desired volume in the highest hierarchy while performing optimal data arrangement.

1 is a block diagram showing a configuration of a storage device according to a first exemplary embodiment of the present invention. It is a flowchart which shows the flow of the storage management method concerning Embodiment 1 of this invention. It is a block diagram which shows the whole structure of the storage management system concerning Embodiment 2 of this invention. It is a figure which shows the example of the hierarchy pool management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the pool management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the volume management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the QoS management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the data optimal arrangement | positioning management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the access frequency management table concerning Embodiment 2 of this invention. It is a figure which shows the example of the volume setting screen concerning Embodiment 2 of this invention. It is a flowchart which shows the flow of the QoS setting process concerning Embodiment 2 of this invention. It is a flowchart which shows the flow of the block replacement | exchange process concerning Embodiment 2 of this invention. It is a figure which shows the example of the execution result of the block replacement process concerning Embodiment 2 of this invention. It is a flowchart which shows the flow of the block replacement process based on the access frequency concerning Embodiment 2 of this invention. It is a figure which shows the example of the setting policy of IOPS concerning a related technique.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. In the drawings, the same elements are denoted by the same reference numerals, and redundant description will be omitted as necessary for the sake of clarity.

<Embodiment 1>
FIG. 1 is a block diagram showing the configuration of the storage system 1000 according to the first embodiment of the present invention. The storage apparatus 1000 includes a storage unit 1100 and a block rearrangement unit 1200. The storage unit 1100 stores performance requirement management information 1110 and block management information 1120. The performance requirement management information 1110 is information for managing settings such as the upper limit value and the lower limit value of the performance requirement for each of a plurality of volumes (logical disks). The performance requirement is, for example, IOPS (Input / Output Per Second) or response time of the storage apparatus 1000. The block management information 1120 is information for managing the data blocks of a plurality of volumes in association with the tiers in the tier pool and the volumes to which they belong.

  The block rearrangement unit 1200 refers to the performance requirement management information 1110 and the block management information 1120, and refers to the first data block associated with the highest tier and the volume for which the lower limit value of the performance requirement is not set, the lower tier, and the performance. Replacement with the second data block associated with the volume for which the lower limit value of the requirement is set is performed.

  FIG. 2 is a flowchart showing the flow of the storage management method according to the first embodiment of the present invention. First, the block relocation unit 1200 of the storage apparatus 1000 refers to the performance requirement management information 1110 and the block management information 1120 (S101). Next, the block relocation unit 1200 associates the first data block associated with the volume for which the uppermost tier and the lower limit value of the performance requirement are not set and the volume for which the lower tier and the lower limit value of the performance requirement are set. The second data block is exchanged (S102).

  As described above, according to this embodiment, it is possible to arrange a block of a desired volume in the highest hierarchy while performing optimal data arrangement.

<Embodiment 2>
The second embodiment is an example of the first embodiment described above.
Here, the problem to be solved by the present embodiment will be described. First, as described above, volume block rearrangement using a storage data optimum arrangement (also called storage tiering or in-casing ILM (Information Lifecycle Management)) function is performed by QoS (Quality of Service) (I / (This is also referred to as O flow rate control.) When used in combination with a function, a technique for preventing a processing delay by placing a volume requiring high performance in a high performance pool is known (for example, Patent Document 1).

  When the QoS function is used based on IOPS, for example, a setting policy definition as shown in FIG. 15 can be given. FIG. 15 is a diagram illustrating an example of an IOPS setting policy according to related technology. In other words, when setting the IOPS upper limit value, the policy is to prevent the volume used by the own business from affecting other business (volume) due to an unexpectedly high load. On the other hand, when the IOPS lower limit value is set, the performance of the own business (volume) can be stably ensured to exceed the IOPS lower limit value without being affected by the sudden high load of other business (volume). It is a policy to make it. When both the IOPS upper limit value and lower limit value are set, it is an intermediate policy.

  Here, it can be considered that the volume for which the IOPS lower limit value is set is more important than the volume for which the IOPS upper limit value is set. The reason is that, when an IOPS lower limit value is set, an IOPS performance that is equal to or greater than the lower limit value must be ensured. Therefore, a higher-speed device for an I / O request that exceeds the IOPS lower limit value from the host computer. This is because it is necessary to place data in Further, when the IOPS upper limit value is set, it is only necessary to suppress the performance below the upper limit value, and therefore it is not always necessary to place data in a high-speed device.

  In other words, if the service level required by the business is the most important level, only the IOPS lower limit value is set, if it is the important level, the IOPS upper limit value and the lower limit value are set, and if it is the general level, only the IOPS upper limit value is set. You can think of it.

