WO2015132946A1

WO2015132946A1 - Storage system and storage system control method

Info

Publication number: WO2015132946A1
Application number: PCT/JP2014/055899
Authority: WO
Inventors: 森　宣仁; 智大川口; アビシェクジョーリ; 山本　彰; 匡人仁科
Original assignee: 株式会社日立製作所
Priority date: 2014-03-07
Filing date: 2014-03-07
Publication date: 2015-09-11

Abstract

When a volume to which a snapshot configuration is applied is migrated from a movement-source storage to a movement-destination storage, a problem of discrepancy in snapshot configuration between the movement source and the movement destination may be caused upon accepting a write from a host. According to the present invention, in order to solve the problem, during migration of a snapshot configuration, an exclusive of a chunk to be migrated is acquired, and snapshot allocate information, which is management information for the chunk to be transitioned, is transmitted to the movement-destination storage. Upon reception of a write request from an higher-level device, an exclusive of a first chunk relating to the write request and a second chunk related to the chunk relating to the write request by the snapshot configuration is acquired, and a write process relating to the write request is executed.

Description

Storage system and storage system control method

The present invention relates to a storage system.

The storage device has a pool volume and a secondary volume that store the difference data between the primary volume and the primary volume. When there is an update to the primary volume, the data before the update of the primary volume is saved to the pool volume. The secondary volume has a function called a snapshot function in which each chunk refers to a primary volume or a chunk of a pool volume, thereby constituting a point-in-time copy of the primary volume.

In Patent Document 1, when the migration source storage forms a snapshot, a plurality of data of the migration source primary volume and a plurality of data of the pool volume are sent to the migration destination storage in units of volumes, and the chunk of the secondary volume is also sent. A technique for moving a volume adapted for a snapshot function to a transfer destination storage by sending position information of a chunk to be referenced in units of volumes is disclosed.

US Patent Publication No. 2013/0297899

However, Patent Document 1 does not disclose executing a write from the host to the migration target volume during the migration of the volume to which the snapshot function is applied. When copying the snapshot configuration of the migration source storage to the migration destination storage, there is a problem that if the write is received from the host, the snapshot configuration will be different between the migration source and the migration destination. In particular, in recent years, storage device virtualization technology has been developed that provides a host with one or more virtual storages across a plurality of storage devices without making the host aware of the physical configuration of the storage device. When migrating a volume by host transparency, there is an increasing need to accept a write from the host during the movement of the volume, and it is necessary to solve the above problem also in this respect.

In addition, an enterprise system is required to have a high availability of an operation rate of 99.999% (down for about 5 minutes per year). For example, assuming that a financial institution performs volume transfer using the snapshot function, if the volume cannot be written during the transition, the financial institution's service that is the basis of economic activity will be suspended. It will interfere with the economic activities of society as a whole. From this point, it is necessary to solve the above problem.

In order to solve the above problems, for example, the configuration described in the claims is adopted.

The present application includes a plurality of means for solving the above-described problems. For example, the following storage system is used.
The storage system includes a first storage device and a second storage device. The first storage device provides a snapshot configuration with the first primary volume, the first pool volume, and the first secondary volume. The first primary volume includes a plurality of chunks. The first pool volume includes a plurality of chunks and stores differential data of the first primary volume. The first secondary volume is a replica of the first primary volume at a predetermined point in time, and includes a plurality of chunks each referring to either the first primary volume chunk or the first pool volume chunk. The first storage device further manages reference destinations of the plurality of chunks of the first secondary volume. The second storage device provides a snapshot configuration with the second primary volume, the second pool volume, and the second secondary volume. The second primary volume includes a plurality of chunks. The second pool volume includes a plurality of chunks and stores differential data of the second primary volume. The second secondary volume is a replica of the second primary volume at a predetermined point in time, and includes a plurality of chunks each referring to either the second primary volume chunk or the second pool volume chunk. The second storage device further manages the reference destinations of the plurality of chunks of the second secondary volume. The first storage device migrates the snapshot configuration from the first storage device to the second storage device. At that time, the exclusion of the chunk to be migrated is acquired, and snapshot allocation information, which is management information of the chunk to be migrated, is transmitted to the second storage device. When a write request is received from the host device, the exclusive of the first chunk related to the write request, the chunk related to the write request, and the second chunk related by the reference is acquired, and the write request related to the write request is acquired. Execute the process.

According to an embodiment of the present invention, it is possible to execute writing and reading from a host while moving a volume to which a snapshot function is applied between storage apparatuses. As a result, it is possible to perform storage device replacement, load balancing, and redundancy while maintaining the snapshot function without host downtime.

It is a conceptual diagram which shows an example of the movement process of the snapshot volume in the 1st Embodiment of this invention. It is a figure which shows an example of the IT system structure in the 1st Embodiment of this invention. It is a figure which shows an example of the system configuration | structure of the host computer and management server in the 1st Embodiment of this invention. It is a figure which shows an example of the system configuration | structure of the storage apparatus in the 1st Embodiment of this invention. It is a figure which shows an example of the virtual volume management information and virtual storage box management information in the 1st Embodiment of this invention. It is a figure which shows an example of a structure of snapshot PVOL, SVOL, and a pool in the 1st Embodiment of this invention. It is a figure which shows an example of the snapshot pair management information in the 1st Embodiment of this invention, snapshot PVOL management information, and snapshot SVOL management information. It is a figure which shows an example of the pool management information in the 1st Embodiment of this invention. It is a figure which shows an example of the movement copy pair management information in the 1st Embodiment of this invention. It is a figure which shows an example of the snapshot allocation information in the 1st Embodiment of this invention. It is a flowchart which shows an example of the volume movement process in the 1st Embodiment of this invention. It is a flowchart which shows an example of the movement copy process in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the snapshot allocation information reception from the migration source storage apparatus in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host write reception to the movement origin VOL in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the write and snapshot allocation information reception from the migration source storage apparatus in the 1st Embodiment of this invention. It is a flowchart which shows an example of the data saving process in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host write reception to the movement destination VOL in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the write reception from the migration destination storage apparatus in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the read reception to the movement origin VOL in the 1st Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host read reception to the movement destination VOL in the 1st Embodiment of this invention. It is a flowchart which shows an example of the movement copy process in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host write reception to the movement origin VOL in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the write reception from the movement origin storage apparatus in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host write reception to the movement destination VOL in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the write reception from the migration destination storage apparatus in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the movement copy process in the 3rd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of host write reception to the movement origin VOL in the 3rd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the host write reception to the movement destination VOL in the 3rd Embodiment of this invention. It is a flowchart which shows an example of the process at the time of the write reception from the migration destination storage apparatus in the 3rd Embodiment of this invention.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below do not limit the invention according to the claims, and all the elements and combinations described in the embodiments are not necessarily essential to the solution of the invention. Absent. In the following description, the processing may be described with the storage device as the subject. However, by executing the processing by the controller, the program, or the CPU of the storage device, the determined processing is appropriately performed with a storage resource (for example, a memory). ) And / or communication interface devices (for example, ports), the processing subject may be a program. The program source may be a program distribution server or a storage medium, for example.

A first embodiment according to the present invention will be described with reference to FIGS.

FIG. 1 is a diagram showing an outline of this embodiment. A real volume 520 is set in the migration source storage apparatus 500A, and a snapshot function is applied to each real volume 520. The real volume 520A is a snapshot PVOL (hereinafter sometimes simply referred to as PVOL), and data is actually stored in a plurality of chunks included in the PVOL. The

real volumes

520B and 520C are snapshot SVOLs (hereinafter sometimes simply referred to as SVOLs). The SVOL manages PVOL data at a certain point as a difference. That is, the SVOL has information such as a pointer to a chunk in the pool 530 in which PVOL chunks and differential data are actually stored.

In the migration destination storage apparatus 500B,

real volumes

520D, 520E, and 520F that are migration destinations of the respective real volumes 520 in the migration source storage apparatus 500A are set, and the snapshot function is applied similarly to the migration source storage. ing. Hereinafter, the volume of the migration source storage apparatus 500A may be referred to as a migration source VOL. Also, when distinguishing PVOL and SVOL, they may be referred to as migration source PVOL and migration source SVOL, respectively. Similarly, the real volume 520 of the migration destination storage apparatus 500B may be described as a migration destination VOL, a migration destination PVOL, and a migration destination SVOL.

The migration source storage apparatus 500A and the migration destination storage apparatus 500B execute the migration of the volume to which the snapshot function is applied by copying data and pointers from the migration source VOL to the migration destination VOL by the migration copy process described later. To do. The host computer 100 has a function of providing one or more virtual storage apparatuses 600 across a plurality of storage apparatuses 500 to the host without making the host computer 100 aware of the physical configuration of the storage apparatus.
The migration source VOL and the migration destination VOL belong to the same virtual storage apparatus 600, and corresponding volumes are set with the same virtual volume ID. That is, the same virtual volume ID is set for each of the

real volumes

520A and 520D, 520B and 520E, 520C and 520F. The virtual volume ID is a virtual identifier set separately from the identifier set for the real volume 520.

When the alternate path software 111 on the host computer 100 requests an identifier for the real volume 520, the storage apparatus 500 obtains an identifier calculated from the virtual storage ID that is the identifier of the virtual storage apparatus 600 and the virtual volume ID. respond. That is, the

real volumes

520A and 520D, 520B and 520E, 520C and 520F respectively respond with the same identifier. The alternate path software 111 recognizes the real volume 520 that responds with the same identifier as an alternate path to one virtual volume 601, and processes I / O (Input / Output) from the OS 110.

