WO2018167950A1

WO2018167950A1 - Storage system and copying control method

Info

Publication number: WO2018167950A1
Application number: PCT/JP2017/010953
Authority: WO
Inventors: 健太井上; 志織井上; 雄太西原; 隆司宮脇; 純二岩崎; 恭平神野
Original assignee: 株式会社日立製作所
Priority date: 2017-03-17
Filing date: 2017-03-17
Publication date: 2018-09-20

Abstract

This storage system manages host I/O history which is history of a host I/O process. When a PVOL has at least one difference region, the storage system carries out a present copying process that includes a sequence determination process. The sequence determination process includes: (a1) determining whether or not there is at least one past I/O destination region (I/O destination region used in a past copying process and identified from host I/O history) among all difference regions of the PVOL; and (a2) setting, when the determination result of (a1) is true, the copying order of all past I/O destination regions among all the difference regions so as to be higher than that of difference regions other than all the past I/O destination regions among all the difference regions.

Description

Storage system and replication control method

The present invention generally relates to replication control in a hierarchical storage system.

Data replication is performed, for example, to change the data arrangement. Regarding data arrangement control, a technique disclosed in Patent Document 1 is known.

Patent No. 5893745

Generally, a storage system has a CM (cache memory) and a plurality of PDEVs (nonvolatile storage devices), and provides a plurality of VOLs (logical volumes). Multiple VOLs are based on multiple PDEVs. Data written to the VOL is written to the PDEV corresponding to the write destination area. The plurality of VOLs include a PVOL and an SVOL that form a replication pair. “PVOL” is a primary VOL and is typically a replication source VOL. “SVOL” is a secondary VOL and is typically a replication destination VOL. The process of reading the replication target data from the PVOL to the CM and transferring the data in order to replicate the data to the SVOL is hereinafter referred to as “data replication”. Processing including data replication is hereinafter referred to as “duplication processing”.

As a storage system, a hierarchical storage system having a plurality of PDEVs with different I / O (Input / Output) performance is known. Hereinafter, a PDEV with relatively high I / O performance is referred to as “high-speed PDEV”, and a PDEV with relatively low I / O performance (at least a PDEV with lower I / O performance than the high-speed PDEV) is referred to as “low-speed PDEV”.

Hereinafter, an I / O request from the host is referred to as a “host I / O request”. In particular, a write request among host I / O requests is referred to as a “host write request”, and a read request among host I / O requests is referred to as a “host read request”. The “host” may be a host system (one or more host computers) connected to the storage system, or a first VM (VM (Virtual Machine) operating as a host system) executed in the storage system. Good. In the latter case, the second VM executed in the storage system may perform host I / O processing (processing according to the host I / O request) and replication processing as a storage controller. In this specification, the “storage system” is one or more physical storage devices. The “storage device” may be any device capable of storing data, and may be a server computer, for example.

In the host write process (process according to the host write request), the storage system temporarily stores write target data associated with the host write in the CM. In the asynchronous replication process (replication process performed asynchronously with the host write process), the storage system includes writing the write target data once to the PVOL and then reading the data from the PVOL (PDEV) to the CM in the subsequent replication process. Perform data replication. For this reason, if the PDEV that stores the replication target data is a low-speed PDEV, it takes time to read data from the PDEV to the CM. As a result, the data replication time increases. This can cause at least one of the following problems.
(1) Assume that the storage system can receive a host I / O request even during data replication. If the data replication time is long, the load time for data replication is long. For this reason, it is considered that the host I / O performance (host I / O processing performance) is more likely to be affected by data replication.
(2) It is assumed that the storage system can receive a host read request designating SVOL during replication processing. When the read target data of the received host read request is not in the SVOL, the storage system needs to replicate the read target data from the PVOL to the SVOL. When the read target data is stored in the low-speed PDEV, it takes a long time to copy the read target data from the PVOL to the SVOL. For this reason, it takes a long time to respond with the read target data. That is, the host I / O performance decreases.
(3) It is assumed that the storage system is configured to back up data from the SVOL to a backup VOL in or outside the storage system. When the backup target data is not in the SVOL, the storage system needs to replicate the backup target data from the PVOL to the SVOL. When the back tap target data is stored in the low speed PDEV, it takes a long time to replicate the backup target data from the PVOL to the SVOL.

The storage system manages the host I / O history, which is the history of host I / O processing. When the PVOL has at least one difference area, the storage system performs the current replication process including the order determination process. The order determination process includes (a1) one or more past I / O destination areas (all I / O destination areas in the past replication process that are specified from the host I / O history) of all the PVOL difference areas. If the determination result of (a2) and (a1) is true, the replication order of all past I / O destination areas in all difference areas is Including higher than the replication order of the differential area other than all the past I / O destination areas.

In a tiered storage system, it can be expected to shorten the data replication time and reduce the impact on the host I / O performance.

The outline | summary of embodiment is shown. 1 shows a configuration of a computer system according to an embodiment. The management information and function in a storage apparatus are shown. The structure of an I / O monitor table is shown. The structure of a past history table is shown. The structure of a slot management table is shown. The structure of a replication order table is shown. The structure of an area management table is shown. The structure of a configuration management table is shown. The flow of replication processing is shown. Details of the flow of order determination (S1001 in FIG. 10) are shown. Details of the flow of influence determination (S1002 in FIG. 10) are shown. Details of the flow of area reservation (S1003 in FIG. 10) are shown. . Details of the flow of order update (S1007 in FIG. 10) are shown. Details of the flow of post-processing (S1009 in FIG. 10) are shown.

Hereinafter, an embodiment will be described.

In the following description, information may be described using an expression such as “xxx table”, but the information may be expressed in any data structure. That is, in order to show that the information does not depend on the data structure, the “xxx table” can be referred to as “xxx information”. In the following description, the configuration of each table is an example, and one table may be divided into two or more tables, or all or part of the two or more tables may be a single table. Good.

In the following description, the “interface unit” includes one or more communication interface devices. The one or more communication interface devices may be one or more similar interface devices (for example, one or more NIC (Network Interface Card)) or two or more different types of interface devices (for example, NIC and HBA (Host Bus Adapter)). ).

In the following description, the “memory unit” is one or more memories. Regarding the memory part, the at least one memory may be a volatile memory. The memory unit is mainly used for processing by the processor unit.

In the following description, the “processor unit” is one or more processors. The at least one processor is typically a microprocessor such as a CPU (Central Processing Unit). Each of the one or more processors may be a single core or a multi-core. The processor may include a hardware circuit that performs part or all of the processing.

