JP2012014313A - Storage device, control method therefor and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a storage device in which power consumption is reduced at a low cost even for a storage like a primary storage, which has a difficulty of classifying and recording less accessed data.SOLUTION: A storage medium recorded with redundant data is discontinued by utilizing the redundancy of data between the storage media constituting a storage. The data retrieval is then performed only by other storage media. Also, data updates are recorded in the storage medium having higher speed and lower power than the discontinued storage medium. The updated data are reflected collectively when the discontinued storage medium is reactivated.

Description

  The present invention relates to a storage apparatus, a control method therefor, and a program. In particular, the present invention relates to a storage device that stops a part of storage devices that form a redundant array.

  With the recent evolution of information communication technology and the expansion of its application fields, the demand for the amount of data handled in the information processing system and the immediacy of the data is increasing year by year. For this reason, the access performance is speeded up so that a large amount of data can be processed in a short time with respect to a storage that stores data represented by an HDD (Hard Disk Drive).

  In addition, a large number of information communication devices including storage are integrated in a data center and operated as a large-scale system. In such a data center, an operation cost proportional to the size of the installation site is required. For this reason, the density of storage apparatuses has been increased for the purpose of reducing the operation cost of the entire system.

  On the other hand, demands for power saving for information communication devices including storage are increasing. The first reason is restrictions on power consumption and heat generation in data centers and the like. Due to the increase in power consumption of storage, the electricity charge during system operation increases. In addition, the upper limit of power consumption that can be accepted as a whole is set for the rack that accommodates the information communication equipment, and when the upper limit of the power consumption of the entire rack is exceeded, it is necessary to separate it into another rack. Such separation leads to an increase in the installation cost of the apparatus and equipment. In addition, an increase in power consumption often leads to an increase in calorific value, resulting in an increase in air conditioning costs. Therefore, it is necessary to reduce the power consumption of each device including the storage. The second reason is an increase in social demand for environmental impact. Companies that use information and communication equipment are required to reduce their environmental impact. Under such circumstances, information and communication devices are also required to reduce the environmental load through reduction of power consumption.

  Under such a background, there are various technologies aiming at power saving for storage. Specific examples thereof include MAID (Massive Array of Idle Disks). MAID is mainly applied to storage using an HDD. The HDD that is powered on consumes a lot of power because the disk on which data is recorded continues to rotate while there is no access. In MAID, power consumption is suppressed by stopping an HDD that is not accessed or is not frequently accessed in the storage. However, when accessing the data stored in the stopped HDD, it is necessary to turn on the HDD and wait for the start-up to complete, so the access performance is greatly reduced. Therefore, in the MAID, the performance degradation is suppressed by limiting the data to HDDs storing data that is not scheduled to be accessed or data with low access frequency and stopping the operation.

  Further, power saving can be realized by using a storage medium that originally has low power consumption as a storage medium constituting the storage. In recent years, an SSD (Solid State Drive) using a flash memory has started to spread in place of an HDD mainly used as a storage medium constituting a storage. SSDs have higher access performance than HDDs and do not have any physically movable parts, so they consume less power regardless of whether they are accessed. Therefore, power saving of the storage can be achieved by simply replacing the HDD with the SSD as it is.

  Furthermore, in the storage apparatus having a plurality of storage media described in Patent Document 1, when the power consumption of the system is monitored and a predetermined value is exceeded while maintaining the disk mirror configuration, the operation of the disk on one side of the mirroring configuration A disk array system is disclosed in which the power consumption of the entire system is reduced by recording data in a nonvolatile memory during this period.

JP 2009-187450 A

  The following analysis has been made from the viewpoint of the present invention.

  As described above, MAID is a technology that focuses on storage with low access frequency. On the other hand, a storage storing frequently accessed data, that is, a primary storage exists on the system. Such primary storage is required to have particularly high access performance. Therefore, a high-speed HDD is usually used for the primary storage. However, while a high-speed HDD has high performance, it consumes more power and generates heat than a normal HDD. For this reason, the restrictions on power and heat generation in the system become more severe as compared with a normal HDD. In other words, primary storage has particularly strict requirements in terms of high density with restrictions on power and heat generation.

  The effect of power saving using MAID is limited for such primary storage. In the first place, power saving is achieved by stopping HDDs that are not accessed by MAID or that are not accessed frequently. In contrast, the primary storage is intended to store frequently accessed data. Therefore, there is no HDD to be stopped by MAID in the primary storage, and the power saving effect cannot be obtained.

  On the other hand, when the SSD is adopted as a storage medium instead of the HDD, the access performance of the entire storage can be improved simultaneously with the power saving. In addition, due to improved access performance, the number of devices themselves can be reduced, and higher system density can be expected. Therefore, replacing all HDDs with SSDs is effective in all of high performance, high density, and power saving. However, this is not practical for cost reasons. This is because the SSD has a smaller storage capacity than the HDD and is expensive, so that the cost required to replace the SSD with the same capacity as the HDD is very high.

  As described above, it is necessary to reduce the power consumption of a storage that is frequently accessed like the primary storage, requires high access performance, and consumes a large amount of power.

  As described above, there are problems to be solved in the prior art.

  In one aspect of the present invention, a storage device with high performance and low power consumption, a control method thereof, and a program are desired.

  According to the first aspect of the present invention, a plurality of storage media that are used as storage destinations of data including redundant data and can be brought into a stopped state independently, and can be accessed at higher speed than the storage media A non-volatile storage medium used as a temporary recording destination of the redundant data when there is a data write request to the storage medium that is stopped among the plurality of storage media, and when the data read request, A storage device is provided that includes a controller that inquires of the non-volatile storage medium whether or not there is redundant data corresponding to the data requested to be read, and reads the redundant data from the non-volatile storage medium if recorded. The

  According to the second aspect of the present invention, a plurality of storage media that are used as storage destinations of data including redundant data and can be brought into a stopped state independently, and can be accessed at higher speed than the storage media A non-volatile storage medium, wherein the redundant data is stored in the non-volatile storage medium in response to a data write request to a storage medium that is stopped among the plurality of storage media. A step of recording, a step of inquiring of the nonvolatile storage medium whether or not there is redundant data corresponding to the data for which the read request has been received, and a nonvolatile storage medium if the redundant data is recorded. And reading the redundant data from the storage device.