Under such a virtual storage environment, the host does not need to recognize the virtual storage apparatus 600 and the

storage apparatuses

500A and 500B that constitute the virtual storage apparatus 600. Therefore, the storage apparatus 500 can be added, replaced, or reconfigured transparently to the host computer 100 without stopping the business.
Thus, during the migration copy process, the OS 110 can issue an I / O without distinguishing the migration source VOL and the migration destination VOL. Further, even if the migration source VOL is deleted when the migration is completed, the I / O can be issued to the migration destination VOL continuously, so that the operation of the OS 110 is not affected.

The migration copy is, for example, replacement from the migration source storage apparatus 500A to the migration destination storage apparatus 500B, load balancing by moving a part of the snapshot configuration of the migration source storage apparatus 500A to the migration destination storage apparatus 500B, and migration source storage apparatus This is performed for redundancy by duplicating the snapshot function between A and the migration destination storage apparatus B.

S0 to S7 show processing when the migration source storage apparatus 500A receives I / O from the host computer 100 during the migration copy process.
S0: Start moving copy processing for each chunk.
S1: A write is issued from the host computer 100 to the real volume 520A.
S2: The migration source storage apparatus 500A receives the write, and excludes the migration copy process for the write target address of the SVOL (real volume 520B) where the pointer is rewritten when the real volume 520A is updated. The address is a numerical value indicating the position of the chunk on the real volume 520.
S3: The migration source storage apparatus 500A executes data update by writing to the real volume 520A.
S4: The migration source storage apparatus 500A stores the data before the write in S3 in the pool 530, and changes the pointer of the SVOL (real volume 520B) to the storage destination address. Note that storing data in the pool 530 and changing the pointer (address) of the SVOL in this way may be referred to as saving.
S5: The migration source storage apparatus 500A transmits snapshot allocation information including information such as a write and a pointer of SVOL (real volume 520B).
S6: The migration destination storage apparatus 500B reflects the write and snapshot allocation information transmitted in S5.
S7: The migration destination storage apparatus 500B executes evacuation required as a result of the write in S6.

FIG. 2 is a block diagram showing an outline of the IT system configuration. This system includes a host computer 100, a management server 200, and one or more storage devices 500, which are connected to each other via a LAN 400. The storage apparatus 500 and the host computer 100 are connected to each other via the SAN 300. A plurality of host computers may exist. In this case, the host computers may be connected to each other via a data transfer LAN in addition to the management LAN 400.

FIG. 3A shows the internal configuration of the host computer 100. The host computer 100 includes one or more CPUs 101, one or more memories 102, one or more SAN interfaces 105, one or more LAN interfaces 106, and one or more storage devices 103, which are connected to each other by an internal bus. 107 is connected. The host computer 100 is connected to the SAN switch 400 via the SAN interface 105. The host computer 100 is connected to the LAN switch 400 via the LAN interface 106. Note that the storage device 103 is not necessarily provided. If not, the volume in the storage device 500 is used as a software storage area.

FIG. 3B shows the internal configuration of the management server 200. The internal configuration of the management server 200 is the same as that of the host computer 100, but does not necessarily have a SAN interface. The management server 200 is connected to the LAN switch 400 via the LAN interface 204.

FIG. 4 shows the internal configuration of the storage apparatus 500. The storage apparatus 500 includes one or more controllers 540 and one or more physical disks 501. The controller 540 includes one or more CPUs 506, one or more memories 508, one or more nonvolatile memories 505, one or more cache memories 507, one or more back-end interfaces 509, and one or more LAN interfaces 504. , One or more SAN interfaces 502 are connected to each other via an internal bus 541. The controller 540 is connected to the physical disk 501 via the back end interface 509.

In addition, the storage apparatus 500 is connected to the management server 200 via the LAN interface 504. The storage apparatus 500 is connected to the host computer 100 via the SAN interface 502. In addition, the physical disk 501 is a SAS (Serial Attached SCSI System) disk, a SATA (Serial Advanced Technology) disk, or a SSD (Solid State) disk type. Note that the PVOL in the real volume 520 shown in FIG. 1 is an area composed of RAID (Redundant Arrays of Inexpensive Disks) using one or more physical disks 501 in many cases. You may implement by techniques, such as Thin Provisioning, which allocates an area | region dynamically.

Also, the SVOL is composed of references to PVOL or pool 530 chunk addresses. The pool 530 is implemented as a RAID area composed of one or more physical disks 501 or a group of one or more real volumes 520 composed of RAID.

In the latter case, the controller 540 treats one or more chunks of the real volume 520 as serial numbers (addresses) so as to treat them as one volume. The memory 508 stores a resource management program 550, a snapshot control program 551, and a migration copy control program 552, and the CPU 506 executes these programs. The resource management program 550 manages the inclusion relationship between the virtual storage apparatus 600 and volumes by referring to and changing information such as virtual volume management information 610 and virtual storage box management information 700 described later.

Further, the snapshot control program 551 refers to and changes information such as snapshot pair management information 800, snapshot PVOL management information 900, snapshot SVOL management information 1000, and pool management information 1100, which will be described later. In addition, processing related to snapshot control such as read and write processing is performed. The migration copy control program 552 executes later-described migration copy processing by referring to and changing the migration copy pair management information 1200 described later and information related to the snapshot.

FIG. 5A shows an example of the virtual volume management information 610. This information is information held by each storage device 500 and is mainly stored in the memory 508. The real volume ID column 611 stores the identifier of the real volume 520. The real volume ID is an identifier for specifying the real VOL 520 in the storage apparatus 500. The resource group ID column 612 stores the identifier of the resource group to which the real volume 520 belongs. The resource group is a unit for grouping the real volumes 520 and has a one-to-N relationship with a storage box described later.

For example, by setting an accessible user in the resource group, it is possible to realize access control in which only a specific user can handle the real volume 520 in the resource group and prohibit access by other users. The virtual volume ID column 613 stores the second identifier set for the real volume 520. The virtual volume ID is an identifier associated with the real volume ID, and the storage apparatus 500 returns the virtual volume ID when requested by the host computer 100 for the identifier.

FIG. 5B shows an example of the virtual storage box management information 700. This information is information held by each storage device 500. The virtual storage box ID column stores the identifier of the virtual storage box. The virtual storage box is a part for each storage device 500 of the virtual storage device 600. That is, the substance of the virtual storage device 600 is a set of virtual storage boxes in which the same virtual storage ID is set in each storage device 500.

In the virtual storage ID column 702, the identifier of the virtual storage device 600 is stored. The resource group ID column 703 stores resource group IDs belonging to the virtual storage box. The real volume 520 can identify the virtual storage box and virtual storage device 600 to which the real volume 520 belongs by combining the virtual volume management information 610 and the virtual storage box management information 700.

FIG. 6 is a diagram showing the concept of the snapshot PVOL (real volume 520A), SVOL (

real volumes

520B, 520C), and pool 530. Each of the PVOL, SVOL, and pool is composed of one or more chunks. Data stored in each chunk of SVOL is managed by any one of PVOL reference, shared reference, and exclusive reference. The PVOL reference refers to a chunk having the same address in the PVOL. When a read is issued from the host computer 100 to the SVOL PVOL reference chunk, the storage apparatus 500 responds with the data of the chunk of the same address in the PVOL.

The shared reference refers to a chunk in the pool 530. When a read is issued from the host computer 100 to a shared reference chunk of the SVOL, the storage apparatus 500 responds with the data of the chunk of the pool 530 referenced by the chunk of the SVOL. An exclusive reference refers to a chunk in the pool 530, as in the case of a shared reference. Shared references and exclusive references are distinguished by the trigger of saving. When a write occurs in a chunk with a PVOL, the storage apparatus 500 stores the old data before the write execution in the pool 530 for the chunk with the same address in the SVOL that references the PVOL, and then refers to it (pointer) Is changed to the address of the save destination pool.

In this way, when referring to data saved when a write occurs in PVOL, it is called a shared reference. This is because, in many cases, a plurality of SVOLs share the saved data. On the other hand, when a write occurs in a chunk having an SVOL, the storage apparatus 500 stores the write data in an unallocated chunk 531 of the pool 530 and stores the write data as the reference destination of the chunk in the SVOL. Change to chunk. In this way, the case of referring to data saved when a write occurs in the SVOL is referred to as an exclusive reference. This is because the saved data becomes data specific to the SVOL. The transition between the PVOL reference, the shared reference, and the exclusive reference in the write opportunity is unidirectional, and the PVOL reference changes to the shared reference or the exclusive reference, and the shared reference to the exclusive reference, but there is no other transition.

FIG. 7A shows an example of the snapshot pair management information 800. This information is information held by each storage device 500 and is mainly stored in the memory 508. A snapshot pair is a set of PVOL and SVOL that constitute a snapshot. One PVOL is composed of one or more SVOLs, each of which is a copy of the PVOL at a different point in time. The snapshot pair ID column 801 stores snapshot pair identifiers. The pair status column 802 stores the status of the snapshot pair. The pair status is any one of PAIR, COPY, PSUS, and SMPL. PAIR is a state in which PVOL and SVOL match. When the pair state is PAIR, all chunks of SVOL are PVOL references. That is, when a read is executed from the host computer 100, the PVOL and SVOL data match in all chunks.

In the case of PAIR, writing to SVOL is prohibited. For example, when a write is issued from the host computer 100 to an SVOL whose pair status is PAIR, an error is returned. COPY is a transient state during copying to transition to PAIR. As with PAIR, writing to SVOL is prohibited. The pair state PSUS is a state in which the update itself is not reflected on either the PVOL or the SVOL while the pair relationship is maintained, and the PVOL and the SVOL are independent. That is, when a write occurs in PVOL, SVOL data is saved.