In the following description, the function may be described using the expression “kkk part”, but the function may be realized by executing one or more computer programs by the processor part. The hardware circuit (for example, FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)) may be used. When the function is realized by the program being executed by the processor unit, since the predetermined processing is appropriately performed using the memory unit and / or the interface unit, the function is at least a part of the processor unit. May be. The processing described with the function as the subject may be processing performed by the processor unit or a device having the processor unit. The program may be installed from a program source. The program source may be, for example, a program distribution computer or a computer-readable recording medium (for example, a non-transitory recording medium). The description of each function is an example, and a plurality of functions may be combined into one function, or one function may be divided into a plurality of functions.

In the following description, “PDEV” means a physical storage device, and typically a non-volatile storage device (for example, an auxiliary storage device). The PDEV may be, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive).

Also, in the following description, “RAID” is an abbreviation for Redundant “Array” of “Independent” (or “Inexpensive)” Disks. The RAID group is composed of a plurality of PDEVs (typically the same kind of PDEVs), and stores data according to the RAID level associated with the RAID group. The RAID group may be referred to as a parity group. The parity group may be, for example, a RAID group that stores parity. The RAID group may be composed of a plurality of storage areas based on a plurality of PDEVs.

In the following description, “VOL” (or “volume”) is an abbreviation for logical volume, and may be a logical storage device. The VOL may be a substantial VOL (RVOL) or a virtual VOL (VVOL). The “RVOL” may be a VOL based on a physical storage resource (for example, one or more RAID groups) included in the storage system that provides the RVOL. The “VVOL” is composed of a plurality of virtual areas (virtual storage areas) and may be a VOL according to capacity virtualization technology (typically ThinThProvisioning). The “pool” may be a logical storage space composed of a plurality of pages (a plurality of logical storage areas). When a page is not allocated to the virtual area (VVOL virtual area) to which the address specified by the host write request belongs, the storage system allocates a page from the pool to the virtual area (write destination virtual area). The storage system may write the write target data accompanying the host write request to the allocated page. The “pool VOL” may be a VOL that is a component of the pool. The pool VOL may be an RVOL.

Also, in the following description, when a description is made without distinguishing the same type of element, a common part in the reference code is used, and when a description is made by distinguishing the same type of element, a reference code may be used.

FIG. 1 shows an outline of the embodiment.

The storage system 102 includes a memory unit 98 including a CM (cache memory) 111, a processor unit 94, and a plurality of PDEVs 10 having different I / O performance. The RAID group 20 is configured by two or more PDEVs 10 having the same (or similar) I / O performance. For example, there are a RAID group 20S composed of two or more SSDs 10S and a RAID group 20A composed of two or more SASs (Serial Attached SCSI) -HDD 10A. The SSD 10S is an example of a first storage device (for example, a high speed PDEV). The SAS-HDD 10A is an example of a second storage device (for example, a low-speed PDEV) that has lower I / O performance than the first storage device.

The processor unit 94 provides a PVOL (primary VOL) 113A and an SVOL (secondary VOL) 113B that constitute a replication pair. The SVOL 113B may exist in a storage system (not shown) different from the storage system 102.

Two or more RAID groups having different I / O performance such as the

RAID groups

20S and 20A are the basis of the logical storage space. A part of the logical storage space is allocated to at least PVOL 113A. In this embodiment, the logical storage space is the pool 40, and each of the PVOL 113A and the SVOL 113B is a VVOL. The pool 40 includes a plurality of tiers 30 having different I / O performance. Each tier 30 is composed of one or more pool VOLs 113P. Each pool VOL 113P is divided into a plurality of pages 60. For example, the tier 30S includes one or more pool VOL 113PS based on the RAID group 20S (two or more SSDs 10S). The tier 30A is composed of one or more pool VOLs 113PA based on the RAID group 20A (two or more SAS-HDDs 10A). The page 60 is an example of a logical storage area. The page 60 in the tier 30S is an example of the first logical storage area. The page 60 in the tier 30A is an example of a second logical storage area.

When the processor unit 94 receives a host I / O request designating an area (address) in the PVOL 113A from the host 101, the processor unit 94 performs host I / O processing on the PVOL 113A in response to the host I / O request. The host I / O processing is, for example, as follows. The host 101 is a computer connected to the storage system 102, but may instead be a VM (Virtual Machine) executed in the storage system 102.
The host I / O request is a host write request, and the page 60 is not allocated to the write destination area 50 (the virtual area 50 to which the specified address belongs) (for example, 50A) specified in the host write request The processor unit 94 writes the write target data to the CM 111 and allocates a free page 60 (for example, 60A) to the write destination area. The processor unit 94 writes the write target data to the allocated empty page 60 (PDEV 10 that is the basis of the empty page 60).
When the host I / O request is a host read request, the processor unit 94 is assigned to the read source area 50 (virtual area 50 to which the specified address belongs) (for example, 50A) specified by the host read request. Read target data is read from the existing page 60 (for example, 60A) (PDEV 10 that is the basis of the page 60) to the CM 111, and the read target data is returned to the host 101.

The processor unit 94 manages a difference area which is an area having a difference from the SVOL 113B in the PVOL 113A. The capacity of the difference area may be the same as the capacity of the virtual area 50 or the page 60 (that is, the allocation unit and the difference management unit (and the data replication unit) may be the same), but in this embodiment, the capacity of the difference area Is smaller than the capacity of the virtual area 50 or the page 60. Specifically, in the present embodiment, an example of the difference area is a difference slot. The virtual area 50 and the page 60 are composed of two or more slots. Differences are managed in slot units, and data replication is performed.

The processor unit 94 is adapted to perform replication processing including data replication. For data replication, for each of one or more difference slots in PVOL 113A, data (replication target data) is read from CM 60 to page 111 allocated to virtual area 50 including the difference slot, and the difference slot in SVOL 113B is read. This is a process of transferring the data to copy the data to the corresponding slot.

The processor unit 94 manages a host I / O history 120 that is a history of host I / O processing. The host I / O history 120 includes an I / O destination slot in the PVOL 113A and a time of host I / O processing for each I / O destination slot.

When the PVOL 113A has at least one difference slot, the processor unit 94 performs the current duplication process including the following (A). (A) is an order determination process including the following (a1) and (a2).
(A1) It is determined whether there is one or more past I / O destination slots among all the differential slots of the PVOL 113A. Each of the one or more past I / O destination slots is an I / O destination slot during past duplication processing and is a slot specified from the host I / O history 120.
(A2) When the determination result of (a1) is true, the duplication order of all past I / O destination slots in all difference slots is set to the difference slot other than all past I / O destination slots in the all difference slots. Higher than the replication order.