  According to the third aspect of the present invention, a plurality of storage media that are used as storage destinations of data including redundant data and can be independently brought into a stopped state, and can be accessed at higher speed than the storage media. A non-volatile storage medium, and a program executed by a computer that controls a storage apparatus, wherein the redundant data is stored in response to a data write request to a stopped storage medium among the plurality of storage media. If the process of recording in the nonvolatile storage medium, the process of querying the nonvolatile storage medium for the presence or absence of redundant data corresponding to the data that has received the read request at the time of a data read request, and the redundant data are recorded, A program for causing the computer to execute a process of reading the redundant data from the nonvolatile storage medium is provided.

  According to each aspect of the present invention, it is possible to provide a storage device with high access performance and low power consumption by partially stopping the operation of the storage medium and using the nonvolatile memory as a cache.

It is a figure for demonstrating the outline | summary of this invention. 1 is a diagram illustrating a configuration of an entire system including a storage apparatus according to an embodiment of the present invention. 1 is a block diagram of an entire storage apparatus according to a first embodiment of the present invention. It is a block diagram which shows the structure of the access request process part 21 of FIG. It is a block diagram which shows the structure of the command execution part 26 in FIG. It is a figure which shows the content of the access request management table. It is a figure which shows the content of the command management table. It is a figure which shows the content of a cache management table. 4 is a flowchart showing the operation of the entire storage apparatus of FIG. It is a flowchart which shows the data reading process of the command control execution part 261 of FIG. 6 is a flowchart showing data write processing of a command control execution unit 261 in FIG. 5. FIG. 6 is a flowchart showing a process in which a command control execution unit 261 in FIG. 6 is a flowchart showing a process in which the command control execution unit 261 of FIG. 5 reflects the update data of the nonvolatile cache 24 on a storage medium. It is the figure shown about the area | region in the address space of HDD in 2nd Embodiment. It is the figure which showed the order of the access request from the host in 2nd Embodiment.

  First, the outline of the present invention will be described with reference to FIG. As described above, it is very difficult to apply MAID to the primary storage and to replace it with a power-saving storage medium such as SSD. In addition, in a storage system having a plurality of storage media, a configuration having data redundancy among a plurality of storage media such as RAID (Redundant Arrays of Independent Disks) is employed in order to ensure fault tolerance for the storage media. Is done. In such a storage, the data main body and the parity information for the data or the mirror (replica) are recorded in different storage media. This means that when data is read from a storage having a redundant configuration among a plurality of storage media, access to the storage medium storing the redundant data is unnecessary. Therefore, if the storage medium storing the target data can be accessed, the data can be read even when the storage medium storing the redundant data is stopped.

  On the other hand, when data is written to the above-mentioned storage, it is necessary to update redundant data such as parity and mirror for the data together with the data body. That is, it is necessary to write to both the storage medium storing the data body and the storage medium storing redundant data. However, if the updated redundant data can be temporarily recorded on another storage medium, the writing process can be performed without accessing the storage medium on which the redundant data is recorded.

  Therefore, in the present invention, in a storage having a redundant configuration between a plurality of storage media, the storage media for storing redundant data such as parity and mirror are operated in a stopped state. At that time, the data to be stored in the stopped storage medium is temporarily stored in another type of storage medium that is faster and saves power than the storage medium constituting the storage. The data temporarily stored in the storage medium is updated by activating the storage medium that is the original storage destination at any time until the free space becomes insufficient.

  Specifically, a high-performance and power-saving medium such as SSD is adopted as a storage medium (hereinafter referred to as a non-volatile cache) for temporarily recording updated data and an address corresponding to the data. When the target data for read access exists on the nonvolatile cache, the data is read, and the data recorded in the nonvolatile cache is written when the storage medium storing the redundant data is accessible.

  In this way, by using a non-volatile cache separately from the storage medium that is mainly used in the storage, the storage medium that can be stopped is stopped among the storage media, thereby realizing power saving. In addition, a high-performance medium is used for the nonvolatile cache, and when a read request is made to the storage, first, the existence of data is confirmed in the nonvolatile cache, and if data exists, the data of the nonvolatile cache is used. Improve overall storage performance.

[First Embodiment]
Next, the first embodiment of the present invention will be described in more detail with reference to the drawings. FIG. 2 is a diagram showing a configuration of the entire system including the storage apparatus according to the embodiment of the present invention. This system includes a host that accesses data, a storage 2 composed of a plurality of storage media, and a network 3 that connects them. The number of hosts connected to the network is not limited to one, and a plurality of hosts can access the storage 2. The network 3 is not limited to a form in which a plurality of hosts can be connected to the storage 2 but may be configured to directly connect the host and the storage 2.

  Next, the configuration of the storage 2 will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of the storage apparatus according to an embodiment of the present invention. The storage 2 includes an access request processing unit 21, a storage medium 22, a redundant data storage medium 23, a nonvolatile cache 24, a storage medium power supply control unit 25, and a command execution unit 26.

  The access request processing unit 21 receives an access request from the host 1 and generates a command corresponding to the request. Further, the access request processing unit 21 restores the original data from the read redundant data. Conversely, it also has a function of generating redundant data to be stored in the redundant data storage medium 23 from the written data.

  Next, the configuration of the access request processing unit 21 will be described in detail. FIG. 4 is a block diagram of the internal configuration of the access request processing unit 21. The access request processing unit 21 includes an access interpretation unit 211, an access request information storage unit 212, a data restoration unit 213, and a redundant data generation unit 214.

  The access interpretation unit 211 receives an access request from the host 1, and generates commands for the storage medium 22 and the redundant data storage medium 23 in response to the request. The access interpretation unit 211 notifies the command execution unit 26 of the generated command, and records management information of the command in the access request information storage unit 212. If the generated command is a write command, the data to be written is passed to the command execution unit 26. After that, when the completion notice for the data read by the Read command or the Write command is received from the command execution unit 26, the completion notice for the read data or write request is returned to the corresponding host 1 as the request source. .

  Further, when the access request from the host 1 is completed, the access interpretation unit 211 deletes the management information of the corresponding access request recorded in the access request information storage unit 212 at an arbitrary timing. The access request information storage unit 212 manages the correspondence between the access request received by the access interpreting unit 211 and the command generated by the access interpreting unit 211 using the access request management table (see FIG. 6).

  The data restoration unit 213 generates a data main body corresponding to the redundant data by combining the redundant data read from the redundant data storage medium 23 with the data read from the other storage medium 22 and the redundant data storage medium 23.

  The redundant data generation unit 214 generates redundant data to be recorded in the redundant data storage medium 23 from data written to the storage 2.