Also, writing to SVOL is permitted. When the pair status is SMPL, it indicates that PVOL and SVOL are not paired and the snapshot pair ID is not used. The PVOL real volume ID column 803 stores the identifier of the PVOL real volume. The SVOL real volume ID column 804 stores the identifier of the SVOL real volume. The pool ID column 805 stores an identifier of the pool 530 for saving data.

If the PVOL real volume ID is the same, the same pool 530 is normally used. This is because if the pools are different, the saved data cannot be shared between the SVOLs, and the capacity use efficiency and the processing speed during the saving are reduced. In this embodiment, the main object is to move the PSUS snapshot pair between the storage apparatuses 500. This is because in PAIR and COPY, the PVOL and SVOL data are completely the same, and no write to the SVOL occurs, so that the migration process can be realized by the process of moving the PSUS snapshot pair.

FIG. 7B shows an example of the snapshot PVOL management information 900. This information is information held by each storage device 500 and is mainly stored in the memory 508. In the PVOL real volume ID column 901, the identifier of the PVOL real volume is registered. In the address column 902, the address of the PVOL chunk is registered. In the exclusive column 903, a value indicating whether the address is exclusive (yes) or not (no) is registered. Here, “exclusive” refers to prohibiting or restricting access to another chunk by another entity while the other entity is accessing the chunk so that a plurality of entities do not access a chunk at the same time.

FIG. 7 (c) shows an example of the snapshot SVOL management information 1000. This information is information held by each storage device 500 and is mainly stored in the memory 508. In the SVOL real volume ID column 1001, the identifier of the SVOL real volume is registered. In the address column 1002, the address of the SVOL chunk is registered. In the reference type column 1003, a value indicating whether the chunk of the address is a PVOL reference, a shared reference, or an exclusive reference is registered. In the reference address column 1004, the address of the chunk of the pool 530 referred to by the chunk of the address is registered. In the case of PVOL reference, since the reference destination address is always the same address of the PVOL, the reference address is “none”. In the exclusive column 1005, a value indicating whether the address is exclusive (yes) or not (no) is registered.

In the shared partner VOL real volume ID column 1006, the real volume ID of the SVOL (shared partner VOL) that refers to the same address on the pool 530 when the reference type is shared reference is registered. In the case of PVOL reference and exclusive reference, the shared partner VOL real volume ID column 1006 is “none”. In addition, even if the shared reference is used, for example, when there is only one SVOL that forms a snapshot pair with the PVOL, “None” is given when there is no shared partner VOL.

FIG. 8 shows an example of the pool management information 1100. This information is information held by each storage device 500 and is mainly stored in the memory 508. In the pool ID column 1101, the identifier of the pool 530 is registered. In the address column 1102, the address of the chunk of the pool 530 is registered. In the assigned column 1103, a value indicating whether or not the address is assigned to the SVOL (yes) or not (no) is registered.

FIG. 9 shows an example of the moving copy pair management information 1200. This information is information held by each storage device 500 and is mainly stored in the memory 508. In the moving copy pair ID column 1201, the identifier of the moving copy pair is registered. In the pair status column 1202, a value of PAIR or COPY indicating the status of the moving copy pair is registered. PAIR is a state in which a later-described move copy process has been completed for all addresses of the move copy pair, and the data of the move source VOL and the move destination VOL match. COPY is a state that is not PAIR.

In the migration source VOL real volume ID column, the real volume ID of the migration source VOL in the migration copy pair and the identifier of the migration source storage apparatus 500A are registered. In the migration destination VOL real volume ID column, the real volume ID of the migration destination VOL of the migration copy pair and the identifier of the migration destination storage apparatus 500B are registered. In the address column 1205, the addresses of the chunks of the migration source VOL and the migration destination VOL are registered. The moving copy status column 1206 registers a value indicating whether or not the moving copy of the chunk at the address has been completed (incomplete).

FIG. 10 shows an example of the snapshot allocation information 1300. This information is management information for each chunk that is generated for each chunk and is transmitted from the migration source storage apparatus 500A to the migration destination storage apparatus 500B during the migration copy for each chunk. In the migration destination VOL real volume ID column 1301, the real volume ID of the migration destination VOL is registered. In the address column 1302, the address of the chunk of the migration destination VOL is registered. In the reference type column 1303, a value indicating whether the reference is a PVOL reference, a shared reference, or an exclusive reference is registered. When the migration destination VOL is PVOL, the value of the reference type column 1303 is empty. In the sharing partner VOL real volume ID column 1304, the real volume IDs of the sharing partner VOL whose migration copy status is “complete” are registered.

When the migration destination VOL is SVOL and is a PVOL reference or occupancy reference, and when the migration destination VOL is a PVOL, “None” is displayed. Even if the reference is shared, if there is no shared partner VOL, or if the migration copy status of any shared partner VOL is incomplete, data to be shared is not stored in the pool 530 of the migration destination storage apparatus 500B. Therefore, it becomes “None”. In the data presence / absence column 1305, a value indicating whether or not there is data in the data column 1306 is registered. The data column 1306 stores the data of the chunk. The data is stored under the condition that “the migration destination VOL is a PVOL”, or “the migration destination VOL is an SVOL and shared reference, and the value of the shared partner VOL real volume ID column 1304 is“ none ”. Or “move destination VOL is an exclusive reference in SVOL”.

FIG. 11 is a flowchart showing an example of the volume movement process. This processing is executed by the migration source storage 500A in the following description of each step, but may be executed by the management server 200. When executed by the management server 200, the management server 200 executes each step by issuing an instruction to each storage device 500 and the host computer 100 via the LAN. When the storage apparatus 500 executes, it is executed by the migration copy control program 552 included in the storage apparatus 500.

Step 1401: A migration destination VOL that belongs to the virtual storage device 600 having the same virtual storage ID as the migration source VOL of the migration source storage device 500A and has the same virtual volume ID is created in the migration destination storage device 500B. At this time, the snapshot pair management information 800, the snapshot PVOL management information 900, and the snapshot SVOL management information 1000 are updated between the migration source storage apparatus 500A and the migration destination storage apparatus 500B. Note that the migration source VOL is one PVOL and all SVOLs that form a snapshot pair with the PVOL. Accordingly, in this step, one or more destination VOLs are created. Further, a snapshot pair having the same configuration as that of the migration source storage apparatus 500A is set for the created migration destination VOL. At this time, the pair state of the snapshot is PSUS. Also, a path is set for the migration destination VOL to the host.

Step 1402: A migration copy pair is set in the migration source VOL and the migration destination VOL. At this time, the pair status of the moving copy pair is COPY.
Step 1403: The migration copy process (described in FIG. 12) is executed.
Step 1404: Release the path to the host of the migration source VOL.
Step 1405: Release the moving copy pair.
In step 1401, “all SVOLs” is set, but “some SVOLs” may be set. In this case, only some SVOLs are moved to the migration destination storage apparatus 500B. Such partial migration is effective, for example, when load distribution is performed between the migration source storage apparatus 500A and the migration destination storage apparatus 500B.

The moved SVOL can be deleted from the migration source storage apparatus 500A after the end of this flowchart. On the other hand, the PVOL needs to remain in both storage apparatuses 500 in order to maintain the snapshot pair in both the migration source storage apparatus 500A and the migration destination storage apparatus 500B. Further, when the migration copy is executed for duplication, both the PVOL and the SVOL must be left in the migration source storage apparatus 500A, so S1404 and S1405 are not required. FIG. 12 shows an example of the migration copy process. It is a flowchart. This process is executed by the migration copy control program 552 of the migration source storage apparatus 500A.

Step 1501: With reference to the moving copy pair management information 1200, one of the moving copy pairs whose pair status is COPY is selected. The moving copy pair selected here is called a target pair.
Step 1502: If the target pair is found, proceed to Step 1503. Otherwise, since all the moving copies have been completed, the process proceeds to step 1512 (end).
Step 1503: Referring to the moving copy pair management information 1200, one address whose moving copy status is “incomplete” is selected from the addresses of the target pair. The address selected here is called a target address. In the following description, the word “address” may refer to the chunk indicated by the address unless there is a risk of misunderstanding.

Step 1504: Referring to the snapshot pair management information 800 and the moving copy pair management information 1200, if the target pair is a moving copy pair between snapshot PVOLs (YES), the process proceeds to step 1507. If it is a moving copy pair between snapshots SVOL (NO), the process proceeds to step 1505.
Step 1505: Referring to the moving copy pair management information 1200, if the target address of the PVOL that forms a snapshot pair with the volume (SVOL) of the target pair is in the moving copy state “completed”, the process proceeds to step 1507. If the moving copy status is “incomplete”, the process proceeds to step 1506.

Step 1506: If the moving copy state of the target address of PVOL is “incomplete”, it is necessary to execute the moving copy of the target address of PVOL first, so the target pair is changed to a moving copy pair between PVOLs. If the moving copy status of the PVOL referenced by the SVOL remains “incomplete” and a moving copy of the referenced SVOL is made first, the destination SVOL before the moving copy status becomes “completed” With reference to PVOL, there is a discrepancy in the snapshot configuration between the source and destination. Step 1506 makes it possible to prevent this discrepancy.

Step 1507: The exclusion of the target address of the migration source VOL is acquired. In all the subsequent processes including this process, exclusion is acquired only in the migration source VOL, and exclusion control is not performed in the migration destination VOL. As shown in step 2002 of FIG. 17 to be described later, when a write command for the migration destination VOL is received, the write command is transferred to the migration source storage apparatus. Therefore, if exclusive control is performed only on the migration source VOL, There is no discrepancy in the snapshot configuration between the source and destination VOLs.
Step 1508: If acquisition of exclusion is successful, the process proceeds to Step 1509. If unsuccessful, the process proceeds to step 1501.