The processor unit 94 performs data duplication in descending order of duplication. As a result, among all the differential slots, the data in the past I / O destination slot preferentially exists in the SVOL. For example, it is assumed that the storage system 102 can receive a host read request designating the SVOL 113B during the replication process. If the host I / O processing for the SVOL 113B during the current replication processing is related to the host I / O processing for the PVOL 113A during the past replication processing, the I / O destination range in those host I / O processing ( The address range is likely to be similar. Therefore, when a host read request for the SVOL 113B is received during the current replication process, there is a high possibility that the read target data according to the host read request already exists in the read source slot specified by the host read request. . As a result, it is expected that the possibility of the host I / O performance being reduced is reduced.

The processor unit 94 sets the replication order of all past I / O destination slots in descending order of the host I / O amount in the past replication processing. “Host I / O amount” refers to either the host I / O frequency (frequency of host I / O requests) or host I / O data amount (amount of data input / output in response to host I / O requests). It's okay. In this embodiment, for each of all past I / O destination slots, the host I / O amount is a cumulative value of the host I / O frequency for the past I / O destination slot during past duplication processing. Hereinafter, the host I / O frequency is referred to as “IOPS”. A slot with a larger host I / O amount is considered to have a higher possibility of being accessed (for example, a possibility of being a read source slot of a host read request designating SVOL 113B). For this reason, it is expected that the higher the host I / O amount, the lower the possibility that the host I / O performance during the current replication process will decrease by increasing the replication order.

If there is a past I / O destination slot belonging to the low speed media among all the differential slots, the processor unit 94 gives priority to a past I / O destination slot having a higher duplication order among the past I / O destination slots belonging to the low speed tier. Therefore, it may shift to a high-speed tier.

Here, the “low-speed tier” is a tier with relatively low I / O performance in the pool 40, for example, tier 30A (tier 2). Hereinafter, pages in the low speed tier may be referred to as “low speed pages”. Further, a PDEV (for example, SAS-HDD 10A) that is a base of the low-speed tier may be referred to as a “low-speed PDEV”.

“High-speed tier” is a tier with relatively high I / O performance in the pool 40, for example, tier 30S (tier 1). Hereinafter, pages in the high-speed tier are sometimes referred to as “high-speed pages”. Also, a PDEV (for example, SSD 10S) that is the basis of the high-speed tier may be referred to as a “high-speed PDEV”.

“The past I / O destination slot belonging to the low-speed tier” is a past I / O destination slot in which the page allocated to the virtual area including the past I / O destination slot is a page in the low-speed tier.

“Shifting the past I / O destination slot belonging to the low-speed tier to the high-speed tier” is an example of data rearrangement, and specifically includes the following.
Relocate (move) data in a low speed page (eg, 60B) allocated to a virtual area (eg, 50B) including the past I / O destination slot to an empty high speed page (eg, 60C) (arrow 91) ),
Assign a reallocation-destination high-speed page to the virtual area (for example, 50B) including the past I / O destination slot instead of the reallocation-source low-speed page (see arrow 92).

Of the past I / O destination slots belonging to the low-speed tier, the past I / O destination slot having the highest replication order is preferentially shifted to the high-speed tier. Is done. This is because there is a high possibility that the reading source is a high-speed tier (high-speed PDEV). In addition, for example, the storage system 102 is configured to back up data from the SVOL 113B to a backup VOL (not shown) inside or outside the storage system 102, and there is no backup target data in the SVOL 113B. When data to be backed up is copied from the PVOL 113A to the SVOL 113B, there is a high possibility that the back tap target data is stored in the high speed page (high speed PDEV). For this reason, it is expected that the time required for data replication of the backup target data from the PVOL 113A to the SVOL 113B is short.

Further, the processor unit 94 performs the following (B) and (C) in addition to or instead of (A) in the current replication process.
(B) If there is one or more differential slots to which one or more low-speed pages are assigned, relocating (moving) data from the low-speed pages to the high-speed pages will reduce the host I / O performance during the current replication process Determine whether or not
(C) When the determination result of (B) is false, for at least one of the one or more difference slots, the data is rearranged (moved) from the low speed page assigned to the difference slot to the empty high speed page. In addition, the high speed page is allocated to the virtual area including the difference slot instead of the low speed page.

Thus, it can be expected that the replication target data is rearranged from the low speed page to the high speed page (that is, the data replication time is shortened) while maintaining the host I / O performance during the current replication processing.

Hereinafter, this embodiment will be described in detail.

FIG. 2 shows a configuration of a computer system according to the present embodiment.

The host 101 is connected to the storage apparatus 202 via a SAN (Storage Area Network) 181. A management system 69 is connected to the storage apparatus 202 via a LAN (Local Area Network) 182. A host 101 is also connected to the LAN 182. At least one of the SAN 181 and the LAN 182 may be another type of communication network. Further, the storage apparatus 202, the host 101, and the management system 69 may be connected to the same communication network. There may be a plurality of at least one of the storage apparatus 202, the host 101, and the management system 69.

The host 101 issues an I / O request specifying either PVOL 113 or SVOL 113B. The host 101 has an I / F (for example, HBA (Host Bus Adapter)) 264 connected to the SAN 181 (storage device 202) and an I / F (for example, Network Interface Card) (NIC) connected to the LAN 182 (storage device 202). ) 263, a memory 262, and a CPU 261 connected to the elements 262 to 264.

The management system 69 may be composed of one or more computers. Specifically, for example, when the management computer has a display device and the management computer displays information on its own display device, the management computer 69 may be the management computer 69. For example, when a management computer (for example, a server) transmits display information to a remote display computer (for example, a client) and the display computer displays the information (when the management computer displays information on the display computer) ), A system including at least the management computer among the management computer and the display computer may be the management system 69. The management system 69 may include an I / F (for example, NIC) 63 connected to the LAN 182 (storage device 202), a memory 62, and a CPU 61 connected to these

elements

62 and 63. That the computer in the management system 69 “displays display information” may be that display information is displayed on a display device included in the computer, or the computer transmits display information to the display computer. (In the latter case, display information is displayed by a display computer). The memory 62 stores information accessed by the CPU 61 and programs executed by the CPU 61. The CPU 61 manages the storage apparatus 202 with reference to information in the memory 62 by reading and executing the program from the memory 62.