  Therefore, when combining the QoS function with the data optimal arrangement function, the request that the most important job with the highest importance, that is, the volume set with only the IOPS lower limit value, should be arranged in the highest hierarchy, which is a faster device. There is.

However, the prior art has the following problems (1) and (2).
(1) Since block relocation between layers of the data optimum arrangement function is performed according to the analysis result of the access frequency in units of volumes, until a frequently accessed block is arranged in a high-speed device of the highest hierarchy Depending on the analysis period, a certain period is required (for example, several days to one week). This is particularly noticeable when there is no past access frequency data immediately after a new volume is created.

  (2) As for the allocation of each block of the volume for which the IOPS lower limit value is set to the highest hierarchy, as a result of the access frequency analysis based on the past I / O history, if the block other than the volume has a higher access frequency The block of the volume is not arranged in the highest hierarchy.

  As described above, both (1) and (2) have a problem that the QoS of the set IOPS lower limit value may not be maintained as a result. In addition, if it is inevitably necessary to place the device immediately or continuously on a high-speed device in the highest tier, a separate pool from the tier pool is required, which causes an additional cost.

An outline of an embodiment of the present invention for solving the above problems will be described below.
When setting the IOPS lower limit value of a volume by the QoS function, an option is prepared for immediately allocating a block of the volume to a block of the highest hierarchy (or rearranging with a block of the assigned highest hierarchy). When this option is specified, the following is performed independently of the access frequency analysis.
-If there is an empty space in the highest hierarchy, relocation (move) is performed from the lower hierarchy block to the highest hierarchy block of the volume.
・ If there is no free space in the top tier, replace the blocks of the volume to which the blocks of the top tier belong without the IOPS lower limit value being replaced with the blocks of the volume in the lower tier. Do.

The above will be described with reference to FIGS.
First, the storage management client program 501 in the management console 500 sets the IOPS upper limit value and lower limit value by the QoS function for the volume constructed on the tiered pool using the data optimum arrangement function. At that time, the user transmits a QoS setting request to the storage apparatus 10 via a user interface such as the storage management client program GUI image of FIG. Here, the request contents include information on a volume name, an IOPS upper limit value, an IOPS lower limit value, an “assign block immediately to Tier 0” flag, and an “assign fixed block to Tier 0” flag. The QoS setting request when the “assign block immediately to Tier 0” flag is ON is an example of an immediate request for relocation with respect to a data block. The “assign block immediately to Tier 0” flag is an example of a flag indicating whether or not to fix to the highest hierarchy.

  When the QoS management unit 110 receives a QoS setting request from the storage management client program 501, the QoS management unit 110 stores the request content in the QoS management table 340.

  Next, when the “assign block immediately to Tier 0” flag is ON (= assign immediately), the QoS management unit 110 activates the data optimum arrangement management unit 120. The data optimum arrangement management unit 120 refers to the data optimum arrangement management table 350 and checks whether there is an empty block in Tier 0 (SSD pool). If there is an empty block, the data optimum arrangement management unit 120 rearranges the block so that the block of the requested volume is allocated to the Tier 0 pool. The data optimum arrangement management unit 120 uses the data optimum arrangement work area 12 which is a part of the area of the memory 11 of the storage apparatus 10 as a work area for data movement between blocks when performing relocation. If there is no free block, the data optimum arrangement management unit 120 refers to the QoS management table 340 and confirms whether a block of a volume for which the IOPS lower limit value is not set exists in Tier 0 (SSD pool). If the data exists, the data optimum arrangement management unit 120 rearranges the blocks so that the block of the volume for which the IOPS lower limit value is not set and the block of the requested volume are exchanged.

  In this way, without waiting for the next access frequency analysis, and even when the actual access frequency is not higher than other volumes, the block of the volume for which the IOPS lower limit value has been set is quickly Can be placed on high-speed devices.

  The “assign block to Tier 0 fixedly” flag will be described later. Next, the configuration of the embodiment of the present invention will be described.

  FIG. 3 is a block diagram showing the overall configuration of the storage management system 1 according to the second embodiment of the present invention. The storage management system 1 includes host computers 510, 520,..., A storage apparatus 10, and a management console 500. Here, the host computer 510 and the like and the storage apparatus 10 are connected by a storage area network (SAN) N1. The management console 500 and the storage apparatus 10 are connected by a local area network (LAN) N2.

