WO2018193608A1

WO2018193608A1 - Storage system, control method for storage device, and storage control device

Info

Publication number: WO2018193608A1
Application number: PCT/JP2017/015988
Authority: WO
Inventors: 征之兒玉; 昌弘鶴谷
Original assignee: 株式会社日立製作所
Priority date: 2017-04-21
Filing date: 2017-04-21
Publication date: 2018-10-25

Abstract

Provided is a storage device having a storage device including a plurality of nonvolatile semiconductor drives and a control device which includes a processor and a memory and controls the storage device. The processor measures a drive write load of each of the nonvolatile semiconductor drives, selects, in the case where a nonvolatile semiconductor drive, the write load of which exceeds a predetermined threshold value is detected, the nonvolatile semiconductor drive as a drive to be subjected to relaxation of restriction on power consumption, and selects, from among the nonvolatile semiconductor drives except the drive to be subjected to the relaxation of the restriction on the power consumption, a nonvolatile semiconductor drive capable of accepting a reduction in power consumption as a drive to be subjected to strengthening of restriction on power consumption. After the power consumption of the nonvolatile semiconductor drive to be subjected to the strengthening of the restriction on the power consumption is reduced, the processor increases the power consumption of the nonvolatile semiconductor drive to be subjected to the relaxation of the restriction on the power consumption.

Description

Storage system, storage device control method, and storage control device

The present invention relates to a flash drive, a storage system including the flash drive, and a control method thereof.

Flash drives and SSDs (Solid State Drives) using NAND Flash Memory are becoming cheaper and larger in capacity due to higher integration of NAND Flash Memory. As a result, even in a storage system that conventionally uses an HDD (Hard Disk Drive), a product in which an SSD is installed instead of the HDD has appeared.

In particular, storage products in which all drives are configured with SSDs are called AFA (All Flash Array), but these products are also equipped with a large number of SSDs. Since these AFAs require high performance for SSDs, SSDs simultaneously access a plurality of NAND Flash Memory chips in order to improve performance.

Unlike the conventional HDD, NAND Flash Memory has a feature that it consumes a large amount of power when performing Write. However, as the number of simultaneous accesses increases as described above, power consumption and heat generation associated with the power increase.

In addition, the SSD has a feature of performing autonomous data movement (for example, garbage collection) within a NAND Flash Memory chip or between chips by a drive controller as an internal process. For this reason, the amount of writing to the NAND Flash Memory chip inside the SSD, that is, the power consumption cannot be controlled only by limiting the write amount by the storage controller of the AFA.

However, AFA has power restrictions due to the installation environment of power supply equipment and air conditioning equipment, power supply units that can be mounted, and cooling performance that discharges heat generated by power consumption. Therefore, by providing a power limiting system as in Patent Document 1 for an SSD mounted on an AFA in the past and performing a uniform power limitation on all SSDs in accordance with the amount of power of the mounted device, The power restrictions of AFA are observed.

JP 2016-212580 A

Flash drives such as SSD can improve performance by increasing simultaneous access to NAND Flash Memory chip during Write. However, since the power consumption increases by increasing the number of simultaneous accesses, it is necessary to determine the upper limit of SSD power according to the storage system to be mounted and to limit the write performance according to the determined upper limit.

In the above-described conventional AFA, power is uniformly limited for the mounted SSD, but in normal operation, writing is rarely performed with the maximum performance for all SSDs.

However, since the power constraint on the SSD and the write processing of the storage controller are not linked, even if there is a power margin as an AFA, the write performance is uniformly affected by the power upper limit setting of the SSD itself. For this reason, the AFA has a problem in that the performance of the entire AFA is limited to the SSD in which the write is concentrated even though there is a power margin.

The present invention is a storage device including a storage device including a plurality of nonvolatile semiconductor drives, and a control device that includes a processor and a memory to control the storage device, and the processor includes the nonvolatile semiconductor drive And when the processor detects a non-volatile semiconductor drive whose write load exceeds a predetermined threshold, the processor selects the non-volatile semiconductor drive as a power consumption restriction relaxation target, and the processor Selects a non-volatile semiconductor drive that can allow a reduction in power consumption as a non-restricted target for power consumption, and the processor limits the power consumption. After reducing the power consumption of the nonvolatile semiconductor drive to be strengthened, Cost increases the power consumption of the power of the restriction relaxation target nonvolatile semiconductor drive.