Step 1509: The snapshot allocation information 1300 is transmitted to the migration destination storage apparatus 500B, and a process completion notification from the migration destination storage apparatus 500B is awaited. Note that the migration destination VOL real volume ID of the target pair is registered in the migration destination VOL real volume ID column 1301 of the snapshot allocation information 1300 based on the migration copy pair management information 1200. A target address is registered in the address column 1302. In the reference type 1303, a reference type is registered based on the snapshot SVOL management information 1000. When the target pair is a moving copy pair between PVOLs, the reference type 1303 is empty. In the shared partner VOL real volume ID column 1304, the real volume ID of the shared partner VOL is registered based on the shared partner VOL real volume ID column 1006 of the snapshot SVOL management information 1000. Although it may not be registered depending on conditions such as the reference type, such conditions have already been described. The conditions for registering the data presence / absence column 1305 and the data column 1306 have also been described.

Step 1510: The moving copy status of the target address of the target pair in the moving copy pair management information 1200 is set to “complete”, and the exclusion is released.
Step 1511: If the moving copy status of all addresses of the target pair is “complete”, the pair status of the moving copy is changed to PAIR, and the process proceeds to Step 1501. Otherwise, nothing is done in this step, and the process proceeds to step 1501 in the same manner.

FIG. 13 is a flowchart showing an example of processing at the time of receiving snapshot allocation information from the migration source storage apparatus. This process is executed by the migration copy control program 552 of the migration destination storage apparatus 500B.

Step 1601: The snapshot allocation information 1300 transmitted at Step 1509 is received.
Step 1602: If the migration destination VOL indicated by the migration destination VOL real volume ID column 1301 of the snapshot allocation information 1300 is a snapshot PVOL (YES), the process proceeds to Step 1603. If it is a snapshot SVOL (NO), the process proceeds to step 1604.
Step 1603: The data of the target address (value of the address string 1302) of the migration destination VOL is updated with the data of the data string 1306. Since this step is a step for PVOL, there is no update related to the snapshot SVOL management information 1000 such as the reference type.

Step 1604: If the real volume ID of the shared partner VOL is stored in the shared partner VOL real volume ID column 1304 of the snapshot allocation information 1300 (YES), the process proceeds to Step 1606. If the real volume ID of the sharing partner VOL is not stored (NO), the process proceeds to step 1605.
Step 1605: The unassigned address of the pool 530 is searched, and the data in the data string 1306 is stored in the chunk of the found address. Further, the value of the reference type column 1003 of the snapshot SVOL management information 1000 of the migration destination VOL is updated with the value of the reference type column 1303 of the snapshot allocation information 1300. Further, the value of the reference address column 1004 is set as the address of the chunk of the pool 530 storing the data.

Step 1606: For the target address of the migration destination VOL, the values of the reference type column 1003 and the shared partner VOL real volume ID column 1006 of the snapshot SVOL management information 1000 are updated with the values of the columns having the same names in the snapshot allocation information 1300, respectively. . Similarly, for the target address of the migration destination VOL, the reference address column 1004 is updated with the value of the column of the same name in the snapshot SVOL management information 1000 of the sharing partner VOL. Further, for the target address of the sharing partner VOL, the value of the sharing partner VOL real volume ID column 1006 of the snapshot SVOL management information 1000 is updated with the value of the column of the same name in the snapshot allocation information 1300. As described above, when the sharing partner VOL exists, the processing is completed only by rewriting the snapshot SVOL management information 1000, and data transfer is not performed.

Step 1607: The migration copy status of the target address of the migration copy pair to which the migration destination VOL belongs is set to “complete”.
Step 1608: If the move copy status of all addresses is “complete” in step 1607, the pair status of the move copy pair is changed to PAIR. Otherwise, do nothing in this step.
Step 1609: A completion notification is transmitted to the migration source storage apparatus 500A.
As described above, the migration copy process is performed in units of chunks using the snapshot allocation information 1300 that is information in units of chunks, and exclusive acquisition at that time is also performed in units of chunks. Compared to the case where the migration copy process is performed for each volume and the exclusion is acquired for each volume, the I / O is acquired for chunks that are not related to the migration copy process even if they are included in the same volume by acquiring the exclusion in a finer unit. O can be accepted.

FIG. 14 is a flowchart showing an example of host write reception processing to the migration source VOL. This processing is executed by the snapshot control program 551 of the migration source storage apparatus 500A. Step 1701: A write command for the migration source VOL is received from the host. The source VOL may be either PVOL or SVOL. Further, the volume and address that are the write target are referred to as a target VOL (or write target VOL) and target address, respectively.

Step 1702: The exclusion of the target address of the write target VOL and the SVOL that shares data with the write target VOL is acquired. Here, “SVOL that shares data with the write target VOL” means “SVOL whose target address is a PVOL reference” when the write target VOL is PVOL, and about the target address when the write target VOL is SVOL. It is a sharing partner VOL. Note that when the target address of the write target VOL is a PVOL reference or exclusive reference, exclusion is acquired only for the write target VOL.

Due to the exclusion in step 1702, during the write process, the migration copy process is no longer executed to the address to which the reference destination is changed by the write due to the write target address and the snapshot configuration. Thus, it is possible to prevent a difference in the snapshot configuration from occurring. Here, if the target address of the write target VOL and the SVOL that shares data with the write target VOL has been acquired by moving copy, for example, does not proceed to step 1703 and execute step 1702 again? Return an error to the host.

Step 1703: Write is executed. That is, the data included in the write command is written to the target address of the target VOL. When the write target VOL is SVOL and the target address is a PVOL reference, an unallocated chunk of the pool 530 is searched, data is written to the found chunk, and the target address is updated with the address of the chunk of the found pool 530. In other cases, since a chunk is assigned to the target address, data is written to the chunk. Note that the writing here is physically performed to the cache memory 507 first, and writing (destage) from the cache memory 507 to the physical disk 501 is executed asynchronously with this processing. This is a general storage apparatus process.

Step 1704: Data saving processing (described in FIG. 16) is executed.

Step 1705: If the moving copy state of the target address of the write target VOL is “completed”, the process proceeds to step 1706. If it is “incomplete”, the process proceeds to step 1709. When the migration copy status is “incomplete”, the write to the write target VOL of the migration source VOL is reflected in the migration destination VOL by the subsequent migration copy, and when the migration copy status is “complete”, the migration source VOL is written. The write to the target VOL is reflected on the destination VOL at step 1706.

Step 1706: The write command for the migration destination VOL that forms a migration copy pair with the write target VOL and the snapshot allocation information 1300 of the SVOL that acquired the exclusion in Step 1702 are transmitted to the migration destination storage apparatus 500B, and a completion notification is awaited. It should be noted that the snapshot allocation information 1300 to be transmitted is not “snapshot allocation information 1300 of the SVOL acquired in step 1702” but “moving copy status of the target address in the SVOL acquired in step 1702 is incomplete. The snapshot allocation information 1300 ”may be used.

Also, the snapshot allocation information 1300 to be transmitted is the information before saving data in step 1704. The change of the reference destination due to the saving caused by the write to the migration source VOL is reflected by the saving performed at the migration destination, so the snapshot allocation information 1300 to be transmitted may be the information before saving the data in step 1704. . Further, by transmitting the snapshot allocation information 1300 together with the write command, the migration copy is executed not only during the migration copy processing of FIGS. 12 and 13 but also when writing to the migration source.

Step 1707: The migration copy state of the target address of the SVOL that sent the snapshot allocation information 1300 in step 1706 is set to “complete”.
Step 1708: As a result of Step 1707, the pair status of the move copy pair in which the move copy status of all addresses is “complete” is set to PAIR.
Step 1709: The exclusion acquired in step 1702 is canceled.
Step 1710: Send completion (RSP) to the host.

FIG. 15 is a flowchart showing an example of processing at the time of receiving write and snapshot allocation information from the migration source storage apparatus 500A. This process is executed by the snapshot control program 551 of the migration destination storage apparatus 500B.

Step 1801: The write command and snapshot allocation information for the migration destination VOL (write target VOL) transmitted from the migration source storage apparatus 500A in Step 1706 are received.

Step 1802: Based on the snapshot allocation information 1300, the snapshot SVOL management information 1000 is updated. This update process is the same as the snapshot allocation information reception process (shown in FIG. 13). However, since the snapshot allocation information 1300 processed in this step includes information of a plurality of SVOLs, the snapshot allocation information reception process is executed for each.

Step 1803: For the SVOL updated in Step 1802, the migration copy state of the target address is set to “complete”.
Step 1804: As a result of Step 1803, the pair status of the move copy pair in which the move copy status of all addresses is “complete” is set to PAIR.
Step 1805: Write is executed.
Step 1806: Data saving processing (described in FIG. 16) is executed.
Step 1807: A completion notification is returned to the migration source storage apparatus 500A.

FIG. 16 is a flowchart showing an example of the data saving process. This process is executed by the snapshot control program 551 of the storage apparatus 500. Further, although this process is executed immediately after the write is executed, it may be executed asynchronously. That is, the data saving process may not be executed immediately after the write is executed (written to the cache memory 507), but may be executed when the destage is executed.

Step 1901: When the write target VOL is PVOL, the process proceeds to Step 1902. If the target VOL is an SVOL, the process proceeds to step 1905.
Step 1902: Among the SVOLs, a search is made for a target address whose reference type is a PVOL reference.

Step 1903: If at least one SVOL is found in Step 1902, an unallocated chunk is searched from the pool 530, and the data before the write execution of the target address is stored in the chunk. The address of the storage chunk is called the new address. If no SVOL is found, nothing is done in this step.