The storage apparatus 202 includes a storage controller 108 and a plurality of PDEVs 10 (10S, 10A,...) Connected to the storage controller 108. A plurality of RAID groups 20 (20S, 20A,...) Are configured by the plurality of PDEVs 10.

The storage controller 108 includes a CPU 112, a host interface (HI / F) 115, a device interface (DI / F) 114, an SM (shared memory) 113, a CM (cache memory) 111, a transfer circuit 118, and a management circuit. An interface (MI / F) 116 is provided. Data transfer among the HI / F 115, the CPU 112, the CM 111, and the DI / F 114 is performed via the transfer circuit 118. The SM 113 and the MI / F 116 are connected to the CPU 112. The I /

Fs

114, 115, and 116 are examples of the interface unit. CM 111 and SM 113 are examples of the storage unit. The CPU 112 is an example of a processor unit. In FIG. 2, only one component of the storage controller 108 is shown, but at least one of these components is provided in the storage controller 108 to ensure high performance and high availability. May be.

The DI / F 114 is connected to each PDEV 10. Communication with the PDEV 10 is performed via the DI / F 114. The HI / F 115 is connected to the host 101. Communication with the host 101 is performed via the HI / F 115. The MI / F 116 is connected to the management system 69. Communication with the management system 69 is performed via the MI / F 116.

In the CM 111, data input / output to / from the PDEV 10 is temporarily stored. In SM113. A program executed by the CPU 112 and information referred to or updated by the CPU 112 (hereinafter, management information) are stored. The CPU 112 performs various controls of the storage device 202.

1 may be the storage apparatus 202, or may be a system including the storage apparatus 202 and the management system 69. Therefore, the example of the processor unit 94 illustrated in FIG. 1 may be the CPU 112 or may include the

CPUs

112 and 61. For example, the CPU 61 of the management system 69 acquires at least a part of the host I / O history 120 from the storage apparatus 202, and will be described with reference to processing described later (eg, with reference to FIGS. 10 to 15) based on the acquired information. At least a part (for example, at least a part of S1001, S1002, and S1007 in FIG. 10) may be executed. Hereinafter, to simplify the description, it is assumed that the storage system 102 is the storage device 202 and the processor unit 9 is the CPU 112.

FIG. 3 shows management information and functions in the storage apparatus 202.

The management information stored in the SM 113 includes a host I / O history 120, a slot management table 303, a replication order table 304, an area management table 305, and a configuration management table 306. The host I / O history 120 includes an I / O monitor table 301 and a past history table 302. Each table will be described later.

The functions realized by the CPU 112 executing the program in the SM 113 include an I / O unit 311, a duplication unit 312, and a rearrangement unit 313.

The I / O unit 311 executes host I / O processing in response to a host I / O request. Further, the I / O unit 311 updates the I / O monitor table 301 in the host I / O history 120 as appropriate. The I / O unit 311 manages the difference between the PVOL 113A and the SVOL 113B.

The replication unit 312 performs a replication process. The duplicating unit 312 includes an order determining unit 321, an influence determining unit 322, an area securing unit 323, an order updating unit 324, a post-processing unit 325, and a control unit 326. The order determining unit 321 determines the duplication order of each difference slot based on the past history table 302. The influence determination unit determines, based on the host I / O history 120, whether or not the rearrangement from the low speed page to the high speed page lowers the host I / O performance during the current replication process. The area securing unit 323 secures an empty page from the high-speed tier. The order update unit 324 updates the replication order determined by the order determination unit 321 based on the host I / O process during the current replication process. The post-processing unit 325 reflects the history of the host I / O process during the current replication process in the past history table 302. The control unit 326 controls calling of functions in the duplication unit 312.

The relocation unit 313 performs data relocation between different tiers 30.

Hereinafter, each table in the SM 113 will be described with reference to FIGS.

FIG. 4 shows the configuration of the I / O monitor table 301.

The I / O monitor table 301 is a table that holds information as a monitoring result of host I / O processing. The I / O monitor table 301 has a record for every predetermined time. Each record holds information such as time 401, slot IOPS 402, difference slot # 403, status 404, and volume IOPS 405. Hereinafter, one fixed time (referred to as “target fixed time” in the description of FIG. 4) is taken as an example.

Time 401 indicates the start time of the target fixed time. The slot IOPS 402 indicates the IOPS (I / O frequency) at the target fixed time for the slot accessed during the target fixed time (targeted for host I / O processing). Slot # 403 indicates the identification number of the slot accessed during the target fixed time. The status 404 indicates the status for the target fixed time. There are three types of status 404 values: “0” (start), “1” (end), and “Null” (between start and end). The volume IOPS 405 indicates IOPS (I / O frequency) at the target fixed time for the VOL including the slot accessed at the target fixed time.

When two or more slots are accessed at a certain target time, two or more slot numbers corresponding to two or more slots are registered as slot # 403, and two or more IOPS respectively corresponding to two or more slots are registered. Registered as slot IOPS 402. The host I / O process is monitored by the I / O unit 311, and a record having information as a monitoring result at the certain time is added to the I / O monitor table 301 every time a certain time elapses. .

FIG. 5 shows the structure of the past history table 302.

The past history table 302 is a table that holds a history of host I / O processing during past replication processing. The past history table 302 has a record for each past replication process. Each record holds information such as a start time 501, an end time 502, a duplication processing time 503, a cumulative slot IOPS 504, a cumulative volume IOPS 505, and a slot # 506. Hereinafter, one past duplication process (referred to as “target past duplication process” in the description of FIG. 5) is taken as an example.

The start time 501 indicates the start time of the target past replication process. The end time 502 indicates the end time of the target past replication process. The replication processing time 503 indicates the time taken for the target past replication processing (difference between the end time 502 and the start time 501).

The cumulative slot IOPS 504 indicates the cumulative value of the slot IOPS during the target past replication process for the slot that is the target of the host I / O process during the target past replication process. For example, if there are “100” and “200” as the slot IOPS 402 during the target past replication process for a certain slot, the cumulative slot IOPS for that slot is “300” (= 100 + 200).

The cumulative volume IOPS 505 indicates the cumulative value of the volume IOPS during the target past replication process for the VOL including the slot that is the target of the host I / O process during the target past replication process. For example, if there are “200” and “400” as the volume IOPS 405 during the target past replication process for the PVOL 113A, the cumulative volume IOPS for the PVOL 113A is “600” (= 200 + 400).

Slot # 506 indicates the identification number of the slot that is the target of the host I / O process during the target past replication process.