  Next, the storage medium included in the storage 2 will be described. The storage medium 22 permanently records the data written in the storage 2. The storage medium 22 performs data read or write access according to a command from the command execution unit 26. The operation state and the stop state can be switched by an instruction from the command execution unit 26.

  The redundant data storage medium 23 permanently records redundant data such as parity or mirror (replication) for the data stored in the storage medium 22. As with the storage medium 22, the redundant data storage medium 23 can perform read or write access of data by a command and can be switched between an operation state and a stop state.

  The nonvolatile cache 24 has higher access performance than the storage medium 22 and the redundant data storage medium 23, and temporarily stores the data received from the host 1 before it is written to the stopped storage medium 22 or the redundant data storage medium 23. That is, the nonvolatile cache 24 has a role of holding data to be written to the storage medium 22 or the redundant data storage medium 23 while the storage medium is stopped while the storage medium is stopped. Further, the non-volatile cache 24 not only holds data written to the stopped storage medium 22 or the redundant data storage medium 23, but also temporarily uses the data in the storage medium 22 and the redundant data storage medium 23 as a cache. The processing performance of read requests and write requests to the storage 2 is improved.

  The storage medium power control unit 25 records the operation or stop of the storage medium 22 and the redundant data storage medium 23, and the power supply is controlled to switch between the operation state and the stop state of these recording media. This completes the description of the storage medium.

  Next, the command execution unit 26 will be described. The command execution unit 26 receives a Read command or a Write command from the access request processing unit 21. If the command is a Read command, it is confirmed whether the requested data is recorded in the nonvolatile cache 24. If the command is recorded, the data at the corresponding address is read from the nonvolatile cache 24 and the data is accessed. It returns to the request processing unit 21. On the other hand, if the requested data does not exist in the nonvolatile cache 24, the data or redundant data corresponding to the data is read from the storage medium 22 or the redundant data storage medium 23. In the case of the Write command, when the storage medium 22 is stopped, the write data body is written into the nonvolatile cache 24 and the write data is held until the storage medium 22 transitions to the operating state. Similarly, when the redundant data storage medium 23 is stopped, the redundant data generated by the redundant data generation unit 214 is written to the non-volatile cache 24, and the redundant data is retained until the redundant data storage medium 23 changes to the operating state. To do.

  When the data body is held in the nonvolatile cache 24, the command execution unit 26 switches the storage medium 22 from the stopped state to the operating state at an arbitrary timing, and writes the write data held in the nonvolatile cache 24 to the storage medium 22. . Similarly, when the redundant data for the write data is held in the nonvolatile cache 24, the redundant data storage medium 23 is switched from the stopped state to the operating state at an arbitrary timing, and the redundant data held in the nonvolatile cache 24 is stored in the redundant data storage. Write to medium 23.

  In addition, the command execution unit 26 does not go through the non-volatile cache 24, and when the command issued to the storage medium 22 and the redundant data storage medium 23 is completed, or when reading or writing to the non-volatile cache 24 is completed, the access request processing unit 21 On the other hand, in the case of the Read command, the read data is notified, and in the case of the Write command, the writing completion is notified.

  Next, the command execution unit 26 will be described in detail. FIG. 5 is a block diagram of the internal configuration of the command execution unit 26. The command execution unit 26 includes a command execution control unit 261, a command management information storage unit 262, and a data update control unit 263.

  The command execution control unit 261 receives a Read command or a Write command from the access request processing unit 21, and stores information necessary for execution management of the access command in the command management information storage unit 262. If the received command is a Read command, the target data is read from the nonvolatile cache 24, the storage medium 22, or the redundant data storage medium 23. If the received command is a write command, the operating state of the storage medium 22 or the redundant data storage medium 23 to which data is written is obtained from the storage medium power supply control unit 25. Redundant data is written directly to the storage medium 22 or the redundant data storage medium 23. On the other hand, when the write destination storage medium 22 or the redundant data storage medium 23 is in a stopped state, the write data or the redundant data is temporarily stored in the nonvolatile cache 24. If there is write data recorded in the non-volatile cache 24 but not stored in the storage medium 22 and the redundant data storage medium 23, the storage medium 22 and the redundant data storage medium 23 are instructed by the data update control unit 263. Write data to. When the access to the storage medium 22, the redundant data storage medium 23, or the nonvolatile cache 24 is completed, the command returns data to the access request processing unit 21 if it is a Read command, and a notification of completion if it is a Write command.

  The command management information storage unit 262 records a command management table for managing commands received from the access request processing unit 21 (see FIG. 7).

  The data update control unit 263 monitors the state of the non-volatile cache 24, and when the free space of the area for storing the update data in the non-volatile cache 24 is insufficient, or the storage medium power The storage medium 22 or the redundant data storage medium 23 is switched to the operating state through the control unit 25 so that the update data of the nonvolatile cache 24 can be written. Further, when the storage medium 22 or the redundant data storage medium 23 in which the update data is to be stored is in an operating state, the update data on the nonvolatile cache 24 is transferred to the storage medium 22 or the redundant data storage medium 23 with respect to the command execution control unit 261. Make a request to write.

  Here, various tables used in the storage 2 will be described. FIG. 6 is a diagram showing the contents of the access request management table described above. The access request management table holds information associated with each access request from the host 1. As information accompanying the access request, a unique ID given for each access request, an access type (represented as R / W in FIG. 6) indicating whether it is a read request or a write request, and an address in the entire storage 2 The address of the first block in the space, the access length, the requesting source host of the access request, and one or more of each storage added to the command requested to process the command execution unit 26 to process the access request The ID of the access command for the medium (hereinafter referred to as command ID) is recorded. The access request management table is referred to by the access interpreter 211 when information associated with each access request from the host 1 is required.

  Next, the command management table will be described. FIG. 7 is a diagram showing the contents of the command management table. The command management table holds information accompanying commands issued to the storage medium 22, redundant data storage medium 23, or nonvolatile cache 24 in order to process a command requested to be processed by the access request processing unit 21. The information accompanying the command includes an ID used to identify each access command for the storage medium in the access request management table, and an access type indicating whether it is a read request or a write request (represented as R / W in FIG. 7). ), The ID indicating the access destination that issues the command, the address of the head block of the access destination, the access length, and the write command, the data written in the storage medium 22 or the redundant data written in the redundant data storage medium 23 Store the data.