Step 1904: For the target address of the SVOL found in Step 1902, the reference type column 1003 of the snapshot SVOL management information 1000 is updated to the shared reference, and the reference address column 1004 is updated to the new address. Further, the value of the shared partner VOL real volume ID column 1006 is updated based on the found SVOL.

Step 1905: If the target address is a shared reference, go to Step 1906. If the target address is a PVOL reference or an exclusive reference, the process proceeds to step 1910.
Step 1906: For the target address of the write target VOL, the reference type column 1003 of the snapshot SVOL management information 1000 is set as an exclusive reference.
Step 1907: The sharing partner VOL is searched for the target address of the write target VOL. That is, the shared partner VOL real volume ID column 1006 of the snapshot SVOL management information 1000 is referred to.

Step 1908: If at least one shared partner VOL is found in Step 1907, an unallocated chunk is searched from the pool 530, and the data before write execution of the target address is stored in the chunk. The address of the storage chunk is called the new address. If no shared partner VOL is found, nothing is done in this step.
Step 1909: The reference address column 1004 of the snapshot SVOL management information 1000 is updated to a new address for the target address of the shared partner VOL found in Step 1907.

Step 1910: For the target address of the write target VOL, the reference type column 1003 of the snapshot SVOL management information 1000 is set as an exclusive reference. If it is already an exclusive reference, nothing is done in this step.

FIG. 17 is a flowchart showing an example of processing at the time of host write reception to the migration destination VOL. This process is executed by the snapshot control program 551 of the migration destination storage apparatus 500B.

Step 2001: Receive a write for the migration destination VOL from the host computer 100.
Step 2002: Transfer the write command to the migration source storage apparatus 500A and wait for the completion notification.
Step 2003: If the migration copy state of the target address of the write target VOL included in the completion notification is “complete”, the process proceeds to step 2004. If “incomplete”, the process proceeds to step 2009.

Step 2004: Update the snapshot SVOL management information 1000 based on the snapshot allocation information 1300 included in the completion notification. This update process is the same as the snapshot allocation information reception process (shown in FIG. 13). However, since the snapshot allocation information 1300 processed in this step includes information of a plurality of SVOLs, the snapshot allocation information reception process is executed for each.

Step 2005: The migration copy status is set to “complete” for the target address of the SVOL updated in step 2004.
Step 2006: As a result of Step 2005, the pair status of the move copy pair in which the move copy status of all addresses is “complete” is set to PAIR.

Step 2007: Write is executed.
Step 2008: A data saving process is executed.
Step 2009: Send completion (RSP) to the host computer 100.
Step 2010: Send exclusive release to the migration source storage apparatus 500A.

FIG. 18 is a flowchart showing an example of processing upon receipt of a write from the migration destination storage apparatus 500B. This processing is executed by the migration source storage apparatus 500A.

Step 2101: A write command is received from the migration destination storage apparatus 500B.
Step 2102: The exclusion of the target address of the write target VOL and the SVOL sharing data with the write target VOL is acquired. Here, “SVOL that shares data with the write target VOL” means “SVOL whose target address is a PVOL reference” when the write target VOL is PVOL, and about the target address when the write target VOL is SVOL. It is a sharing partner VOL. Note that when the target address of the write target VOL is a PVOL reference or exclusive reference, exclusion is acquired only for the write target VOL.

Step 2103: Write is executed.
Step 2104: A data saving process is executed.
Step 2105: If the migration copy status of the target address of the write target VOL is “complete”, the process proceeds to step 2106. If “incomplete”, the process proceeds to step 2109.

Step 2106: A completion notification is returned to the migration destination storage apparatus 500B, and an exclusive release instruction is awaited. The completion notification to be transmitted includes “the migration copy state of the target address of the write target VOL is“ complete ”” and the snapshot allocation information 1300 of the SVOL that acquired the exclusion in step 2102. Note that the snapshot allocation information 1300 is information before execution of the data saving process in step 2104.

Further, the snapshot allocation information 1300 to be transmitted is not “snapshot allocation information 1300 of the SVOL that acquired the exclusion” but “snapshot allocation information 1300 of the target address that has not been moved and copied at the target address”. But you can.

Step 2107: The migration copy status is set to “complete” for the target address of the SVOL that has transmitted the snapshot allocation information 1300 in step 2106.

Step 2108: As a result of Step 2107, the pair status of the move copy pair in which the move copy status of all addresses is “completed” is set to PAIR. Step 2109: A completion notification is returned to the migration destination storage apparatus 500B, and the exclusive release is released. Wait for instructions. The completion notification to be transmitted includes that “the migration copy state of the target address of the write target VOL is“ incomplete ””.

Step 2110: The acquired exclusion is released. This step is not executed until there is an exclusion release instruction from the migration destination storage apparatus 500B as described in step 2106 and step 2109.

17 and 18, even if a write request is received at the movement destination, it becomes possible to execute the write without causing a discrepancy in the snapshot configuration.

FIG. 19 is a flowchart showing an example of processing upon receipt of a read to the migration source VOL. This process is executed by the snapshot control program 551 of the migration source storage apparatus 500A.

Step 2201: A read (sometimes referred to as a read command) is received from the host computer 100 or the migration destination storage apparatus 500B. The volume and address to be read are called a read target VOL and a target address, respectively.
Step 2202: The exclusion of the target address of the read target VOL is acquired.
Step 2203: Read is executed. That is, the data stored in the target address of the read target VOL is read.
Step 2204: The exclusion acquired at step 2202 is canceled.
Step 2205: Send completion (RSP). The RSP also includes read result data.

FIG. 20 is a flowchart showing an example of processing at the time of receiving a lead to the destination VOL. This process is executed by the snapshot control program 551 of the migration destination storage apparatus 500B.

Step 2301: A read is received from the host computer 100.
Step 2302: If the moving copy status of the target address of the read target VOL is “complete”, the process proceeds to step 2303. If it is “incomplete”, the process proceeds to step 2304.
Step 2303: Read is executed.
Step 2304: Read is transferred to the migration source storage apparatus 500A and waits for RSP. The RSP that is responded includes the data of the read result.
Step 2305: Send RSP to the host. The RSP includes read result data.

As described above, in the first embodiment, the moving copy is performed in units of chunks using the snapshot allocation information 1300, and at this time, the exclusion is acquired in units of chunks. As a result, it is possible to continue to accept I / O for chunks not related to the moving copy process. Then, when there is a write from the host to the migration source or migration destination volume of the migration copy, write data is written to the migration source volume to determine whether the migration copy of the write target address is complete, If the migration copy is complete, send the data related to the write request to the migration destination volume to reflect the write to the migration source in the migration destination, and if the migration copy is not complete, move the write data The transmission to the movement source is reflected on the movement destination by the movement copy performed later, without transmitting it first.

As a result, since the write is also reflected in the moving copy pair for which the moving copy has been completed, it is possible to prevent a difference in the snapshot configuration between the moving source and the moving destination in the moving copy pair for which the moving copy has been completed. Also, by acquiring the exclusive of the address of the move copy target, the write target address, and the address related to the write target address by the snapshot configuration, even if the write is accepted during the move copy, the move source and move It is possible to prevent the snapshot configuration from being different from the previous one.

This enables storage devices to be replaced while maintaining the snapshot function without host downtime. In addition, the snapshot copy information 1300 is transmitted together with the write transmission from the migration source to the migration destination, so that the migration copy process is performed when writing to the migration source or the migration destination. Both write processing and move copy processing acquire chunk exclusion. By performing a move copy of a chunk that has been locked by a write process, the number of times that a lock is acquired can be reduced compared to when a copy of a chunk is acquired again and a move copy is performed later. As a result, the processing time for moving copy can be shortened.

A second embodiment according to the present invention will be described with reference to FIGS. 2 to 11, 13, 16, and 19 to 25. FIG. The difference between the present embodiment and the first embodiment is that the exclusion of the target address is not acquired in the moving copy process. As a result, the influence of the migration copy process on the write performance / read performance of the migration source VOL and the migration destination VOL can be reduced. Further, exclusive acquisition in the migration source storage apparatus 500A when a write to the PVOL occurs is performed in a smaller range than in the first embodiment. Thereby, the write performance / read performance of PVOL is improved as compared with the first embodiment. Hereinafter, descriptions of FIGS. 21 to 27 which are differences from the first embodiment will be described.

FIG. 21 is a flowchart illustrating an example of the move copy process. This processing is executed by the migration source storage apparatus 500A.
Step 2401: One moving copy pair whose pair status is COPY is selected. The moving copy pair selected here is called a target pair.

Step 2402: If the target pair is found, proceed to Step 2403. Otherwise, since all the moving copies have been completed, the process proceeds to step 2412 (end).

Step 2403: Select one address whose migration copy status is “incomplete” from the addresses of the target pair. The address selected here is called a target address.

Step 2404: If the target pair is a moving copy pair between snapshot PVOLs (YES), the process proceeds to step 2407. If it is a moving copy pair between snapshots SVOL (NO), the process proceeds to step 2405.

Step 2405: If the target address of the PVOL that forms the snapshot pair with the volume (SVOL) of the target pair is in the migration copy state “completed”, the process proceeds to step 2407. If the moving copy status is “incomplete”, the process advances to step 2406.

Step 2406: When the moving copy state of the target address of PVOL is “incomplete”, it is necessary to execute the moving copy of the target address of PVOL first, so the target pair is changed to a moving copy pair between PVOLs.