  The storage apparatus 10 includes a storage management program 100, a memory 11, a performance log 21, a tiered pool management table 310, a pool management table 320, a volume management table 330, a QoS management table 340, and a data optimum arrangement management table. 350, an access frequency management table 360, a plurality of SSDs (Solid State Drives) S1, S2,... Sm (m is an integer of 2 or more), a plurality of HDDs (Hard Disk Drives) H1, H2,. ..Hn (n is an integer of 2 or more), a plurality of hierarchical pools, and a plurality of volumes LD1, LD2, LD3, and LD4.

  The storage management program 100, the performance log 21, the tiered pool management table 310, the pool management table 320, the volume management table 330, the QoS management table 340, the data optimum arrangement management table 350, and the access frequency management table. 360 is stored in a management storage area (not shown) in the storage apparatus 10. A control unit (not shown) of the storage apparatus 10 reads and executes the storage management program 100 into a memory (not shown; different from the memory 11). Thereby, the storage apparatus 10 functions as the QoS management unit 110, the data optimum arrangement management unit 120, and the performance monitoring unit 130.

  The memory 11 is a cache memory prepared in the storage apparatus 10. The memory 11 is an area for temporarily storing data of Read and Write I / O requests received from the host computers 510, 520,. The memory 11 includes a data optimum arrangement work area 12 in a part thereof.

  The data optimum arrangement work area 12 is used to temporarily store data to be moved when the data optimum arrangement management unit 120 replaces (moves) the data blocks between the hierarchies by rearrangement, or from the movement source and the movement destination. This is a work area used for storing necessary information regarding the blocks.

  The performance log 21 is a log file from which the performance monitoring unit 130 outputs various performance information such as IOPS and response time of the storage apparatus 10. The performance log 21 is information that is input during an access frequency analysis that is periodically executed by the performance monitoring unit 130.

  The hierarchical pool is a pool of a plurality of pools (for example, T0, T1, T2) virtually combined as one pool. In FIG. 4, as an example of a tier pool, a three-tier tier pool is shown in which a tier 0 (top tier) is an SSD pool, a tier 1 is a SAS pool, and a tier 2 is a near-line SAS (hereinafter NL-SAS) pool. In each layer, different hardware device characteristics and RAID (Redundant Arrays of Inexpensive Disks) types are arranged. Thus, according to the access frequency analysis, blocks that are often accessed are placed in high-speed and expensive devices such as Tier 0, and blocks that are not frequently accessed are placed in low-speed and inexpensive devices such as Tier 2 for efficient storage. Can be used. In the example of FIG. 4, the tier pool is three tiers, but may be two tiers or four or more tiers.

  The volumes LD1 to LD4 are cut out from the hierarchical pool and allocated to the host computers 510, 520,.

  The QoS management unit 110 receives a QoS setting request issued from the storage management client program 501 and returns a request result to the storage management client program 501 and a function for setting information such as an IOPS upper limit value and lower limit value for each volume. And a function for controlling the IOPS value for each volume based on information on the set IOPS upper limit value and lower limit value in cooperation with the performance monitoring unit 130. The QoS management unit 110 is an example of a reception unit.

  The data optimum arrangement management means 120, based on a request from the storage management client program 501 on the management console 500, a function of “allocating a block immediately to Tier 0” and a function of “allocating a block fixedly to Tier 0” and performance monitoring And a function of rearranging the blocks of the hierarchical pool based on the access frequency analysis result of the means 130. The data optimum arrangement management unit 120 is an example of the block rearrangement unit 1200.

  The performance monitoring unit 130 has a function of outputting various performance information such as IOPS of the storage apparatus 10 to the performance log 21, a function of periodically executing an access frequency analysis of the storage apparatus 10 using the performance log 21 as input information, and Have

  FIG. 4 is a diagram showing an example of the tier pool management table 310 according to the second embodiment of the present invention. The tier pool management table 310 has columns for ID, tier pool name, capacity, and number of tiers. ID is the primary key. The tier pool name is an identifier that uniquely identifies the tier pool. The capacity represents the total capacity of the tier pool. The number of tiers represents the number of tiers of each pool constituting the tier pool.

  From FIG. 4, for example, it can be seen that the tier pool 1 has a capacity of 100 GB and is composed of three tiers.

  FIG. 5 is a diagram showing an example of the pool management table 320 according to the second embodiment of the present invention. The pool management table 320 has columns of ID, tier pool name, tier number, pool name, and capacity. ID is the primary key. The tier pool name is an identifier that uniquely identifies the tier pool. The Tier number is a natural number starting from 0, and represents the tier to which the tier pool belongs. The pool name is an identifier that uniquely identifies the pool. The capacity represents the capacity of the pool.