  Next, the non-volatile cache management table will be described. FIG. 8 shows the contents of the non-volatile cache management table. The non-volatile cache management table is managed by the non-volatile cache 24 or the command execution unit 26, and manages data for updating the storage medium 22 and redundant data storage medium 23 stored in the non-volatile cache 24 or updated data. This table holds information necessary for As information necessary for management, since the data has been updated, a dirty flag indicating whether update processing is required for the storage medium 22 and the redundant data storage medium 23 when the data is evicted from the nonvolatile cache 24 (updated in the case of 1). ), The address of the first block of data, the access length, and the address of the first block of the data storage destination on the nonvolatile cache 24 are stored. Further, data that has already been updated with respect to the storage medium 22 and the redundant data storage medium 23 is used as a cache at the time of read processing.

  Next, the operation of the storage apparatus according to this embodiment will be described with reference to the drawings. FIG. 9 shows the operation of the entire storage 2 when an access request is made from the host 1 to the storage 2.

  In step S101, the access interpretation unit 211 interprets an access request from the host 1, and generates an access command (Read command or Write command) for the storage medium 22 or the redundant data storage medium 23.

  In step S102, it is determined whether or not the command generated in step S101 is a write command for the redundant data storage medium 23. If it is not a write command, the process proceeds to step S104. If it is a write command, the process proceeds to step S103.

  In step S103, the access interpretation unit 211 uses the redundant data generation unit 214 to generate redundant data that is actually written to the redundant data storage medium 23 from the write data received in step S101.

  In step S104, the access interpretation unit 211 stores the management information of the access request received in step S101 and the correspondence information of the generated command in the access request information storage unit 212. The information stored at this time is the content shown in the access request management table of FIG. Further, the access interpretation unit 211 requests the command execution unit 26 to process the command generated in step S101. If the command is a write command, the write target data is passed to the command execution unit 26.

  In step S105, the command execution control unit 261 receives the command passed to the command execution unit 26 in step S104, and records management information of the command in the command management information storage unit 262 in the form of a command management table.

  In step S106, if the command received by the command execution control unit 261 is a write command, the process proceeds to step S111. If the command is not a write command, the process proceeds to step S107.

  In step S107, since the command received by the command execution control unit 261 is a Read command, access processing for the Read command is executed. At this time, the command execution control unit 261 executes Read command processing on the storage medium 22, the redundant data storage medium 23, or the nonvolatile cache 24.

  Subsequently, in step S108, the command execution control unit 261 passes the read result to the access request processing unit 21 when the processing of the issued Read command is completed. In addition, the command execution control unit 261 deletes the above-described read command management information recorded in the command management information storage unit 262 at an arbitrary timing.

  In step S109, when the Read command completed in steps S107 and S108 is a command for the storage medium 22, the process proceeds to step S113. If the Read command completed in steps S107 and S108 is a command for the redundant data storage medium 23, the process proceeds to step S110.

  In step S110, the data restoration unit 213 of the access request processing unit 21 restores the data requested to be read from the storage 2 from the redundant data received in step S108.

  Subsequently, in step S111, when the command received by the command execution control unit 261 is a write command, the command execution control unit 261 executes an access process for the write command to execute the storage medium 22, the redundant data storage medium 23, or the nonvolatile data storage medium. Data is stored in the cache 24. Further, when the storage of the write data is completed in step S111, the command execution control unit 261 notifies the access request processing unit 21 of the completion of the write process.

  In step S112, the command execution control unit 261 deletes the above-mentioned write command management information recorded in the command management information storage unit 262 at an arbitrary timing.

  In step S113, the access interpretation unit 211 receives a read result or a write completion notification from the command execution control unit 261, and searches the access request management table for a request source corresponding to the command result. The access interpretation unit 211 notifies the host 1 that is the request source of the corresponding access request of the data that is the result of the Read access request or the completion notification of the Write access request. When the access interpretation unit 211 completes the command notification or completion notification processing to the host 1, the management information of the command stored in the access request information storage unit 212 is deleted at an arbitrary timing.

  The above steps S101 to S113 are processing of the entire storage 2 when an access request is made from the host 1 to the storage 2.

  Next, the procedure in step S107 described above will be described with reference to FIG.

  In step S <b> 201, the command execution control unit 261 issues a Read command received from the access request processing unit 21 to the nonvolatile cache 24 and inquires whether there is data to be read.

  In step S202, the non-volatile cache 24 refers to the non-volatile cache management table and confirms whether or not the read destination data exists. If there is access data for the Read command in the nonvolatile cache 24, the process proceeds to step S203. If not, the process proceeds to step S204.

  In step S203, the command execution control unit 261 receives data from the nonvolatile cache 24, and ends the process of step S107.

  In step S204, since the read command access destination data does not exist in the nonvolatile cache 24, the command execution control unit 261 stores the read command in the storage medium 22 or the redundant data storage medium 23 according to the read command access destination. Issue a Read command.

  In step S205, the command execution control unit 261 receives the result of the Read command for the storage medium 22 or the redundant data storage medium 23 issued in step S204. Further, the result of the Read command or redundant data is stored in the nonvolatile cache 24 as necessary, and the nonvolatile cache management table is updated. However, recording of the result of the Read command in the nonvolatile cache 24 is an arbitrary process.

  The processing from step S201 to step S205 is performed, and the processing of step S107 for processing the read command of the command execution control unit 261 is completed.

  Next, the procedure in step S111 described above will be described with reference to FIG.

  In step S <b> 301, the command execution control unit 261 acquires the operation state of the write destination storage medium 22 or the redundant data storage medium 23 through the storage medium power control unit 25.

  In step S302, it is determined whether the access destination of the write command is stopped. If the write command access destination storage medium 22 or the redundant data storage medium 23 is operating, the process proceeds to step S303. If the write command access destination storage medium 22 or the redundant data storage medium 23 is stopped, the process proceeds to step S304.

  In step S303, the command execution control unit 261 issues a Write command to the write destination storage medium 22 or the redundant data storage medium 23 and ends the process.

  In step S304, the command execution control unit 261 inquires whether there is an available area in the nonvolatile cache 24 for storing the write data received from the access request processing unit 21.

  In step S305, when there is no free area on the nonvolatile cache 24 as a result of the inquiry, the process proceeds to step S306. On the other hand, if there is a free area, the process proceeds to step S307.

  In step S306, the command execution control unit 261 causes the write command processing to wait until an area available for writing the write data received from the access request processing unit 21 to the nonvolatile cache 24 is secured. Further, the data update control unit 263 is notified that the free space is insufficient. After the end of standby, the process is restarted from the determination in step S305.