Step 2407: The snapshot allocation information 1300 is transmitted to the migration destination storage apparatus 500B, and a process completion notification from the migration destination storage apparatus 500B is awaited. The values registered in each column of the snapshot allocation information 1300 are the same as those in the first embodiment. If the method of executing the data saving process asynchronously with the write is taken, the target address of the related volume (if the target address of the target pair is a PVOL reference, the target address of the PVOL, the target before executing this step) If there is a shared partner VOL at the target address of the pair, the destage and data save processing of the shared partner VOL target address) is executed. Further, the destaging and data saving processing of the related volume is also executed immediately before detecting the difference in the snapshot allocation information 1300 in step 2408.

Step 2408: The snapshot allocation information 1300 transmitted in step 2407 is compared with the snapshot allocation information 1300 based on the latest information of the migration source storage apparatus 500A at the time of execution of this step. If there is a difference, the process returns to step 2407. If they are the same, the process proceeds to step 2409. In this step, while the snapshot allocation information 1300 is being transmitted to the migration destination storage apparatus 500B in step 2407, the data and management information of the target address of the target pair in the migration source storage apparatus 500A are written and saved by the data saving process. If changed, execute to detect and correct the change.

For the same reason, even if the migration destination storage apparatus 500B receives the snapshot allocation information 1300, it cannot be determined whether the migration copy status should be “completed”. Step 1607 and Step 1608 are omitted. Note that the change of the migration copy status in the migration destination storage apparatus 500B is executed in step 2411.

Step 2408 is repeated until there is no difference between the snapshot configuration of the migration source address and the migration destination address in step 2408. Therefore, during the migration copy process, the migration copy process target address and the snapshot copy configuration target address are changed. Even if a write process occurs at an address related to, it is possible to prevent the snapshot configuration from being different between the source address and the destination address. Further, as in the first embodiment, by acquiring exclusion for each chunk and performing a moving copy using the snapshot allocation information 1300 that is information in units of chunks, I / O to a chunk unrelated to the moving copy can be performed. Can be accepted.

Step 2409: The migration copy state of the target address of the target pair is set to “complete”.
Step 2410: If the moving copy status of all addresses of the target pair is “complete”, the pair status of the moving copy is changed to PAIR, and the processing proceeds to step 2411. Otherwise, nothing is done in this step, and the process proceeds to step 2411 in the same manner.

Step 2411: The migration copy status and the pair status change in Step 2409 and Step 2410 are transmitted to the migration destination storage apparatus 500B. That is, the migration destination storage apparatus 500B is notified to set the migration copy status of the target address of the target pair to “complete”. If the moving copy status of all addresses is “completed”, a notification is also made to set the moving copy pair status to PAIR. Note that the processing of the migration destination storage apparatus 500B that has received this notification merely updates the migration copy pair management information 1200, and therefore the flowchart is omitted.

FIG. 22 is a flowchart showing an example of processing at the time of host write reception to the migration source VOL. This processing is executed by the migration source storage apparatus 500A.

Step 2501: A write command for the migration source VOL is received from the host. The source VOL may be either PVOL or SVOL. Further, the volume and address that are the write target are referred to as a target VOL (or write target VOL) and target address, respectively.

Step 2502: If the write target VOL is PVOL, or “the write target VOL is SVOL and the target address is an exclusive reference”, the process proceeds to Step 2503. Otherwise (in SVOL, the target address is a PVOL reference or shared reference), the process proceeds to step 2504.
Step 2503: The exclusion of the target address of the write target VOL is acquired.
Step 2504: If the target address of the write target VOL is a PVOL reference, the process proceeds to Step 2505. In the case of shared reference, the process proceeds to step 2506.

Step 2505: The exclusion of the target address of the write target VOL and “PVOL that forms a snapshot pair with the write target VOL” is acquired.
Step 2506: The exclusion of the target address of the write target VOL and the “volume that is the shared partner VOL for the target address of the write target VOL” is acquired.

The exclusive acquisition in steps 2502 to 2506 is different from the exclusive acquisition in step 1702 of the first embodiment. In the PVOL and the SVOL that refers to the PVOL, in the first embodiment, when the PVOL is written, the PVOL and the SVOL are exclusive, and when the SVOL is written, the SVOL is exclusive. On the other hand, in the second embodiment, when there is a write in the PVOL, the exclusion of the PVOL is acquired. When there is a write in the SVOL, the exclusion of the PVOL referred to as the SVOL is acquired.

In the exclusion of the first embodiment, only the exclusion of the write target is acquired when writing to the SVOL, so that the performance of writing to the SVOL can be improved. On the other hand, in the exclusion of the second embodiment, only the write target exclusion is acquired when writing to the PVOL, so that the performance of writing to the PVOL can be improved. In the second embodiment, exclusive acquisition may be used within the range described in step 1702 of the first embodiment.

Step 2507: Write is executed.
Step 2508: Data saving processing (described in FIG. 16) is executed.
Step 2509: If the migration copy status of the target address of the write target VOL is “completed”, the process proceeds to step 2510. If “incomplete”, the process proceeds to step 2511.
Step 2510: A write command for the migration destination VOL that forms a migration copy pair with the write target VOL is transmitted to the migration destination storage apparatus 500B, and a completion notification is awaited.
Step 2511: The acquired exclusion is released.
Step 2512: Send completion (RSP) to the host.

FIG. 23 is a flowchart showing an example of processing upon receipt of a write from the migration source storage apparatus. This processing is executed by the migration destination storage apparatus 500B.

Step 2601: The write command for the migration destination VOL (write target VOL) transmitted from the migration source storage apparatus 500A in step 2510 is received.
Step 2602: Write is executed.
Step 2603: Data saving processing (described in FIG. 16) is executed.
Step 2604: Send completion (RSP) to the migration source storage apparatus 500A.

FIG. 24 is a flowchart showing an example of processing upon receipt of a host write to the migration destination VOL. This processing is executed by the migration destination storage apparatus 500B.

Step 2701: A write for the migration destination VOL is received from the host computer 100.
Step 2702: The write command is transferred to the migration source storage apparatus 500A, and a completion notification is awaited.
Step 2703: If the migration copy status of the target address of the write target VOL included in the completion notification is “complete”, the process proceeds to step 2704. If it is “incomplete”, the process proceeds to step 2706.

Step 2704: Write is executed.
Step 2705: Data saving processing is executed.
Step 2706: Send completion (RSP) to the host computer 100.
Step 2707: An exclusion release is transmitted to the migration source storage apparatus 500A.

FIG. 25 is a flowchart showing an example of processing upon receipt of a write from the migration destination storage apparatus 500B. This processing is executed by the migration source storage apparatus 500A.

Step 2801: A write command is received from the migration destination storage apparatus 500B.
Step 2802 to Step 2806: Steps for obtaining exclusion, but since the processing is the same as Step 2502 to Step 2506 in FIG.
Step 2807: Write is executed.
Step 2808: Data saving processing is executed.

Step 2809: If the migration copy status of the target address of the write target VOL is “completed”, the process proceeds to step 2810. If “incomplete”, the process proceeds to step 2811.
Step 2810: A completion notification is returned to the migration destination storage apparatus 500B, and an exclusive release instruction is awaited. Note that the completion notification to be transmitted includes “the migration copy state of the target address of the write target VOL is“ complete ””.

Step 2811: A completion notification is returned to the migration destination storage apparatus 500B, and an exclusive release instruction is awaited. The completion notification to be transmitted includes that “the migration copy state of the target address of the write target VOL is“ incomplete ””.

Step 2812: The acquired exclusion is released. This step is not executed until there is an exclusion release instruction from the migration destination storage apparatus 500B as described in step 2810 and step 2811.

Unlike the first embodiment, the second embodiment does not acquire the exclusion of the target address of the moving copy at the time of the moving copy process, and instead takes a snapshot until there is no difference in the snapshot configuration between the target pairs of the moving copy. The transmission of the allocation information 1300 is repeated. Therefore, the write to the target address of the moving copy is not delayed due to exclusion by the moving copy, and the snapshot allocation information is used when the write command is transferred from the source VOL to the destination VOL when writing to the source. 1300 is not transmitted. Therefore, the load during the write process is reduced as compared with the first embodiment. In other cases, as described above, the exclusive acquisition at the time of writing to the movement source has a higher write performance to the PVOL than the first embodiment.

A third embodiment according to the present invention will be described with reference to FIGS. 2 to 9, 11, 16, 16, 19, 20, 23, and 26 to 29. FIG. The main difference between the present embodiment and the first and second embodiments lies in the moving copy processing method. In the first and second embodiments, in the migration copy process, the snapshot VOL is migrated by transmitting the snapshot allocation information 1300 including chunk data and pointers of the migration source VOL from the migration source storage to the migration destination storage. Executed.

On the other hand, in the present embodiment, the migration of the snapshot VOL is executed by transmitting the chunk data of the migration source VOL as a write command from the migration source storage apparatus 500A to the migration destination storage apparatus 500B. This processing method eliminates the need for processing for interpreting the snapshot allocation information 1300 in the migration destination storage apparatus 500B as compared with the first and second embodiments.

That is, if the migration destination storage apparatus 500B has a basic function as a storage apparatus, that is, a function for processing a write from the host and a snapshot function, the snapshot VOL can be migrated. However, in the first and second embodiments, the order of the target VOLs of the migration copy process is not particularly considered. In the present embodiment, it is necessary to perform the process of reproducing in the migration destination storage apparatus 500B in order from the older generation SVOL. Therefore, there is a limitation on the processing order.

In the first and second embodiments, when the snapshot pair is set in step 1401 of FIG. 11, the same PSUS as that of the migration source VOL is set. However, in this embodiment, all are set to PAIR, and the transition is made to PSUS during the migration copy process. Let Hereinafter, descriptions of FIGS. 26 to 29 which are differences from the first and second embodiments will be described.

FIG. 26 is a flowchart showing an example of the moving copy process. This processing is executed by the migration source storage apparatus 500A.