  For example, from the top row of FIG. 5, it can be seen that Tier number 0 of the tier pool 1 is the SSD pool 1 and its capacity is 20 GB.

  FIG. 6 is a diagram showing an example of the volume management table 330 according to the second embodiment of the present invention. The volume management table 330 has columns of ID, volume name, capacity, tier pool name, number of blocks, and volume block number list. ID is the primary key. The volume name is an identifier that uniquely identifies the volume. The capacity represents the capacity of the volume name. The tier pool name represents the tier pool to which the volume belongs. The number of blocks represents the total number of blocks constituting the volume. The volume block number list is a serial number in the volume, and is a list of information on all blocks constituting the volume in CSV format.

  In FIG. 6, for example, from the top row, the volume 1 has a capacity of 70 GB and belongs to the tier pool 1, the total number of blocks is 14, and it is divided into blocks of LD1-1 to LD1-14 and placed on the tier pool 1. It can be seen that they are arranged.

  FIG. 7 is a diagram showing an example of the QoS management table 340 according to the second embodiment of the present invention. The QoS management table 340 has columns of ID, volume name, IOPS upper limit value, IOPS lower limit value, Tier 0 immediate allocation flag, Tier 0 fixed allocation flag, and hierarchical pool name. ID is the primary key. The volume name is an identifier that uniquely identifies the volume. The IOPS upper limit value represents the IOPS upper limit value set for the volume. The value when the upper limit value is not set is “−”. The IOPS lower limit value represents the IOPS upper limit value set for the volume. The value when the lower limit is not set is “−”. The Tier 0 immediate allocation flag is a flag indicating whether the block of the volume is immediately allocated to Tier 0 and rearranged regardless of the result of the access frequency analysis. If immediate allocation is to be performed, “YES” is set. Otherwise, “NO” is set. The Tier 0 fixed allocation flag is a flag indicating whether or not a block of the volume is fixedly allocated to Tier 0 regardless of the result of the access frequency analysis. If fixed allocation is performed, “YES” is set. Otherwise, “NO” is set. The tier pool name represents the tier pool to which the volume belongs. The QoS management table 340 is an example of performance requirement management information 1110.

  For example, from the top row in FIG. 7, 2000 IOPS is set as the IOPS upper limit value for volume 1, but the IOPS lower limit value is not set, and Tier 0 immediate allocation and Tier 0 fixed allocation are not performed. In other words, it can be seen that volume 1 belongs to tier pool 1.

  FIG. 8 is a diagram showing an example of the data optimum arrangement management table 350 according to the second embodiment of the present invention. The data optimum arrangement management table 350 has columns of ID, pool name, in-pool block number, block start address, block end address, hierarchical pool name, volume name, and volume block number. ID is the primary key. The pool name is an identifier that uniquely identifies the pool. The pool block number is a serial number in the pool. The block start address and the block end address respectively represent head and end address information in which data of the block number of the pool is stored. The tier pool name represents the tier pool to which the volume belongs. The volume name is an identifier that uniquely identifies the volume. The volume block number is a unique serial number within the volume, and is a block number constituting the volume. A combination of the pool management table 320 and the data optimum arrangement management table 350 is an example of the block management information 1120.

  In FIG. 8, for example, from the first row, the data of block number 1 of the SSD pool 1 is stored at addresses 0x00000000 to 0x0FFFFFFF, the SSD pool 1 belongs to the tier pool 1, and the current volume 1 It can be seen that the data of the volume block number LD1-1 is stored. When data is not stored, both the volume name and the volume block number are “−”.

  FIG. 9 is a diagram showing an example of the access frequency management table 360 according to the second embodiment of the present invention. The access frequency management table 360 has columns for ID, hierarchical pool name, pool name, pool block number, volume name, volume block number, average IOPS value, and moved flag. ID is the primary key. The tier pool name is an identifier that uniquely identifies the tier pool. The pool name is an identifier that uniquely identifies the pool. The pool block number is a serial number in the pool. The volume name is an identifier that uniquely identifies the volume. The volume block number is a unique serial number within the volume, and is a block number constituting the volume. The average IOPS value represents the average IOPS value for the block of the volume during the analysis period as a result of the access frequency analysis. The moved flag is set to “YES” when the block movement is completed. The initial value is “NO”. The access frequency management table 360 is an example of access frequency management information.

  For example, from the top row in FIG. 9, the data of the block number LD3-2 of the volume 3 is allocated to the block of the block number 6 in the pool of the SAS pool 1 constituting the tiered pool 1, and the average IOPS value is 500 IOPS. It can also be seen that block movement is incomplete.