  In step S307, the command execution control unit 261 issues a Write command to the nonvolatile cache 24, writes the write data, and ends the process.

  The processes from step S301 to step S307 are performed, and the process of step S111 for performing the process for the write command of the command execution control unit 261 is completed.

  Next, a procedure for the data update control unit 263 to solve the capacity shortage of the nonvolatile cache 24 will be described with reference to FIG. The data update control unit 263 eliminates the lack of capacity of the nonvolatile cache 24 by discarding the read data held in the nonvolatile cache 24 or writing the write data to the storage medium 22 or the redundant data storage medium 23.

  In step S401, the data update control unit 263 selects and deletes any one of the data whose dirty flag in the nonvolatile cache management table is 0, that is, data that does not need to be updated.

  In step S402, it is determined whether or not the erasure is successful. If successful, the process proceeds to S403, and if unsuccessful, the process proceeds to step 404.

  In step S403, it is determined whether or not the free capacity of the nonvolatile cache 24 is sufficient. If the free capacity of the nonvolatile cache 24 is sufficient, the data update control unit 263 terminates the capacity shortage elimination process. When the free capacity of the non-volatile cache 24 is insufficient, the process returns to the process of step S401 to continue the process for eliminating the shortage of capacity of the non-volatile cache 24.

  In step S404, the data update control unit 263 selects the arbitrary one from the data that needs to reflect the update to the storage medium 22 or the redundant data storage medium 23, that is, the dirty flag of the nonvolatile cache management table is 1, and executes the command. The control unit 261 is instructed to execute update reflection.

  In step S405, the command execution control unit 261 acquires the operation state of the storage medium 22 or the redundant data storage medium 23 that is the storage destination of the data selected in step S404 through the storage medium power control unit 25.

  In step S406, it is determined whether the storage medium 22 or the redundant data storage medium 23 that is the storage destination of the selected data is stopped. If it is stopped, the process of step S407 is executed. If it is in operation, the command execution control unit 261 executes the processing after step S408.

  In step S407, the command execution control unit 261 switches the storage destination storage medium 22 or the redundant data storage medium 23 to the operating state through the storage medium power control unit 25.

  In step S408, the command execution control unit 261 deletes the write command data temporarily held in the nonvolatile cache 24 after completing the execution of the write command instructed in step S404.

  In step S409, it is determined whether or not the free capacity of the nonvolatile cache 24 is insufficient. If it is determined that there is an insufficiency, the process returns to step S404, and the process of eliminating the capacity shortage of the nonvolatile cache 24 is continued. If it is determined that the free space is sufficient, the process proceeds to step S410.

  In step S410, the command execution control unit 261 does not lose the redundancy required for reading data from the storage medium 22 and the redundant data storage medium 23 through the storage medium power control unit 25, and a single or a plurality of arbitrary storage media To stop. At this time, the selection of the storage medium to be stopped from the storage medium 22 and the redundant data storage medium 23 is not necessarily the same as the storage medium stopped before the execution of step S401.

  The processing from the above steps S401 to S410 is performed, and the data update control unit 263 discards the read data stored in the non-volatile cache 24 or writes the write data to the storage medium 22 or the redundant data storage medium 23 so that the non-volatile cache The process of eliminating the 24 capacity shortage is completed.

  Next, referring to FIG. 13, the data update control unit 263 writes the write data recorded in the nonvolatile cache 24 to the storage medium 22 or the redundant data storage medium 23, so that a certain time has elapsed or the storage 2 A procedure for reflecting the update data temporarily held in the non-volatile cache according to the conditions such as the load of will be described.

  In step S501, the data update control unit 263 selects the arbitrary one from the data that needs to reflect the update to the storage medium 22 or the redundant data storage medium 23, that is, the dirty flag of the nonvolatile cache management table is 1, and executes the command. The control unit 261 is instructed to reflect the update.

  In step S502, the command execution control unit 261 acquires the operation state of the storage medium 22 or the redundant data storage medium 23 that is the storage destination of the data selected in step S501 through the storage medium power control unit 25.

  In step S503, it is determined whether or not the storage medium that is the data storage destination is stopped. If it is stopped, step S504 is executed. If it is in operation, the process proceeds to step S505.

  In step S504, the command execution control unit 261 switches the storage destination storage medium 22 or the redundant data storage medium 23 to the operating state through the storage medium power control unit 25.

  In step S505, the command execution control unit 261 deletes the write command data temporarily stored in the nonvolatile cache 24 after completing the execution of the write command instructed in step S501.

  In step S506, it is determined whether or not there is data in the nonvolatile cache 24 in which the dirty flag on the nonvolatile cache management table is set, that is, there is unupdated data. If it exists, the process returns to the process of step S501 and the update reflection process of the write data on the nonvolatile cache 24 is continued. If it does not exist, the process proceeds to step S507.

  In step S507, the command execution control unit 261 does not lose the redundancy necessary for reading data from the storage medium 22 and the redundant data storage medium 23 through the storage medium power control unit 25, and a single or a plurality of arbitrary storage media To stop. At this time, the selection of the storage medium to be stopped from the storage medium 22 and the redundant data storage medium 23 is not necessarily the same as the storage medium stopped before the execution of step S501.

  The processing from step S501 to step S507 described above is performed, and the update reflection processing of the write data in the nonvolatile cache 24 to the storage medium 22 or the redundant data storage medium 23 by the data update control unit 263 is completed.

  In addition, although the storage medium 22 and the redundant data storage medium 23 are each shown as one unit, these have two or more storage media 22 and redundant data storage media 23 as long as data redundancy is satisfied in the entire storage 2. The structure which has may be sufficient.

  With the configuration as described above, by stopping the storage medium for storing redundant data, the power consumption of the entire storage can be reduced. Further, unlike the power consumption reduction by MAID, the storage medium having redundant data such as parity and mirror is stopped in the array composed of a plurality of storage media. Therefore, it has a certain power consumption reduction effect regardless of the data arrangement on the storage. Further, it is not necessary to access data stored in the stopped storage medium. Unlike MAID, when accessing data recorded in the stopped storage medium, the storage medium can be accessed from power-on. There is no performance penalty due to waiting time.

  In addition, the access performance is improved by adding a non-volatile cache. In Read access, when there is data to be read on the nonvolatile cache, the performance is improved by reading the data from the nonvolatile cache instead of the storage medium. On the other hand, in write access, an update to a storage medium storing redundant data such as parity and mirror is recorded in a nonvolatile cache. Therefore, update data stored in the non-volatile cache can be processed asynchronously with access processing to the storage. As a result, the performance is improved as compared with the case of directly writing to the storage medium.