In the target past replication process, when two or more slots are accessed, two or more slot numbers respectively corresponding to the two or more slots are registered as slot # 506, and two or more cumulative numbers corresponding to the two or more slots, respectively. The slot IOPS is registered as the cumulative slot IOPS 504. In the cumulative volume IOPS 505, the cumulative volume IOPS may be indicated for each of two or more VOLs. In this case, a VOL # may be associated with each cumulative volume IOPS. Information in the record of the past history table 302 is added by the duplicating unit 312 (post-processing unit 325). Also, the record referenced in the current replication process may be deleted by the replication unit 312 at the end of the current replication process.

FIG. 6 shows the configuration of the slot management table 303.

The slot management table 303 holds information regarding slots in the PVOL 113A. The slot management table 303 has a record for each slot in the PVOL 113A. Each record holds information such as area # 601, slot # 602, and difference flag 603. Hereinafter, one slot (referred to as “target slot” in the description of FIG. 6) is taken as an example.

Area # 601 indicates the identification number of the virtual area including the target slot. Slot # 602 indicates the identification number of the target slot. The difference flag 603 indicates whether or not the target slot is a difference slot. “1” means that the target slot is a differential slot. “0” means that the target slot is not a differential slot. In the present embodiment, it is assumed that the size of the virtual area and the size of the slot are constant.

When the slot subjected to the host I / O processing becomes a difference slot, the I / O unit 311 sets the difference flag 603 of the slot to “1”. If the slot is no longer a difference slot due to duplication processing, the duplication unit 312 sets the difference flag 603 of the slot to “0”.

FIG. 7 shows the configuration of the replication order table 304.

The replication order table 304 holds information related to the slot replication order in the PVOL 113A. The slot management table 303 has a record for each slot in the PVOL 113A. Each record holds information such as slot # 701 and duplication order 702. Hereinafter, one slot (referred to as “target slot” in the description of FIG. 7) is taken as an example.

Slot # 701 indicates the identification number of the target slot. The duplication rank 702 indicates the duplication rank of the target slot. If the target slot is not a difference slot, the value of the replication rank 702 is “Null”. In this embodiment, the smaller the value of the duplication rank 702, the higher the duplication rank. Examples of the “target slot is not a differential slot” include at least one of the following.
-The target slot has not been accessed (targeted for host I / O processing).
The target slot is set as the write destination during the current replication process (although it is set as the interrupt target), but the next slot is the difference slot.

FIG. 8 shows the configuration of the area management table 305.

The area management table 305 holds information on high-speed pages. The area management table 305 has a record for each high-speed tier. Each record holds information such as pool VOL # 801, upper map # 802, and usage flag 803. Hereinafter, a single high-speed page (referred to as “target high-speed page” in the description of FIG. 8) is taken as an example.

Pool VOL # 801 indicates the identification number of the pool VOL that includes the target high-speed page. The upper map # 802 indicates the identification number of the target high-speed page. The use flag 803 indicates whether or not the target high-speed page is secured. “0” means that the target high-speed page is not secured (for example, being accessed or already assigned to any virtual area). “1” means that the target high-speed page is secured.

FIG. 9 shows the configuration of the configuration management table 306.

The configuration management table 306 holds information regarding the configuration of the storage apparatus 202. For example, the configuration management table 306 may include a mapping table (not shown) that manages the allocation relationship between virtual areas and pages.

The configuration management table 306 includes a PDEV management table 316, for example. The PDEV management table 316 holds information regarding the PDEV type. The PDEV management table 316 has a record for each PDEV type. Each record holds information such as PDEV type 901, capacity 902, RAID type 903, tier # 904, device IOPS 905, and relocation throughput 906. Hereinafter, one target PDEV type (referred to as “target PDEV type” in the description of FIG. 9) is taken as an example.

The PDEV type 901 indicates the target PDEV type. The capacity 902 indicates the storage capacity of the PDEV of the target PDEV type. The RAID type 903 indicates the RAID type (RAID level and RAID configuration) of the RAID group including the PDEV of the target PDEV type. Tier # 904 indicates the identification number of the tier that includes the pool VOL based on the RAID group that includes the PDEV of the target PDEV type. The smaller the value of tier # 904, the higher (higher) tier. The device IOPS 905 is an example of the I / O performance of the target PDEV type PDEV, and indicates a processable IOPS (for example, maximum value). The relocation throughput 906 is an example of the I / O performance of the PDEV of the target PDEV type, and indicates the throughput (for example, the maximum value) when the PDEV becomes the data relocation source or the data relocation destination.

Hereinafter, the details of the duplication processing performed in this embodiment will be described with reference to FIGS.

FIG. 10 shows the flow of replication processing. The duplication processing starts, for example, periodically or when the duplication unit 312 receives a duplication processing start request from at least one of the host 101 and the management system 69. Even during the current replication process, the host I / O request is received, the host I / O process is executed, and the I / O monitor table 301 is updated.

The order determining unit 321 determines the order (S1001). That is, the order determination unit 321 determines the duplication order of all the difference slots in the PVOL 113A (duplication order of each difference slot) based on the past history table 302.

The influence determination unit 322 makes an influence determination (S1002). That is, the influence determination unit 322 determines whether or not the rearrangement from the low speed page to the high speed page decreases the host I / O performance during the current replication process.

When the determination result of S1002 is false (S1003: N), that is, when it is determined that the host I / O performance does not deteriorate, the area securing unit 323 secures the area (S1004). For example, the area securing unit 323 secures N (N is a natural number) free high-speed pages. N is the number of free high speed pages required.

The relocation unit 313 is called. The rearrangement unit 313 rearranges (moves) the data in the low-speed page allocated to the virtual area including the difference slot to the high-speed page secured in S1004 (S1005).

The control unit 326 determines based on the I / O monitor table 301 whether a host I / O process has occurred during the current replication process (S1006). When the determination result in S1006 is false (S1006: N), the control unit 326 performs data replication (S1008).

If the determination result in S1006 is true (S1006: Y), the order update unit 324 updates the order (S1007). That is, the order update unit 324 updates the replication order determined in S1001 based on the host I / O process history during the current replication process.