  The host computers 510, 520,... Use the storage apparatus 10 via the storage area network N1. Specifically, an I / O request is issued to the volume of the storage apparatus 10 and an I / O processing result is received from the storage apparatus 10.

  The management console 500 includes a storage management client program 501. The storage management client program 501 has a function of issuing a QoS function setting request to the storage apparatus 10 and receiving a request result. Specifically, the storage management client program 501 issues a processing request such as setting an IOPS upper limit value or an IOPS lower limit value for each volume of the storage apparatus 10.

  Subsequently, the operation of the present embodiment will be described. In the following, based on the configuration examples of FIGS. 3 to 10, a processing flow for performing block rearrangement to the highest hierarchy of the hierarchy pool based on the QoS setting request from the management console will be described.

  FIG. 11 is a flowchart showing the flow of the QoS setting process according to the second embodiment of the present invention. First, the storage management client program 501 in the management console 500 transmits a QoS setting request to the storage apparatus 10 (step S11). The QoS setting request is a content instructed by the user via the GUI interface of the storage management client program 501 in FIG.

  FIG. 10 is a diagram showing an example of a volume setting screen according to the second embodiment of the present invention. The volume name in FIG. 10 is used to select a created volume from the list box, and “Volume 4” is selected here. The capacity displays information acquired from the storage apparatus 10 by the storage management client program 501 to the user according to the selected volume name. Here, “10 GB” is displayed. When the IOPS upper limit value and the IOPS lower limit value are set, specific IOPS values are designated. If not set, “Not set” is displayed as an initial value. Here, the IOPS upper limit value is not set, and “1000” is set as the IOPS lower limit value. “Assign block“ immediately ”to Tier 0” is a check box, and is designated when performing block setting reallocation processing to Tier 0 at the same time that QoS setting is performed for the specified volume. In FIG. 10, “allocate block“ immediately ”to Tier 0” is designated. “Assign block“ fixedly to Tier 0 ”” is a check box, and is specified when a block is permanently assigned to Tier 0 regardless of the result of access frequency analysis for the specified volume (QoS). In the first access frequency analysis after setting, if the block of the specified volume is in a tier other than Tier 0, relocation processing to Tier 0 will be performed). In FIG. 10, “assign a block“ fixedly to Tier 0 ”” is designated.

  Therefore, in step S11 of FIG. 11, the storage management client program 501 sends "Volume 4" as the volume name and "-" as the IOPS upper limit value to the storage apparatus 10 as a QoS setting request (meaning that no setting is made). Then, “1000” as the IOPS lower limit value, “YES” as the Tier 0 immediate allocation flag, and “YES” as the Tier 0 fixed allocation flag are transmitted to the storage apparatus 10.

  In the storage apparatus 10, the QoS management unit 110 receives the QoS setting request (step S21). Thereafter, the QoS management unit 110 sets ReturnCode on the memory 11 to “OK” (step S22). The ReturnCode is a memory area that can be set and referred to by a plurality of means on the storage apparatus 10 in common.

  Next, the QoS management unit 110 sets the received QoS setting request in the QoS management table 340 (step S23). FIG. 7 shows that the fourth line of the QoS management table 340 (the line whose ID is 1) has “−” as the IOPS upper limit (no upper limit is set), “1000” as the IOPS lower limit, and “1000” as the Tier 0 immediate allocation flag. In this example, “YES” and “YES” are registered as the Tier 0 fixed allocation flag. In this case, the QoS management unit 110 has received an immediate request for relocation with respect to the data block of the volume 4.

  If the Tier 0 immediate allocation flag is “NO” (step S24 is NO), the QoS management unit 110 proceeds to step S30. When the Tier 0 immediate allocation flag is “YES” (YES in step S24), the QoS management unit 110 activates the data optimum arrangement management unit 120 and takes over the subsequent processing (step S25). At the time of activation, volume name information (“volume 4” in this example) is passed to the data optimum arrangement management unit 120 as an argument. Then, the data optimum arrangement management unit 120 performs the replacement of the first data block and the second data block as follows in response to the immediate request.

  The data optimum arrangement management unit 120 refers to the volume management table 330, the QoS management table 340, the pool management table 320, and the data optimum arrangement management table 350 to check whether there is an empty block in Tier0 (step S26). That is, the data optimum arrangement management unit 120 refers to the block management information and determines whether or not there is an empty block that is a data block associated with the highest hierarchy and not storing data.