[Second Embodiment]
Subsequently, assuming a HDD as a medium for the storage medium 22 and the redundant data storage medium 23, more specific operation will be described. The storage medium 22 and the redundant data storage medium 23 of the storage 2 are each configured as an array of RAID 1 composed of one HDD. As an example, the storage medium 22 and the redundant data storage medium 23 are HDDs in which addresses are assigned in order from 0 in block units.

  When the HDD corresponding to the storage medium 22 is defined as HDD-A and the HDD corresponding to the redundant data storage medium 23 is defined as HDD-B, an array of RAID 1 is configured by the HDD-A and HDD-B. Thus, the address space provided by the storage 2 matches the address space of the HDD-A. This is because RAID1 writes data having the same contents to all HDDs that construct RAID1. That is, when the host accesses the block from address X to address Y, the access destination matches the block from address X to address Y of HDD-A and HDD-B.

  HDD-A and HDD-B are in a mirror relationship with each other, and the storage 2 is operated with the HDD-B stopped. Therefore, even when the HDD-B is stopped, all data can be read from the HDD-A.

  The nonvolatile cache 24 is assumed to store no data in the initial state. Further, it is assumed that the nonvolatile cache 24 can simultaneously store data for 300 blocks of HDD-A and HDD-B at the maximum.

  Here, four areas from areas A to D shown in FIG. 14 are defined, and operations when reading and writing are performed on these areas will be described. It is assumed that access requests from the host 1 to the storage 2 are made in the order shown in FIG.

  First, a write operation for the area A will be described. When the access request processing unit 21 of the storage 2 receives a write request for the area A from the host 1, it generates a Write command for the HDD-A and HDD-B. The writing process to the area A is executed for HDD-A and HDD-B in the procedure from step S301 to step S307 in FIG.

  When attention is paid to the processing for the HDD-A, since the HDD-A is already operating, the update data for the area A is directly written to the HDD-A. On the other hand, since HDD-B, which is a mirror for HDD-A, is stopped, the same update data as HDD-A is recorded in nonvolatile cache 24. Since there is no data stored in the non-volatile cache 24 in the initial state, there is an empty area of 300 blocks. Therefore, since update data for 100 blocks for the area A can be stored, the update data for the area A of the HDD-B is stored in the nonvolatile cache 24.

  Next, an operation when data is read from the area A will be described. In this case, the Read command for the area A of the HDD-A or HDD-B is generated as in the case of the write request for the area A. Since the area A has been updated by the above-described write processing, only the update data for the area A on the HDD-A and the area A of the HDD-B recorded on the nonvolatile cache 24 is valid. In the Read command processing of FIG. 10, when data exists on the nonvolatile cache 24, the data on the nonvolatile cache 24 is preferentially read. Therefore, the update data for the area A on the nonvolatile cache 24 is read and returned as a result to the host 1.

  Next, when the reading process is performed on the area B, the data of the area B is not stored on the nonvolatile cache 24. Therefore, it is necessary to read data from area B on HDD-A or HDD-B. However, since HDD-B is in a stopped state, the data in area B is read from HDD-A. As described above, even when the data for which the read request is received is not recorded in the nonvolatile cache 24 and the medium storing the redundant data is stopped, the redundant data is read from the operating storage medium. If possible, data corresponding to redundant data is read from the operating storage medium, and necessary data is restored.

  Next, the procedure for performing the update data reflection process at time t1 at the stage when the process of reading from the region B is completed will be described. The update data reflection process is performed according to the procedure shown in FIG. In the nonvolatile cache 24, only the update data of the area A for the stopped HDD-B is recorded. An update process request for the area A of the HDD-B stopped by the data update control unit 263 is made to the command execution control unit 261. The command execution control unit 261 causes the power supply of the stopped HDD-B to be in an operating state through the storage medium power supply control unit 25. After the HDD-B is switched to the operating state, the command execution control unit 261 writes the update data on the nonvolatile cache 24 to the area A on the HDD-B. When the writing is completed, the command execution control unit 261 resets the dirty flag of the update data in the area A on the non-volatile cache 24 to 0, and no other update data exists on the non-volatile cache 24. Then, the storage medium power control unit 25 returns to the stopped state.

  The write request for the area B and the area C is performed in the same procedure as that for the area A described above. At this time, the update data for the areas B and C is 100 blocks each.

  Since the capacity of data stored in the nonvolatile cache 24 at time t1 is only the data for 100 blocks in the area A, storing the update data of the areas B and C for the stopped HDD-B stores the nonvolatile cache 24. The total capacity of data stored as a whole is 100 + 100 + 100 = 300 blocks. That is, there is no free capacity in the nonvolatile cache 24 when the access request to the area C is completed. In this state, a processing procedure for a write request for the region D will be described.

  Since there is no free capacity of the non-volatile cache 24 at the time of the writing process to the area C, it is necessary to create a free area in the non-volatile cache 24 in order to store update data for the area D of the HDD-B in the non-volatile cache 24. is there. At this time, as shown in step S306 in FIG. 11, the writing process to the area D is put on standby until a necessary free area is secured in the nonvolatile cache 24. The process of securing a free area necessary for securing the update data of the area D on the nonvolatile cache 24 is performed according to the processing procedure of FIG.

  In the free space securing process, as shown in step S401 in FIG. 12, the data update control unit 263 refers to the nonvolatile cache management table and the dirty flag on the nonvolatile cache 24 is 0, that is, updates to HDD-A and HDD-B. Select data that does not need to be processed. Data corresponding to the above-described conditions is only data in the area A in which the update to the HDD-B is reflected at the time t1 in FIG. Since the data in the area A stored in the nonvolatile cache 24 is the same as the data stored in the HDD-A and HDD-B, it can be deleted from the nonvolatile cache 24 and from the nonvolatile cache management table.

  When the data in area A is erased, the free space increases to 100 blocks. However, in order to store the update data of the area D, an empty area of 200 blocks is required as shown in FIG. Therefore, as shown in and after step S404 in FIG. 12, the dirty flag on the non-volatile cache 24 is 1, that is, the data that needs to be updated for the HDD-A and HDD-B is updated and then deleted. Thus, an empty area of the nonvolatile cache 24 is secured. Update data for the areas B and C of the HDD-B is stored as data that needs to be updated stored in the nonvolatile cache 24.