In this process, “1. Write all data of the migration source SVOL of the oldest generation to the migration destination PVOL, 2. Change the pair status of the snapshot pair of the migration destination SVOL corresponding to the migration source SVOL to PSUS, 3. After that, even if there is a write in the migration destination PVOL, the data of the migration source VOL and the migration destination SVOL remain the same. 3. For the migration source SVOL that is one younger than the oldest generation, the oldest generation SVOL 4. When the pair status of the snapshot pair of the migration destination SVOL corresponding to the migration source SVOL is changed to PSUS, even if the migration destination PVOL is subsequently written, 4. Make sure that the data of the migration source VOL and the migration destination SVOL match. Perform the movement of all generations in 3 repeats the processing of 4 from "that Te.

In other words, after the pair status of all the SVOL snapshot pairs in the migration destination is set to PAIR, the difference data managed by the migration source SVOL (all data of the SVOL in the oldest generation SVOL) is of the old generation Are sequentially written in the destination PVOL. At that time, when writing of the difference data of a generation (all data of the SVOL in the oldest generation SVOL) to the migration destination PVOL is completed, the pair status of the snapshot pair of the migration destination SVOL corresponding to the generation is changed. By setting it to PSUS, even when writing to the destination PVOL by the subsequent migration copy, the source SVOL and the destination SVOL of the generation remain the same, and the source of the source in order from the oldest generation in the destination The snapshot configuration is reproduced.

Hereinafter, each step of FIG. 26 will be described.
Step 2901: From the copy copy pair whose pair status is COPY, select the copy source with the oldest snapshot generation (oldest generation) of the transfer source VOL. The moving copy pair selected here is called a target pair. Note that PVOL is selected as the latest generation. The snapshot generation is an order when one or more snapshot pairs of a certain PVOL are arranged in the order of changing to PSUS.

That is, the oldest snapshot pair that changed to PSUS is the oldest generation, and the latest generation is arranged in the order of change to PSUS. Information about this generation is held as a generation number in the snapshot pair management information 800, for example. This generation number is incremented (incremented) every time a snapshot pair changes to PSUS, and is registered as the generation number of the snapshot pair. That is, the smaller the generation number, the older the generation and the larger the generation number, the new generation.

Step 2902: If the target pair is found, proceed to Step 2903. Otherwise, all moving copies have been completed, and the process proceeds to step 2912 (end).
Step 2903: From the addresses of the target pair, one address whose migration copy status is “incomplete” is selected. The address selected here is called a target address.

Step 2904: Acquire exclusion of the target address.
Step 2905: If exclusive acquisition is successful in step 2904, the flow proceeds to step 2906. If unsuccessful, the process proceeds to step 2903.
Step 2906: It is detected whether there is a difference between the data of the target address of the target pair as compared to the previous generation (previous generation). If there is a difference, the process proceeds to step 2907. If there is no difference, the process proceeds to Step 2908. If the previous generation does not exist, the address is processed as if there is a difference.

Step 2907: A write command of the target address data (difference data) is transmitted to the destination PVOL, and the process is awaited. Thereby, the data of the target address of the transfer source VOL of the target pair matches the target address of the transfer destination PVOL.

Step 2908: The moving copy status of the target address of the target pair is set to “complete”.

Step 2909: If the moving copy status of all addresses of the target pair is “complete”, the pair status of the moving copy is changed to PAIR, and the process proceeds to step 2910. Otherwise, nothing is done in this step, and the process proceeds to step 2910 in the same manner.

Step 2910: If the pair status of the moving copy pair is PAIR as a result of Step 2909, the pair status of the moving copy of the target pair is set to PAIR, and the pair status of the snapshot of the moving destination SVOL is set to PSUS. Notify you to change. When the target pair is a PVOL pair, only the notification for changing the pair status of the migration copy of the target pair to PAIR is transmitted, and the notification for changing the snapshot pair status to PSUS is omitted. Note that the processing of the migration destination storage apparatus 500B that has received this notification merely updates the migration copy pair management information 1200 and the snapshot SVOL management information 1000, and therefore the flowchart is omitted.
Step 2911: The acquired exclusion is released.

FIG. 27 is a flowchart showing an example of processing upon receipt of a host write to the migration source VOL. This processing is executed by the migration source storage apparatus 500A. As a rough process, if the migration copy of the write target VOL is completed (becomes PAIR), a write command is transmitted to the migration destination VOL that forms a migration copy pair with the write target VOL, and the migration source VOL Match the data of the destination VOL. If the migration copy of the write target VOL has not been completed and the migration copy status of the target address is “complete”, the write is transferred to the migration destination PVOL because it is before the PSUS transition in the migration copy process, and the migration source VOL And the data of the destination PVOL are matched. If the migration copy of the write target VOL has not been completed and the migration copy status of the target address is “incomplete”, the write is not transferred to the migration destination storage apparatus 500B. Hereinafter, each step will be described.

Step 3001: A write command for the migration source VOL is received from the host. The source VOL may be either PVOL or SVOL. Further, the volume and address that are the write target are referred to as a target VOL (or write target VOL) and target address, respectively.

Step 3002 to Step 3006: In these steps, an appropriate range of exclusion is acquired. Details are omitted because they are the same as steps 2502 to 2506.

Step 3007: Write is executed.

Step 3008: The data saving process (described in FIG. 16) is executed.

Step 3009: If the pair status of the migration copy of the write target VOL is PAIR, the process proceeds to Step 3012. In this case, if the write target VOL is SVOL, the snapshot pair of the corresponding migration destination SVOL is PSUS. Otherwise, go to step 3010. In this case, if the write target VOL is SVOL, the snapshot pair of the corresponding movement destination SVOL is PAIR.

Step 3010: If the moving copy status of the target address of the write target VOL is “complete”, the process proceeds to step 3011. If “incomplete”, the process proceeds to step 3013. This is because in the case of “incomplete”, the write to the migration source can be reflected in the migration destination by the migration copy process to be executed later on the target address of the write target VOL.

Step 3011: A write command is transmitted to the movement destination PVOL, and the process is awaited. The migration destination storage apparatus 500B that has received the write command executes the write reception process shown in FIG.

Step 3012: A write command is transmitted to the movement destination VOL and the process is completed. The migration destination storage apparatus 500B that has received the write command executes the write reception process shown in FIG. Note that the difference between this step and step 3011 is that the transmission target of the write command is “a migration destination VOL that forms a migration copy pair with a write target VOL” or “a migration destination PVOL that forms a snapshot pair with the migration destination VOL”. . In this step, if the write target VOL is SVOL, the snapshot pair of the corresponding destination SVOL is PSUS. Therefore, in order to reflect the write at the movement source on the movement destination, it is only necessary to transmit a write command to “the movement destination VOL that forms a movement copy pair with the write target VOL”. On the other hand, if the write target VOL is SVOL in step 3011, the snapshot pair of the corresponding destination SVOL is PAIR. Therefore, in order to reflect the write at the movement source on the movement destination, a write command may be transmitted to “the movement destination PVOL that forms a snapshot pair with the movement destination VOL”. When the write target VOL is a PVOL, in either case, the write command transmission target is the destination PVOL.

Step 3013: The acquired exclusion is released.
Step 3014: Send completion (RSP) to the host.

FIG. 28 is a flowchart showing an example of processing upon receipt of a host write to the migration destination VOL. This processing is executed by the migration destination storage apparatus 500B.

Step 3101: A write for the migration destination VOL is received from the host computer 100.
Step 3102: The write command is transferred to the migration source storage apparatus 500A, and a completion notification is awaited.
Step 3103: If the pair status of the migration copy of the write target VOL is PAIR, the process proceeds to Step 3104. In this case, if the write target VOL is SVOL, the snapshot pair of the SVOL is PSUS. Otherwise, go to step 3106. In this case, if the write target VOL is an SVOL, the snapshot pair of the SVOL is PAIR.

Step 3104: Write is executed.
Step 3105: Data saving processing is executed.
Step 3106: Send completion (RSP) to the host computer 100.
Step 3107: An exclusive release is transmitted to the migration source storage apparatus 500A.
FIG. 29 is a flowchart illustrating an example of processing upon receipt of a write from the migration destination storage apparatus. This processing is executed by the migration source storage apparatus 500A.

Step 3201: A write command is received from the migration destination storage apparatus 500B.
Step 3202 to Step 3206: Steps for obtaining exclusion, but since the processing is the same as Step 2502 to Step 2506, description thereof will be omitted.
Step 3207: Write is executed.
Step 3208: The data saving process is executed.

Step 3209: If the pair status of the migration copy of the write target VOL is PAIR, the process proceeds to Step 3212. In this case, writing to the destination VOL is performed in step 2104 in FIG. Otherwise, go to step 3210. In this case, if the write target VOL is SVOL, the snapshot pair of the corresponding movement destination SVOL is PAIR.

Step 3210: If the moving copy status of the target address of the write target VOL is “complete”, the process proceeds to step 3211. If “incomplete”, the process proceeds to step 3212. This is because in the case of “incomplete”, the write to the migration source can be reflected in the migration destination by the migration copy process to be executed later on the target address of the write target VOL.

Step 3211: A write command is transmitted to the destination PVOL, and the process is awaited. In this step, if the write target VOL is SVOL, the snapshot pair of the corresponding destination SVOL is PAIR. Therefore, in order to reflect the write at the movement source on the movement destination, a write command may be transmitted to the movement destination PVOL. The migration destination storage apparatus 500B that has received the write command executes the write reception process shown in FIG.

Step 3212: A completion notification is returned to the migration destination storage apparatus 500B, and an exclusive release instruction is awaited.