  This is done as follows. First, the data optimum arrangement management unit 120 refers to the volume management table 330 in FIG. 6 and acquires “2” as the number of blocks in the volume 4. Next, the data optimum arrangement management unit 120 refers to the QoS management table 340 in FIG. 7 and acquires “tier pool 1” as the tier pool name of the volume 4. Next, the data optimal arrangement management unit 120 refers to the pool management table 320 of FIG. 5 and acquires “SSD pool 1” as Tier 0 of “tier pool 1”, that is, a pool with a Tier number of 0. Next, the data optimum arrangement management unit 120 refers to the data optimum arrangement management table 350 in FIG. 8 and, among the entries whose pool name is “SSD pool 1”, both the volume name and the volume block number are “−”. Search for entries. When both the volume name and the volume block number are “−”, it means that the block of the volume is not allocated to the block in the pool.

  If there is an empty block in Tier 0 (YES in step S26), the data optimum arrangement management unit 120 allocates the block of volume 4 to the empty block in Tier 0 and rearranges it (step S27). Specifically, the data optimum arrangement management unit 120 searches for an entry whose volume name is “volume 4” in the data optimum arrangement management table 350, and sets the volume name and volume block number of the entry to the free space of “SSD pool 1”. Copy to entry. Further, the data optimum arrangement management unit 120 sets “-” to each of the volume name and the volume block number of the original “volume 4” entry. Physical data movement from the NL-SAS disk to the SSD disk is performed in response to the entry change in the data optimum arrangement management table 350. However, explanation of the data movement is omitted because it is an existing technology.

  Next, the operation when there is no empty block in Tier 0 (step S26 is NO) will be described. In the example of the optimum data arrangement management table 350 in FIG. 8, there are four entries of “SSD pool 1” with IDs 1 to 4, but in both cases, the volume name and volume block number are not “−”. Since the block has already been allocated, step S26 is a NO flow.

  The data optimum arrangement management unit 120 checks whether there is a volume for which the IOPS lower limit value is not set in Tier 0 with reference to the QoS management table 340 and the data optimum arrangement management table 350 (step S28). This is carried out as follows. First, it can be seen from the QoS management table 340 that there are two volumes “volume 1” and “volume 2” for which the IOPS lower limit value is not set. Next, the data optimum arrangement management unit 120 determines whether there is an entry with the volume name “volume 1” or “volume 2” for the “SSD pool 1” that is the pool of Tier 0 from the data optimum arrangement management table 350. Search in order.

  If there is no volume for which the IOPS lower limit value has not been set in Tier 0 (NO in step S28), a value of “Tier 0 immediate allocation impossible” is set in Return Code (step S29).

  The operation (part A of FIG. 11) in the case where there is a volume for which the IOPS lower limit value has not been set in Tier 0 (step S28 is YES) will be described with reference to FIG. FIG. 12 is a flowchart showing the flow of the block replacement process according to the second embodiment of the present invention. In the example of FIG. 8, since the entries with IDs 1, 2, and 3 correspond, step S28 is a flow of YES.

  Since it is known that the total number of blocks of the volume 4 for which QoS setting is requested is 2, the following loop processing is performed for this number (step S40).

  First, the data optimum arrangement management unit 120 searches the data optimum arrangement management table 350 for a moveable entry, that is, an entry for a volume for which the IOPS lower limit value has not been set, among the entries in Tier 0, that is, the SSD pool 1 (step S41). In the example of FIG. 8, an entry having an ID of 1 is found as the first entry.

  If no movable entry is found (NO in step S42), it is checked whether the value of the partially assigned flag is “ON” (step S48). The partially assigned flag is an area on the memory, and the initial value is “OFF”. When the partially assigned flag is “ON” (step S48 is YES), the value of “partially assigned to Tier 0 is OK” is set in ReturnCode (step S49). If the partially assigned flag is “OFF” (NO in step S48), a value of “Immediate assignment impossible to Tier 0” is set in ReturnCode (step S50).

  If a movable entry is found (YES in step S42), the data optimum arrangement management unit 120 copies the information of the entry to be moved to the data optimum arrangement work area 12, and temporarily saves it (step S43). As a result, the entry information including the volume name “volume 1” and the volume block number “LD1-1” is saved in the data optimum arrangement work area 12.

  Next, the data optimum arrangement management unit 120 moves the requested volume block to the Tier 0 block saved in step S43 (step S44). In the example of FIG. 8, for the first block of “volume 4”, that is, the entry whose ID is 19 with the volume block number “LD4-1”, the volume name “volume 4” and the volume block number The two pieces of information “LD4-1” are moved to the entry whose ID is 1.