  Next, an operation in a case where the update is reflected on the data in the area B will be described. An update process request for the area B of the HDD-B stopped by the data update control unit 263 is made to the command execution control unit 261. The command execution control unit 261 causes the power supply of the stopped HDD-B to be in an operating state through the storage medium power supply control unit 25. Thereafter, the command execution control unit 261 writes the update data on the nonvolatile cache 24 to the area B on the HDD-B. When the writing is completed, the command execution control unit 261 erases the data in the nonvolatile cache 24 and the area B of the nonvolatile cache management table. By erasing the data in the area B, the free area in the nonvolatile cache 24 is increased to 100 + 100 = 200 blocks to satisfy the capacity necessary for storing the update data in the area D. Therefore, without erasing the data on the other nonvolatile cache 24, the power supply of the HDD-B is stopped through the storage medium power control unit 25, and the process of securing the necessary free space in the nonvolatile cache 24 is finished.

  The writing process for the area D after the necessary free area is secured on the nonvolatile cache 24 is the same procedure as the writing process for the areas A, B, and C. As described above, each access process shown in FIG. 15 is completed.

  During this time, the HDD-A of the storage 2 is in the operating state, while the HDD-B is always in the stopped state except during the process of securing the free capacity of the nonvolatile cache by reflecting the update data of the nonvolatile cache. It is in operation.

  Note that the use of an SSD as the nonvolatile cache 24 can be assumed. This SSD consumes less power than the HDD used in the storage medium 22 and the redundant data storage medium 23. Therefore, the power consumption of the storage 2 can be reduced by operating the HDD-B in a stopped state in principle.

  Further, in the read processing for the area A, the access performance is improved by using the nonvolatile cache 24 having higher access performance than the HDD-A and HDD-B as the cache. Since the non-volatile cache is used only as a temporary storage destination for data to be stored in the HDD-A and HDD-B, its capacity may be smaller than that of the HDD-A and HDD-B. Considering the cost difference between the HDD and the SSD, it is assumed that the non-volatile cache has a small capacity. In addition, as described above, read performance is improved by using an SSD having high access performance as a cache in response to a read request to the storage apparatus. In addition, write performance is also improved. This is due to the fact that data that is originally written randomly to any address in the HDD is written to a non-volatile cache that is faster than the HDD and faster in random access. Although the case where only the operation of the HDD-B mirrored with respect to the HDD-A is stopped has been described, a plurality of nonvolatile caches are not required even when there are a plurality of media for recording redundant data. . This is because the information is centrally managed by the non-volatile cache management table. Further, in the present embodiment, the description has been made assuming a RAID 1 mirrored HDD, but the operation of such an HDD as long as redundancy can be maintained even for an array such as RAID 4 or 5 having parity data. Can be stopped.

  A part or all of the above embodiments can be described as in the following supplementary notes, but is not limited thereto.

  (Supplementary note 1) A plurality of storage media that are used as storage destinations of data including redundant data and can be independently brought into a stopped state, and can be accessed at a higher speed than the storage media. Corresponds to the non-volatile storage medium used as a temporary recording destination of the redundant data when a data write request is made to a stopped storage medium of the medium, and the data received at the time of the data read request A storage device comprising: a controller that inquires of the non-volatile storage medium whether or not there is redundant data to be read and, if recorded, reads the redundant data from the non-volatile storage medium.

  (Additional remark 2) Furthermore, the said control part is when the redundant data corresponding to the data which received the said read request are not recorded on the said non-volatile storage medium, and the recording medium which stored the redundant data has stopped Is a storage device that reads data corresponding to redundant data from an operating storage medium among the plurality of storage media.

  (Supplementary Note 3) Among the plurality of storage media, a storage medium used as a storage destination for redundant data has been stopped since the storage device was started, and the control unit uses the storage medium as a storage destination for the redundant data. A storage apparatus for copying the redundant data recorded on the nonvolatile storage medium to a storage medium used as a storage destination of the redundant data when the storage medium is activated.

  (Supplementary Note 4) When the control unit runs out of the recording area of the nonvolatile storage medium, the control unit starts a storage medium used as a storage destination of the redundant data, and stores the copied data in the nonvolatile storage A storage device that is erased from media.

  (Additional remark 5) The said control part manages whether the data of the said non-volatile storage medium need to be copied to the storage medium used as the storage place of the said redundant data with a table, and the recording of the said non-volatile storage medium A storage device for erasing data that does not need to be copied from the non-volatile storage medium when the area is insufficient.

  (Additional remark 6) The said table is the storage apparatus currently recorded on the said non-volatile storage medium.

  (Supplementary note 7) A storage apparatus in which the control unit stops a storage medium used as a storage destination of the redundant data after the redundant data is copied.

  (Supplementary Note 8) A storage apparatus provided with a memory that can hold data by a power source that is always connected to the storage apparatus even when the storage apparatus is powered off, instead of the nonvolatile storage medium.

  (Supplementary Note 9) A plurality of storage media that are used as storage destinations of data including redundant data and can be independently brought into a stopped state, and a nonvolatile storage medium that can be accessed at higher speed than the storage medium, A storage device control method comprising: a step of recording the redundant data in the nonvolatile storage medium in response to a data write request to a storage medium that is stopped among the plurality of storage media; A step of inquiring of the nonvolatile storage medium whether or not there is redundant data corresponding to the data for which the read request has been received, and a step of reading the redundant data from the nonvolatile storage medium if the redundant data is recorded And a storage device control method.

  (Additional remark 10) Furthermore, when the redundant data corresponding to the data that has received the read request is not recorded in the nonvolatile storage medium, and the recording medium storing the redundant data is stopped, the plurality of A control method for a storage apparatus, comprising a step of reading data corresponding to redundant data from an operating storage medium among storage media.

  (Supplementary Note 11) Starting the storage medium used as the storage destination of the redundant data that has been stopped since the storage device was started, and after starting the storage medium used as the storage destination of the redundant data And a step of copying the redundant data recorded on the nonvolatile storage medium to a storage medium used as a storage destination of the redundant data.

  (Additional remark 12) Furthermore, when the recording area of the non-volatile storage medium runs short, starting the storage medium used as the storage destination of the redundant data; and the data after copying is stored in the non-volatile storage medium And erasing from the storage device.