Step 3213: The acquired exclusion is released. This step is not executed until there is an exclusion release instruction from the migration destination storage apparatus 500B as described in step 3212.

As described above, in the third embodiment, the migration copy process is executed by a write command. Therefore, the migration destination storage apparatus 500B has basic functions as a storage apparatus, that is, a function for processing a write from the host and a snapshot function. Thus, the snapshot VOL can be moved. In this embodiment, since all snapshot pairs of the migration destination VOL are set to PAIR and transition is made to PSUS during the migration copy process, in the write process during the migration copy process, whether the pair status of the migration copy of the write target VOL is PAIR or not. To determine whether the snapshot pair of the corresponding destination VOL is PAIR or PSUS, and if the write target VOL is SVOL, write to the PVOL at the destination or write to the SVOL Decide what to do.

Finally, the usefulness of the present invention will be described again. The usefulness of the present invention is that the migration source snapshot configuration can be transferred to the migration destination while maintaining the snapshot configuration without stopping I / O from the host. Suppose that the technology for migrating a normal volume while accepting I / O is applied to the migration of a snapshot configuration. In this case, not only the difference data but also all the data constituting each SVOL is copied from the migration source to the migration destination for the SVOLs of all generations constituting the replication of the PVOL by chunk reference to the PVOL and the pool volume. There is a need. Since the SVOL generation is as large as 1000 generations, a full copy, which is a copy of all data constituting this volume, takes a lot of time.

The technology of the present invention makes it possible to accept I / O from the host during the migration while performing the migration from the migration source of the snapshot configuration to the migration destination in a short time without performing the above-described full copy. To reduce the time for moving copy, for example, when moving copy is for replacement purposes, it is possible to reduce the maintenance cost of the old device that is the moving source, or to operate the new device that is the moving destination device early. There is an advantage that performance can be improved and new functions can be used at an early stage. Therefore, for the first time, the present invention makes it possible to achieve high availability of non-stop I / O during volume migration using the snapshot function while retaining the advantages of the above-described moving copy in a short time. .

The present invention can also be characterized by the following viewpoints, for example.
The storage system includes a first storage device and a second storage device. The first storage device provides a snapshot configuration with the first primary volume, the first pool volume, and the first secondary volume. The first primary volume includes a plurality of chunks, and has a first real volume ID and a first virtual volume ID associated with the first real volume ID. The first pool volume includes a plurality of chunks and stores differential data of the first primary volume. The first secondary volume is a replica of the first primary volume at a predetermined point in time, and includes a plurality of chunks each referring to either the first primary volume chunk or the first pool volume chunk, 2 real volume IDs and a second virtual volume ID associated with the second real volume. The first storage device further manages the reference destinations of the plurality of chunks of the first secondary volume, and upon receiving an identifier request for the first primary volume, responds with the first virtual volume ID, When the request for the identifier for the secondary volume is received, the second virtual volume ID is returned. The second storage device provides a snapshot configuration with the second primary volume, the second pool volume, and the second secondary volume. The second primary volume includes a plurality of chunks and has a third real volume ID and a first virtual volume ID associated with the third real volume ID. The second pool volume includes a plurality of chunks and stores differential data of the second primary volume. The second secondary volume is a replica of the second primary volume at a predetermined point in time and includes a plurality of chunks each referring to either the second primary volume chunk or the second pool volume chunk, 4 real volume IDs and a second virtual volume ID associated with the fourth real volume ID. The second storage device further manages the reference destinations of the plurality of chunks of the second secondary volume, and upon receiving an identifier request for the second primary volume, responds with the first virtual volume ID, When the request for the identifier for the secondary volume is received, the second virtual volume ID is returned. The first storage device migrates the snapshot configuration from the first storage device to the second storage device. At this time, when the first chunk of the first secondary volume refers to the second chunk of the first primary volume, the first chunk is migrated after the second chunk is migrated. A storage system characterized by that.

100 Host computer 200 Management server 500 Storage device 520 Real volume 530 Pool 600 Virtual storage device

Claims

A first primary volume containing a plurality of chunks;
A first pool volume that includes a plurality of chunks and stores differential data of the first primary volume;
A first secondary volume that is a replica of the first primary volume at a predetermined time point, each including a plurality of chunks that refer to either the first primary volume chunk or the first pool volume chunk Volume,
Providing a first snapshot configuration comprising:
Managing reference destinations of the plurality of chunks of the first secondary volume;
A first storage device;
A second primary volume containing a plurality of chunks;
A second pool volume that includes a plurality of chunks and stores differential data of the second primary volume;
A second secondary volume that is a replica of the second primary volume at a predetermined time point, each including a plurality of chunks that refer to either the second primary volume chunk or the second pool volume chunk Volume,
Providing a second snapshot configuration having:
Managing reference destinations of the plurality of chunks of the second secondary volume;
A second storage device;
A storage system comprising:
The first storage device
When migrating the first snapshot configuration from the first storage device to the second storage device,
Acquires the exclusion of the chunk to be migrated, and sends snapshot allocation information, which is management information of the chunk to be migrated, to the second storage device,
During the transition, when receiving a write request from the host device,
Acquiring exclusive of the first chunk related to the write request, the chunk related to the write request and the second chunk related by the reference, and executing the write processing related to the write request,
A storage system characterized by that.
The snapshot allocation information related to the chunk of the first secondary volume is:
Included in the third secondary volume is information identifying the destination volume of the transmission, information identifying the destination chunk of transmission, information indicating the type of reference of the chunk of the first secondary volume, and , Referring to the same chunk of the first pool volume as the chunk of the first secondary volume, and if there is a chunk that has been migrated, information for identifying the third secondary volume, and the third If there is no secondary volume, the chunk data of the first secondary volume, and
The storage system according to claim 1.
The first storage device
When the data related to the write request is written to the first chunk and the migration of the first chunk is completed, the data related to the write request is transmitted to the second storage device, and the first chunk If chunk migration is incomplete, data related to the write request is not transmitted to the second storage device.
The storage system according to claim 1.
The first storage device transmits the snapshot allocation information of the second chunk to the second storage device together with the data related to the write request.
The storage system according to claim 3.
The second storage device
When a write request is received from the host device, the write request is transferred to the first storage device,
When the transfer of the chunk related to the transferred write request has been completed, the data related to the write request is written to the chunk related to the write request of the second storage device,
The storage system according to claim 1, wherein when the transfer of the chunk related to the transferred write request is incomplete, the data related to the write request is not written to the chunk related to the write request of the second storage device.
The second storage device
When the read request is received from the higher-level device, if the migration of the chunk related to the read request is completed, the data related to the request is read from the second storage device,
If the migration of the chunk related to the read request is incomplete, transfer the read request to the first storage device,
When the first storage device receives the transferred read request from the second storage device, the first storage device reads data related to the read request from the first storage device, and transmits the data to the second storage device.
The storage system according to claim 1.
A first primary volume containing a plurality of chunks;
A first pool volume that includes a plurality of chunks and stores differential data of the first primary volume;
A first secondary volume that is a replica of the first primary volume at a predetermined time point, each including a plurality of chunks that refer to either the first primary volume chunk or the first pool volume chunk Volume,
Providing a first snapshot configuration comprising:
Managing reference destinations of the plurality of chunks of the first secondary volume;
A first storage device;
A second primary volume containing a plurality of chunks;
A second pool volume that includes a plurality of chunks and stores differential data of the second primary volume;
A second secondary volume that is a replica of the second primary volume at a predetermined time point, each including a plurality of chunks that refer to either the second primary volume chunk or the second pool volume chunk Volume,
Providing a second snapshot configuration having:
Managing reference destinations of the plurality of chunks of the second secondary volume;
A second storage device;
A storage system control method comprising:
When migrating the first snapshot configuration from the first storage device to the second storage device,
Acquires the exclusion of the chunk to be migrated, and sends snapshot allocation information, which is management information of the chunk to be migrated, to the second storage device,
During the transition, when receiving a write request from the host device,
Acquiring exclusive of the first chunk related to the write request, the chunk related to the write request and the second chunk related by the reference, and executing the write processing related to the write request,
A storage system control method.
The snapshot allocation information related to the chunk of the first secondary volume is:
Included in the third secondary volume is information identifying the destination volume of the transmission, information identifying the destination chunk of transmission, information indicating the type of reference of the chunk of the first secondary volume, and , Referring to the same chunk of the first pool volume as the chunk of the first secondary volume, and if there is a chunk that has been migrated, information for identifying the third secondary volume, and the third If there is no secondary volume, the chunk data of the first secondary volume, and
The storage system control method according to claim 7.
When the data related to the write request is written to the first chunk and the migration of the first chunk is completed, the data related to the write request is transmitted to the second storage device, and the first chunk If chunk migration is incomplete, data related to the write request is not transmitted to the second storage device.
The storage system control method according to claim 7.
Along with the data relating to the write request, the snapshot allocation information of the second chunk is transmitted to the second storage device.
The storage system control method according to claim 9.
Upon receiving a write request from the host device to the second storage device, the write request is transferred to the first storage device,
When the transfer of the chunk related to the transferred write request has been completed, the data related to the write request is written to the chunk related to the write request of the second storage device,
If the transfer of the chunk related to the transferred write request is incomplete, the data related to the write request is not written to the chunk related to the write request of the second storage device.
The storage system control method according to claim 7.
Upon receiving a read request from the host device to the second storage device, if the migration of the chunk related to the read request is completed, the data related to the request is read from the second storage device,
If the migration of the chunk related to the read request is incomplete, transfer the read request to the first storage device,
When the first storage device receives the read request transferred from the second storage device, the first storage device reads data related to the read request from the first storage device,
Send to the second storage device,
The storage system control method according to claim 7.