  Next, the data optimum arrangement management unit 120 moves the data in the data optimum arrangement work area to the block before the requested volume migration (step S45). In the example of FIG. 8, based on the entry information saved in the data optimum arrangement work area 12, the volume name “volume 1” and the volume block number “LD1-1” are copied to the entry with ID 19, Rearrange (move) blocks.

  Next, the data optimum arrangement management unit 120 sets the value of the partially assigned flag on the memory to “ON” (step S46).

  The above processing is repeated for the total number of blocks in the volume 4 (step S47). As a result, if there are blocks that can be moved to Tier 0 by the number of blocks of the requested volume, they can be replaced and rearranged. FIG. 13 is a diagram illustrating an example of the execution result of the block replacement process according to the second embodiment of the present invention.

  After the above processing is completed, the data optimum arrangement management unit 120 returns ReturnCode representing the QoS processing request result to the storage management client program 501 (step S30). Thereafter, the storage management client program 501 receives the QoS processing request result (step S12). In this way, the data block of the volume for which the IOPS lower limit value is set can be immediately assigned to the highest tier (Tier0) of the tier pool without waiting for the result of the access frequency analysis.

  In addition, when the “assign fixed block to Tier 0” flag in the QoS setting request is ON (= fixed assignment), the operation is performed as follows according to the access frequency analysis periodically performed by the performance monitoring unit 130. To do.

FIG. 14 is a flowchart showing the flow of the block replacement process based on the access frequency according to the second embodiment of the present invention.
First, the performance monitoring unit 130 updates the access frequency management table 360 from the performance log 21 (S61). That is, the performance monitoring unit 130 calculates the access frequency in units of blocks in the analysis target tier pool from the performance information of the storage device 10 accumulated in the performance log 21 in the past certain period (eg, one week) as the average IOPS value. The access frequency is stored in the access frequency management table 360 in descending order.

  Next, the data optimum arrangement management unit 120 refers to the QoS management table 340 and confirms whether there is a volume that “allocates blocks fixedly to Tier 0” (S62). If it exists, the data optimum arrangement management unit 120 determines whether the block of the volume exists other than Tier 0 (S63). Here, it exists in the SAS pool or the NL-SAS pool that it exists other than Tier 0.

  If it exists other than Tier 0, the data optimum arrangement management means 120 refers to the access frequency management table 360, and the access frequency of the blocks existing in Tier 0 is excluded, except for the volume that “allocates blocks fixedly to Tier 0”. The lowest block is specified (S64). Then, the data optimum arrangement management unit 120 replaces the identified block (third data block) with the block (fourth data block) of the volume “allocating blocks fixedly to Tier 0” existing other than Tier 0. In this manner, the blocks are rearranged (S65). When rearrangement is performed, the data optimum arrangement management unit 120 uses the data optimum arrangement work area 12 as a work area for data movement between blocks.

  In this way, as a result of the access frequency analysis, even if the access frequency to the block of the volume is lower than the access frequency to the block of the other volume, the block of the volume is fixedly assigned to the highest hierarchy. , It can be prevented from being rearranged in the lower hierarchy.

  In other words, the above is as follows. That is, the storage unit further stores access frequency management information for managing the access frequency of the plurality of volumes in units of data blocks, and whether or not the performance requirement management information is fixed to the highest hierarchy for each volume. The block relocation unit further associates a flag indicating whether the access frequency is low and the flag is the highest when relocating each data block of the plurality of volumes based on the access frequency management information. A volume indicating that it is not fixed to the upper hierarchy and a third data block associated with the highest hierarchy, a volume indicating that the flag is fixed to the highest hierarchy, and a fourth data associated with the lower hierarchy Replace with block.

  As described above, there are cases where the QoS of the set IOPS lower limit value cannot be maintained while using both the optimal data placement function and the QoS function, and when it is desired to immediately or continuously place the QoS on the highest-level device. Therefore, it is possible to solve the problem that a separate pool from the tiered pool is required and extra cost is required.

  That is, according to the embodiment of the present invention, a volume in which an IOPS lower limit value is specified without being affected by block rearrangement between hierarchies according to an access frequency analysis result while using both the data optimum arrangement function and the QoS function. Can be immediately assigned to a high-speed device (= highest hierarchy). The reason for this is that, in the embodiment of the present invention, even when there is no free space in the highest tier of the tier pool, the data block of the volume set with the IOPS lower limit value existing in the lower tier and the top tier This is because by replacing the data block of a volume that does not have an existing IOPS lower limit value, the block of the volume can be immediately relocated to the highest tier (Tier0) of the tier pool. Thereby, the performance more than the set IOPS lower limit can be secured quickly.