The control unit 326 performs data replication according to the replication order indicated by the replication order table 304 (S1008). Data replication is performed for each of all the differential slots in the PVOL 113A. Data duplication is performed in descending order of duplication. For example, the control unit 326 selects the difference slot with the highest replication order among the difference slots for which data replication is not performed, reads data from the selected difference slot to the CM 111, and stores the data in the CM 111 as the difference slot. Is written in the slot in the SVOL 113B corresponding to The control unit 326 repeats this process until there is no difference slot for which data replication is not performed. Specifically, the replication order in S1008 is as follows.
In S1001, if the difference slot that is the replication source in the current replication process does not include the past I / O destination slot (difference slot that was set as the I / O destination in the past replication process), the replication slot in S008 The order is from the smallest slot # of the difference slot. In other words, in this case, the smaller the slot #, the higher the replication order.
If the determination result in S1002 is true (S1003: Y), that is, if it is determined that the host I / O performance is degraded, the replication order in S1008 is the replication order determined in S1001.
In S1007, if no I / O destination slot in the current replication process is a past I / O destination slot, the replication order in S1008 is the replication order determined in S1001.
In S1007, if at least one I / O destination slot in the current replication process is a past I / O destination slot, in S1008, the replication order is the replication order after the update in S1007.

After S1008, the post-processing unit 325 performs post-processing (S1009). That is, the post-processing unit 325 uses the monitoring result of the host I / O process during the current replication process (the result specified from the I / O monitor table 301) as the history of the host I / O during the past replication process. Register in the history table 302.

FIG. 11 shows the details of the sequence determination (S1001 in FIG. 10).

The order determination unit 321 registers the current time as the time 401 and “0” (start) as the status 404 in the record of the I / O monitor table 301. The record is the first record for the current replication process (S1101).

The order determining unit 321 refers to the slot management table 303 and identifies a difference slot in the PVOL 113B (if there is no difference slot, the current replication process may be canceled). The order determining unit 321 determines whether or not all the identified difference slots include at least one past I / O destination slot based on the past history table 302 (S1102). If slot # 506 in the record corresponding to the past duplication processing does not include slot # of any difference slot, the determination result in S1102 is false. If the determination result in S1102 is false (S1102: N), the process proceeds to S1008 in FIG.

When the determination result of S1102 is true (S1102: Y), that is, when slot # 506 in the record corresponding to the past duplication processing includes slot # of at least one difference slot (that is, at least one difference) When the slot is a past I / O destination slot), the order determination unit 321 determines the cumulative slot IOPS for each of the corresponding difference slots (hereinafter, past I / O destination difference slot), and the cumulative slot IOPS 504 in the past history table 302. (S1103).

The order determination unit 321 sets the duplication rank of the past I / O destination difference slot among all the difference slots higher than the duplication rank of the difference slot that is not the past I / O destination difference slot among all the difference slots. At that time, the duplication order of the difference slot that is not the past I / O destination difference slot may be the order of the smallest slot #. In addition, the order determining unit 321 sets the duplication order of the past I / O destination difference slot in the descending order of the cumulative slot IOPS (S1104).

The order determining unit 321 registers the duplication rank of all the difference slots in the duplication rank table 304 (S1105).

FIG. 12 shows details of the flow of influence determination (S1002 in FIG. 10).

The influence determination unit 322 specifies the replication processing time 503 and the cumulative volume IOPS 505 of the PVOL 113A for the past replication processing based on the past history table 302 (S1201). Further, the influence determination unit 322 identifies the current volume IOPS of the PVOL 113A based on the I / O monitor table 301 (S1202). The “current volume IOPS” may be a representative volume IOPS (for example, a maximum value or an average value) identified from the I / O monitor table 301.

The influence determination unit 322 calculates the upper free capacity based on the configuration management table 306 (S1203). The above-described mapping table in the configuration management table 306 indicates, for example, for each page, a tier # to which the page belongs and a page status indicating whether or not the page is free. “Upper free capacity” is the total capacity of free high-speed pages.

The influence determination unit 322 calculates the usage amount during replication (S1204). The “utilizing amount during copying” is the maximum value of the total capacity of free high-speed pages that are considered to be newly allocated to the PVOL 113A during the current copying process. For example, the usage during replication = current volume IOPS × I / O size × duplication processing time. “Current volume IOPS” is the value specified in S1202. The “I / O size” is a data amount (for example, slot size) input / output by one I / O. The “duplication processing time” is the duplication processing time 503 specified in S1201.

The influence determination unit 322 determines whether or not upper free space ≦ use amount during replication (S1205). If the determination result in S1205 is true (S1205: Y), the process proceeds to S1008 in FIG. If the upper free capacity is equal to or less than the used capacity during replication, the reason for S1205: Y is that if all the host I / O processing targets during the current replication process are new virtual areas, the allocation is made to that virtual area. This is because there may be a shortage of possible high-speed pages (that is, high host I / O performance may not be achieved).

When the determination result in S1205 is false (S1205: N), the influence determination unit 322 calculates rearrangement IOPS (S1206). “Relocation IOPS” is an example of IOPS consumed by data relocation. For example, rearrangement IOPS = relocation throughput / I / O size. The “relocation throughput” is the PDEV relocation throughput 906 (see FIG. 9) that is the basis of the high-speed tier. The “I / O size” is a data amount (for example, slot size) input / output by one I / O.

The influence determination unit 322 calculates effective IOPS (S1207). “Valid IOPS” is an IOPS that can be processed during data relocation. For example, valid IOPS = device IOPS−relocation IOPS. “Device IOPS” is a PDEV device IOPS 905 (see FIG. 9) that is the basis of the high-speed tier. “Relocation IOPS” is the value calculated in S1206.

The influence determination unit 322 calculates the average volume IOPS (S1208). “Average volume IOPS” is the average of IOPS of PVOL 113A. For example, average volume IOPS = cumulative volume IOPS ÷ duplication processing time. Both “cumulative volume IOPS” and “duplication processing time” are the values specified in S1201.

The influence determination unit 322 determines whether or not effective IOPS ≦ average volume IOPS (S1209). If the determination result in S1209 is true (S1209: Y), the process proceeds to S1008 in FIG. When effective IOPS ≦ average volume IOPS, the reason for S1209: Y is that the IOPS (host I / O performance) required for the PVOL 113A cannot be satisfied by performing data relocation from the low speed page to the high speed page. Because you get.

If the determination result in S1209 is false (S1209: N), the influence determination ends.

In this embodiment, if both the determination results of S1205 and S1209 are false, it is a determination result that the rearrangement from the low-speed page to the high-speed page does not deteriorate the host I / O performance during the current replication process. At least one of S1205 and S1209 is an example of determining whether or not the rearrangement from the low-speed page to the high-speed page lowers the host I / O performance during the current replication process. S1209 may be performed before S1205. Further, when at least one determination result of S1205 and S1209 is false, it may be a determination result that the host I / O performance does not deteriorate.