  (Supplementary Note 13) In the step of erasing the copied data from the nonvolatile storage medium, it is determined whether or not the data in the nonvolatile storage medium needs to be copied to a storage medium used as a storage destination of the redundant data. A control method for a storage device, which, based on a table to be managed, erases data that does not need to be copied from the nonvolatile storage medium when the recording area of the nonvolatile storage medium is insufficient.

  (Supplementary note 14) A storage apparatus control method including a step of stopping a storage medium used as a storage destination of the redundant data after the redundant data is copied.

  (Supplementary Note 15) A plurality of storage media that are used as storage destinations of data including redundant data and can be independently brought into a stopped state, and a nonvolatile storage medium that can be accessed at higher speed than the storage medium, A program that is executed by a computer that controls a storage apparatus including: the redundant data is recorded in the nonvolatile storage medium in response to a data write request to a stopped storage medium among the plurality of storage media Processing, processing to query the nonvolatile storage medium for the presence or absence of redundant data corresponding to the data for which the read request has been received, and if the redundant data is recorded, from the nonvolatile storage medium A program for causing the computer to execute a process of reading redundant data.

  (Supplementary Note 16) Further, when the redundant data corresponding to the data that has received the read request is not recorded in the nonvolatile storage medium, and the recording medium storing the redundant data is stopped, the plurality of A program that executes a process of reading data corresponding to redundant data from an operating storage medium.

  (Supplementary Note 17) After starting the storage medium used as the storage destination of the redundant data, the processing for starting the storage medium used as the storage destination of the redundant data that has been stopped since the startup of the storage device And a process of copying the redundant data recorded on the nonvolatile storage medium to a storage medium used as a storage destination of the redundant data.

  (Additional remark 18) Furthermore, when the recording area of the non-volatile storage medium is insufficient, a process of starting a storage medium used as a storage destination of the redundant data, and the data after copying are stored in the non-volatile storage medium A program for executing the process of erasing from the program.

  (Supplementary Note 19) In the process of erasing the copied data from the nonvolatile storage medium, it is determined whether or not the data in the nonvolatile storage medium needs to be copied to a storage medium used as a storage destination of the redundant data. A program for erasing data that does not need to be copied from the non-volatile storage medium when the recording area of the non-volatile storage medium is insufficient based on a table to be managed.

  (Supplementary note 20) A program including a process of stopping a storage medium used as a storage destination of the redundant data after the redundant data is copied.

  It should be noted that the disclosures of the above patent documents and the like are incorporated herein by reference. Within the scope of the entire disclosure (including claims) of the present invention, the embodiment can be changed and adjusted based on the basic technical concept. Various combinations and selections of various disclosed elements are possible within the scope of the claims of the present invention. That is, the present invention of course includes various variations and modifications that could be made by those skilled in the art according to the entire disclosure including the claims and the technical idea. For example, instead of a non-volatile cache, power may be supplied only to a specific medium from a power source that is always connected to the storage device, and the contents may be retained.

1, 1a Host 2 Storage 3 Network 21 Access request processing unit 22 Storage medium 23 Redundant data storage medium 24 Non-volatile cache 25 Storage medium power supply control unit 26 Command execution unit 211 Access interpretation unit 212 Access request information storage unit 213 Data restoration unit 214 Redundancy Data generation unit 261 Command execution control unit 262 Command management information storage unit 263 Data update control unit

Claims (10)

A plurality of storage media that are used as storage destinations for data including redundant data and can be brought into a stopped state independently;
A non-volatile storage medium that can be accessed at a higher speed than the storage medium and that is used as a temporary recording destination of the redundant data when a data write request is made to a storage medium that is stopped among the plurality of storage media When,
At the time of data read request, the nonvolatile storage medium is inquired about the presence or absence of redundant data corresponding to the data that has received the read request, and if it is recorded, a controller that reads the redundant data from the nonvolatile storage medium;
A storage apparatus comprising:   Further, the control unit is configured such that when the redundant data corresponding to the data that has received the read request is not recorded in the nonvolatile storage medium, and the recording medium storing the redundant data is stopped, The storage device according to claim 1, wherein data corresponding to redundant data is read from the storage medium in operation among the storage media. Among the plurality of storage media, a storage medium used as a storage destination of redundant data has been stopped since the storage device was started,
A storage medium in which the redundant data recorded in the nonvolatile storage medium is used as the storage destination of the redundant data when the control unit starts up the storage medium used as the storage destination of the redundant data The storage device according to claim 1 or 2, wherein the storage device is copied to the storage device.   The control unit activates a storage medium used as a storage destination of the redundant data when the recording area of the nonvolatile storage medium is insufficient, and erases the copied data from the nonvolatile storage medium The storage apparatus according to claim 3.   The control unit manages whether or not data in the nonvolatile storage medium needs to be copied to a storage medium used as a storage destination of the redundant data, and a recording area of the nonvolatile storage medium is insufficient. 5. The storage apparatus according to claim 4, wherein data that does not need to be copied is erased from the nonvolatile storage medium.   The storage apparatus according to claim 5, wherein the table is recorded in the nonvolatile storage medium.   The storage apparatus according to any one of claims 3 to 6, wherein the control unit stops a storage medium used as a storage destination of the redundant data after copying the redundant data.   The storage apparatus according to any one of claims 1 to 7, further comprising a memory capable of holding data by a power source that is always connected to the storage apparatus even when the storage apparatus is powered off, instead of the nonvolatile storage medium. . A plurality of storage media that are used as storage destinations for data including redundant data and can be brought into a stopped state independently;
A non-volatile storage medium accessible at a higher speed than the storage medium;
A storage apparatus control method comprising:
In response to a data write request to a stopped storage medium among the plurality of storage media, recording the redundant data in the nonvolatile storage medium;
A step of inquiring the nonvolatile storage medium for the presence or absence of redundant data corresponding to the data for which the read request has been received at the time of a data read request;
If the redundant data is recorded, reading the redundant data from the nonvolatile storage medium;
A method for controlling a storage apparatus. A plurality of storage media that are used as storage destinations for data including redundant data and can be brought into a stopped state independently;
A non-volatile storage medium accessible at a higher speed than the storage medium;
A program for causing a computer to control a storage device comprising:
In response to a data write request to a stopped storage medium among the plurality of storage media, a process of recording the redundant data in the nonvolatile storage medium;
A process for inquiring of the nonvolatile storage medium whether or not there is redundant data corresponding to the data for which the read request has been received at the time of a data read request;
If the redundant data is recorded, a process of reading the redundant data from the nonvolatile storage medium;
A program for causing the computer to execute.

Images