  In addition, by excluding the volume for which the IOPS lower limit value has been set from the block relocation target based on the analysis result of the access frequency, the block of the volume is fixedly allocated to the high-speed device in the highest hierarchy regardless of the access frequency. Thus, the IOPS lower limit performance can be ensured.

<Other embodiments>
According to the present invention, the present invention can be applied to the use of a storage apparatus or storage management software having both a QoS function based on IOPS upper limit value / lower limit value control in volume units and a data optimum arrangement function in block units.

  Note that the present invention is not limited to the above-described embodiment, and can be changed as appropriate without departing from the spirit of the present invention.

  In the above-described embodiments, the present invention has been described as a hardware configuration, but the present invention is not limited to this. The present invention can also realize arbitrary processing by causing a CPU (Central Processing Unit) to execute a computer program.

  In the above example, the program can be stored and supplied to a computer using various types of non-transitory computer readable media. Non-transitory computer readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic recording media (for example, flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (for example, magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / W, DVD (Digital Versatile Disc), semiconductor memory (for example, mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (Random Access Memory)) are included. The program may also be supplied to the computer by various types of transitory computer readable media. Examples of transitory computer readable media include electrical signals, optical signals, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire and an optical fiber, or a wireless communication path.

1000 storage device 1100 storage unit 1110 performance requirement management information 1120 block management information 1200 block relocation unit 1 storage management system 10 storage device 11 memory 12 data optimal allocation work area 100 storage management program 110 QoS management unit 120 data optimal allocation management unit 130 Performance monitoring means 21 Performance log 310 Hierarchical pool management table 320 Pool management table 330 Volume management table 340 QoS management table 350 Data optimal allocation management table 360 Access frequency management table LD1 Volume LD2 Volume LD3 Volume LD4 Volume T0 Hierarchical pool T1 Hierarchical pool T2 Hierarchy Pool S1 SSD
H1 HDD
N1 Storage area network N2 Local area network 500 Management console 501 Storage management client program 510 Host computer 520 Host computer

Claims (7)

Performance requirement management information for managing the setting of the lower limit value of the performance requirement for each of the multiple volumes,
For each data block of the plurality of volumes, block management information for managing the tier in the tier pool in association with the volume to which it belongs,
A storage unit for storing
Referring to the performance requirement management information and the block management information, the first data block associated with the highest tier and the volume for which the lower limit value of the performance requirement is not set, the lower tier and the lower limit value of the performance requirement are A block rearrangement unit that replaces the second data block associated with the set volume;
A storage device comprising: A reception unit that receives an immediate request for relocation with respect to the second data block;
The storage apparatus according to claim 1, wherein the block rearrangement unit replaces the first data block and the second data block in response to the immediate request. The block rearrangement unit
Referring to the block management information, determining whether there is a free block that is associated with the highest hierarchy and is a data block in which no data is stored;
If it is determined that the empty block exists, the second data block is allocated to the empty block;
The storage apparatus according to claim 1 or 2, wherein when it is determined that the empty block does not exist, the first data block and the second data block are exchanged. The storage unit further stores access frequency management information for managing an access frequency of the plurality of volumes in units of data blocks,
The performance requirement management information further associates a flag indicating whether or not each volume is fixed to the highest tier,
The block rearrangement unit
When relocating each data block of the plurality of volumes based on the access frequency management information, the access frequency is low, and the flag indicates that the flag is not fixed to the highest hierarchy, and is associated with the highest hierarchy The third data block thus selected is replaced with a volume indicating that the flag is fixed to the highest hierarchy and a fourth data block associated with the lower hierarchy. The storage device described in 1. The storage apparatus according to any one of claims 1 to 4, wherein the performance requirement includes IOPS (Input / Output Per Second). Performance requirement management information for managing the setting of the lower limit value of the performance requirement in each of the plurality of volumes, and block management for managing the data blocks of the plurality of volumes in association with the tiers in the tier pool and the volumes to which the volumes belong. See information,
A first data block associated with a volume in which a lower limit value of the highest tier and the performance requirement is not set, and a second data block associated with a volume in which the lower tier and the lower limit value of the performance requirement are set; Storage management method to replace Performance requirement management information for managing the setting of the lower limit value of the performance requirement in each of the plurality of volumes, and block management for managing the data blocks of the plurality of volumes in association with the tiers in the tier pool and the volumes to which the volumes belong. Information reference processing, first data block associated with the highest tier and volume for which the lower limit value of the performance requirement is not set, and lower volume and associated with the volume for which the lower limit value of the performance requirement is set A process of replacing the second data block;
A storage management program that causes a computer to execute.

Images