It is possible to improve the write performance of the entire storage system by optimizing the power of the nonvolatile semiconductor drive according to the upper limit of the write amount of the storage system and the power that can be supplied.

It is a block diagram which shows the Example of this invention and shows an example of a storage system. It is a block diagram which shows the Example of this invention and shows an example of the program and data which are stored in the memory of a storage controller. It is a figure which shows the Example of this invention and shows an example of an enclosure electric power upper limit table. It is a figure which shows the Example of this invention and shows an example of an enclosure cooling capacity table. It is a figure which shows the Example of this invention and shows an example of an enclosure cooling state table. It is a figure which shows the Example of this invention and shows an example of the mounted drive information table. It is a figure which shows the Example of this invention and shows an example of a parity group structure information table. It is a figure which shows the Example of this invention and shows an example of a drive operation mode state table. It is a block diagram which shows the Example of this invention and shows an example of a structure of a flash drive. It is a figure which shows the Example of this invention and shows an example of the operation mode table stored in the memory of a flash drive. FIG. 3 is a diagram illustrating a physical structure of a storage system and a configuration of a parity group according to an embodiment of this invention. 5 is a flowchart illustrating an example of processing performed by the storage controller according to the embodiment of this invention. It is a figure which shows the Example of this invention and shows an example of an enclosure cooling state table. It is a figure which shows the Example of this invention and shows an example of a drive operation mode state table. It is a figure which shows an Example of this invention and shows an example of the enclosure cooling state table after changing the operation mode. It is a figure which shows an Example of this invention and shows an example of the drive operation mode state table after changing an operation mode. FIG. 3 is a diagram showing an embodiment of the present invention and showing a physical structure of a storage system and a configuration of a parity group when a flash drive failure occurs. It is a figure which shows the Example of this invention and shows an example of the drive operation mode state table at the time of failure of a flash drive. It is a figure which shows the Example of this invention and shows an example of the drive operation mode state table during RAID structure reconstruction.

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an example of the storage system of the present invention. FIG. 1 shows a configuration of a storage system in which a host 100 and a storage system 101 are connected.

The host 100 executes the application and issues commands such as Read and Write to the storage system 101. Although not shown, the host 100 is a computer including an interface for connecting to the storage system 101, a CPU (Central Processing Unit), a memory, and the like.

The storage system 101 includes a storage controller 102 and a drive enclosure 103. The storage controller 102 has one or more controller units 104 mounted therein.

In the example shown in the figure, it is composed of two controller units 104-1 and 104-2. In the following description, when the controller unit is not specified, the symbol “104” in which “−” and the subsequent parts are omitted is used. The same applies to the reference numerals of other components.

The controller unit 104 includes a front-end interface (FE I / F in the figure) 110, a CPU 111, a memory 112, a back-end interface (BE I / F in the figure) 113, and a control I / F 114. As the front end interface 110, a protocol such as FC (Fibre Channel) or IP can be adopted. As the back-end interface 113, SAS (Serial Attached SCSI) or NVMe (Non-Volatile Memory Host Host Controller Interface) can be employed.

The drive enclosure 103 includes enclosure interfaces (ENC I / F in the figure) 120-1 and 120-2, power supply and fan units 121-1 and 121-2, and a plurality of flash drives (nonvolatile semiconductor drives) 122. Is done.

In this embodiment, for redundancy, controller units 104-1 and 104-2, enclosure interface (ENC I / F) 120-1 and 120-2, power supply and fan units 121-1 and 121-2 are each set to two. However, other configurations are possible.

Also, the number of mounted flash drives 122 can be changed. A plurality of drive enclosures 103 can be connected to one storage controller 102.

The front-end interface (FE I / F) 110 of the storage controller 102 is connected to the host 100 via a network (not shown) and mediates communication with the host 100.

The CPU 111 executes various processes using each program, management information, etc. stored in the memory 112. In addition, the CPUs 111 mounted on each controller unit 104 can communicate with each other to ensure data consistency and perform processing and data synchronization.