FIG. 13 shows details of the flow of area reservation (S1003 in FIG. 10).

The area securing unit 323 refers to the top record of the area management table 305 (S1301).

The area securing unit 323 determines whether or not the current I / O has occurred (already allocated) on the high-speed page corresponding to the reference record, for example, by referring to the mapping table described above (S1302).

When the determination result in S1302 is false (S1302: N), the area securing unit 323 sets the use flag 803 in the reference record to “1” (usable) (S1303). As a result, a free high-speed page corresponding to the reference destination record is secured.

On the other hand, if the determination result in S1302 is true (S1302: Y), the area securing unit 323 sets the use flag 803 in the reference record to “0” (unusable) (S1304).

After S1303 or S1304, the area securing unit 323 determines whether the number of secured free high-speed pages has reached the required number N (N is a natural number) (S1305). Here, the required number may be one of the following. Whether the required number is one of the following may be determined in advance, or the area securing unit 323 may be the maximum of the number of free high-speed pages (corresponding to the upper free capacity in S1203 in FIG. 12) or less. May be selected.
The number of virtual areas that contain all differential slots.
-Number of virtual areas including all past I / O destination difference slots.
The number of virtual areas including the past I / O destination difference slot having the highest replication order among the virtual areas including all the past I / O destination difference slots.

When the determination result of S1305 is false (S1305: N), the area securing unit 323 increments the position (number) of the reference destination record (that is, the next record is the reference destination) (S1306), and the next record S1302 is performed.

If the determination result in S1305 is true (S1305: Y), the area reservation is completed.

FIG. 14 shows the details of the sequence update (S1007 in FIG. 10).

The order updating unit 324 determines whether or not at least one current I / O destination difference slot is a past I / O destination difference slot (S1401). If the determination result in S1401 is false (S1401: N), the process proceeds to S1008 in FIG. Note that the “current I / O destination difference slot” is a difference slot (targeted as an I / O destination) that is a target of host I / O processing during the current replication processing.

If the determination result in S1401 is true (S1401: Y), the order update unit 324 specifies a difference slot that can be related to the current I / O destination difference slot corresponding to S1401: Y (S1402). The “difference slot that can be related to the current I / O destination difference slot” may be, for example, a difference slot that stores data associated with data in the current I / O destination difference slot.

The order update unit 324 relatively raises the replication rank of at least one of the current I / O destination difference slot corresponding to S1401: Y and the difference slot specified in S1402 (S1403). As a result, among the difference slots corresponding to the past I / O destination difference slot, the difference slot corresponding to the current I / O destination difference slot is set to a higher replication order than the replication order determined in S1001. The order update unit 324 registers the replication order after the update in S1403 in the replication order table 304.

The order update unit 324 assigns at least one of the current I / O destination difference slot corresponding to S1401: Y and the difference slot specified in S1402 to the virtual area including the difference slot. It is determined whether the page is a low speed page (S1404). If the determination result in S1404 is true (S1404: Y), the rearrangement unit 313 rearranges the data in the low-speed page to an empty high-speed page (S1405). As a result, the replication target data in the difference slot whose replication order has been increased exists in the high-speed PDEV at the time of data replication.

FIG. 15 shows details of the flow of post-processing (S1009 in FIG. 10).

The post-processing unit 325 registers status 404 “1” (end) in the final record corresponding to the current replication process in the I / O monitor table 301 (S1501).

The post-processing unit 325 refers to the first record corresponding to the current replication process in the I / O monitor table 301 (S1502).

The post-processing unit 325 determines whether or not the status 404 in the referenced record is “0” (S1503).

If the determination result in S1503 is true (S1503: Y), the referenced record is the first record. The post-processing unit 325 stores the time 401 in the record in the SM 113 (for example, the work area) as the start time, and stores the slot IOPS, slot #, and volume IOPS in the record in the SM 113 (S1504).

If the determination result in S1503 is false (S1503: N), the post-processing unit 325 determines whether the status 404 in the referenced record is “1” (S1505).

If the determination result in S1505 is false (S1505: N), the referenced record is a record between the first record and the last record. The post-processing unit 325 reflects the slot IOPS, slot #, and volume IOPS in the record on the slot IOPS, slot #, and volume IOPS already stored in the SM 113, respectively. That is, the accumulated slot IOPS for each difference slot is updated, and the accumulated volume IOPS for the PVOL 113A is updated. Thereafter, the post-processing unit 325 increments the position (number) of the reference destination record (that is, the next record is the reference destination) (S1507), and performs S1505 for the next record.

If the determination result in S1505 is true (S1505: Y), the referenced record is the last record. The post-processing unit 325 stores the time 401 in the record as the end time in the SM 113 (for example, a work area), and additionally stores the slot IOPS, slot #, and volume IOPS in the record in the slot IOPS, slot # already stored in the SM 113. And reflected in the volume IOPS (S1508).

At the end of S1508, the start time, end time, accumulated slot IOPS of each difference slot, and accumulated volume IOPS of PVOL 113A are fixed. The post-processing unit 325 adds a start time 501, an end time 502, a copy processing time 503 (difference between the end time and the start time), a cumulative slot IOPS to the record of the past history table 302 (a record corresponding to the current copy process). (Accumulated slot IOPS of each difference slot), accumulated volume IOPS 505 (at least the accumulated volume IOPS of PVOL 113A), and slot # 506 (list of difference slot #) are registered (S1509). In the next replication process, the record having the information registered in S1509 is referred to.

As mentioned above, although one embodiment was described, this is an illustration for explaining the present invention, and the scope of the present invention is not limited to this embodiment. The present invention can be implemented in various other forms. For example, the PVOL 113A may not be a VVOL.

102 ... Storage system

Claims

A plurality of storage devices based on a plurality of logical storage areas and having different I / O performance;
A memory unit that is one or more memories including a cache memory;
Providing at least a primary volume of the primary volume and the secondary volume, performing host I / O processing on the primary volume in response to a host I / O request specifying an area in the primary volume, A processor unit that manages a secondary volume and a differential area that is a differential area;
The plurality of storage devices are:
One or more first storage devices;
One or more second storage devices having lower I / O performance than the one or more first storage devices;
At least a part of the plurality of logical storage areas is an area allocated to at least a write destination area of at least the primary volume,
The plurality of logical storage areas are:
A plurality of first logical storage areas provided by the one or more first storage devices;
A plurality of second logical storage areas provided by the one or more second storage devices,
The processor unit performs a replication process including data replication, and the data replication is performed from a logical storage area allocated to the differential area for each of one or more differential areas of the primary volume. A process of reading the data to the cache memory and transferring the data to copy the data to an area corresponding to the differential area of the secondary volume;
The processor unit manages a host I / O history that is a history of host I / O processing,
The host I / O history includes an I / O destination area in the primary volume and a host I / O processing time for each I / O destination area,
When the primary volume has at least one difference area, the processor unit performs the current replication process including (A) below.
(A) An order determination process including the following (a1) and (a2) is performed.
(A1) It is determined whether there is one or more past I / O destination areas among all the differential areas of the primary volume.
Each of the one or more past I / O destination areas is an I / O destination area during past duplication processing and is an area specified from the host I / O history.
(A2) When the determination result of (a1) is true, the replication order of all past I / O destination areas of all the difference areas is set to other than all past I / O destination areas of the all difference areas. Make it higher than the replication order of the difference area,
Storage system.
In (a2), the replication order of all past I / O destination areas is in descending order of the host I / O amount in the past replication processing.
The storage system according to claim 1.
For each of all the past I / O destination areas, the host I / O amount is a cumulative value of the host I / O frequency for the past I / O destination area during the past replication process.
The storage system according to claim 2.
The processor unit further performs the following (B) and (C) in the current replication process.
(B) When there is one or more difference areas to which at least a part of all the difference areas and one or more second logical storage areas are assigned, data is transferred from the second logical storage area to the first logical storage area. To determine whether or not rearranging will reduce the host I / O performance during the current replication process,
(C) When the judgment result of (B) is false, data is transferred from the second logical storage area allocated to the differential area to the free first logical storage area for at least one of the one or more differential areas. And assigning the first logical storage area to the difference area instead of the second logical storage area,
The storage system according to claim 1.
When it corresponds to at least one of the following (b1) and (b2), the judgment result of (B) is false.
(B1) The following (b11) is more than the following (b12),
(B11) the total capacity of the free first logical storage area,
(B12) A used amount during replication that is the maximum value of the total capacity of the free first logical storage area that is considered to be newly allocated to the primary volume during the current replication process;
(B2) The following (b21) is more than the following (b22),
(B21) Host I / O amount per unit time for the primary volume,
(B22) The amount of host I / O that can be processed per unit time during data relocation,
The storage system according to claim 4.
(B12) is an amount based on the following (b121) and (b122) specified from the I / O history.
(B121) Duplication processing time of the past duplication processing,
(B122) Host I / O amount per unit time for the primary volume,
The storage system according to claim 5.
(B21) is an amount based on the following (b211) and (b212) specified from the I / O history,
(B211) Duplication processing time of the past duplication processing,
(B212) Host I / O amount per unit time for the primary volume during the past replication processing,
(B22) is an amount based on the following (b221) and (b222).
(B221) The amount of I / O that can be processed per unit time by the one or more first storage devices,
(B222) I / O amount consumed by the data relocation,
The storage system according to claim 5.
In (C), the processor unit preferentially sets a difference area having a relatively high replication order determined in (A) as a data relocation target.
The storage system according to claim 4.
The processor unit further performs the following (D) in the current replication process:
(D) When the determination result of (B) is false, an order update process including the following (d1) and (d2) is performed.
(D1) Based on the host I / O history, it is determined whether or not the I / O destination area in the host I / O process during the current replication process is any past I / O destination area. ,
(D2) When the determination result of (d1) is true, the replication order of the past I / O destination area as the I / O destination area, which is relatively higher than the replication order determined in (A) ,
The storage system according to claim 4.
The order update process further includes the following (d3) and (d4),
(D3) It is determined whether or not the second logical storage area is allocated to the past I / O destination area for which the determination result of (d1) is true.
(D4) When the determination result of (d3) is true, for the past I / O destination area, data is transferred from the second logical storage area allocated to the past I / O destination area to the free first logical storage area. Rearrange and assign the first logical storage area to the past I / O destination area instead of the second logical storage area.
The storage system according to claim 9.
The processor unit is
If the determination result in (B) is true, data replication is performed in the replication order determined in (A),
If the determination result in (B) is false, data replication is performed in the replication order updated in (D).
The storage system according to claim 9.
In (C), the empty first logical storage area of the data rearrangement destination is an empty first logical storage area secured by the processor unit.
The storage system according to claim 4.
The host I / O history includes a history of host I / O processing during past replication processing,
The processor unit further performs the following (E):
(E) The host I / O processing history in the current replication process is newly reflected in the host I / O history as the past replication processing host I / O processing history.
The storage system according to claim 1.
Each difference area is a unit of data replication,
Each logical storage area is a unit of allocation and is a unit larger than the unit of data replication.
The storage system according to claim 1.
A replication control method in a storage system,
The storage system
A plurality of storage devices based on a plurality of logical storage areas and having different I / O performance;
A memory unit that is one or more memories including a cache memory;
The storage system provides at least a primary volume of a primary volume and a secondary volume, performs host I / O processing for the primary volume in response to a host I / O request specifying an area in the primary volume, and It is designed to manage a difference area that is a difference between the secondary volume and the primary volume,
The plurality of storage devices are:
One or more first storage devices;
One or more second storage devices having lower I / O performance than the one or more first storage devices;
At least a part of the plurality of logical storage areas is an area allocated to at least a write destination area of at least the primary volume,
The plurality of logical storage areas are:
A plurality of first logical storage areas provided by the one or more first storage devices;
A plurality of second logical storage areas provided by the one or more second storage devices,
The storage system performs replication processing including data replication, and the data replication is performed for each of one or more differential areas of the primary volume from a logical storage area allocated to the differential area. A process of reading the data to the cache memory and transferring the data to copy the data to an area corresponding to the differential area of the secondary volume;
The replication control method includes:
A host I / O history, which is a history of host I / O processing, is managed, and the host I / O history includes the I / O destination area in the primary volume and the host I / O processing time for each I / O destination area. Including
If the primary volume has at least one difference area, the current replication process including (A) below is performed.
(A) An order determination process including the following (a1) and (a2) is performed.
(A1) It is determined whether or not there is one or more past I / O destination areas among all the differential areas of the primary volume. Each of the one or more past I / O destination areas is in a past replication process. An I / O destination area that is specified from the host I / O history.
(A2) When the determination result of (a1) is true, the replication order of all past I / O destination areas of all the difference areas is set to other than all past I / O destination areas of the all difference areas. Make it higher than the replication order of the difference area,
Replication control method.