The memory 112 stores a program executed by the CPU 111, management information used by the program, and the like. The memory 112 may be used for other purposes, such as storing a cache of data accessed by the host 100.

The memory 112 is known to be composed of SDRAM (Synchronous Dynamic Random Access Memory), but may be composed of memory elements other than those described above.

The back-end interface (BE I / F) 113 of each controller unit 104 is connected to the flash drive 122 via the enclosure interface (ENC I / F) 120-1 and 120-2, and the CPU 111 accesses the flash drive 122. To mediate.

The control I / F 114 of each controller unit 104 is connected to the power supply and fan unit 121 mounted on the drive enclosure 103 via enclosure interfaces 120-1 and 120-2, and mediates control of fan cooling performance.

The enclosure interface 120 mediates communication between the controller unit 104 and the power supply / fan unit 121 and the flash drive 122. In addition, when the communication performed by the control I / F 114 can be mediated by the back-end interface 113, the installation and communication of the control I / F 114 are not necessary. Further, the enclosure interface 120 is unnecessary in the configuration in which the back-end interface 113 of the controller unit 104 and the flash drive 122 can be directly connected. For example, the case where the storage controller 102 and the drive enclosure 103 are mounted as a single enclosure corresponds to the above case.

The power supply and fan unit 121 supplies power to the entire drive enclosure 103 and performs cooling to prevent temperature rise of the flash drive 122 by blowing air with an air cooling fan (not shown). The air cooling fan included in the power supply and fan unit 121 functions as a cooling device.

The rotation speed of the air cooling fan can be controlled from the controller unit 104 via the control I / F 114 and the enclosure interface 120. It is assumed that a predetermined cooling performance equal to or less than the maximum cooling performance can be obtained by changing the rotation speed of the air cooling fan.

Note that it is possible to implement stepwise control of the rotation speed of the air cooling fan. Note that the power supply and the fan unit 121 are not necessarily integrated, and may be separately mounted. In this embodiment, an example in which cooling is performed by an air cooling fan has been described. However, any cooling device capable of controlling the cooling performance may be used. For example, a water cooling type cooling device may be employed.

The flash drive 122 is a device in which a NAND flash memory is mounted as a nonvolatile storage area. The flash drive 122 may use a highly reliable technology such as RAID (Redundant Arrays of Independent Disks) by grouping a plurality of flash drives into units called parity groups.

FIG. 2 is a block diagram showing programs and information stored in the memory 112 of the controller unit 104.

The memory 112 includes a storage control program 150 for controlling the flash drive 122 of the drive enclosure 103 as a storage device, a power performance optimization program 200, a write load monitoring program 201, an optimum power allocation calculation program 202, and drive operation. A mode change notification program 203 and a cooling performance control program 204 are stored.

In addition, the memory 112 includes, as data used by the control program, an enclosure power upper limit table 210, an enclosure cooling capacity table 211, an enclosure cooling state table 212, an installed drive information table 220, and a parity group configuration information table 221. The IO pattern learning table 222 and the drive operation mode state table 223 are also held. Note that the memory 112 may also hold control programs other than those described above, control data, and the like.

The storage control program 150 receives a read request or a write request from the host 100 and controls reading / writing of the flash drive 122 in the drive enclosure 103.

The power performance optimization program 200 links the write load monitoring program 201, the optimal power allocation calculation program 202, the drive operation mode change notification program 203, and the cooling performance control program 204 to optimize the power and performance of the storage system 101. Turn into.

The write load monitoring program 201 monitors the write processing load on each flash drive 122 and monitors the occurrence of bottlenecks. As the write processing load, for example, IOPS (Input / Output Operations Per Second) of the Write command can be used. Alternatively, the writing speed (MB / S) for each flash drive 122 may be used.

The write load monitoring program 201 measures IOPS (load information) for each flash drive 122, and determines that the load of the flash drive 122 in which the IOPS exceeds a preset first threshold is “high load”. Further, the write load monitoring program 201 treats the flash drive 122 determined to be “high load” as a bottleneck.

The optimal power allocation calculation program 202 operates when a bottleneck is detected by the write load monitoring program 201. This program is based on the write load information of the flash drive 122, the number of flash drives 122 mounted in the drive enclosure 103, the operation mode set for each flash drive 122, and the upper limit of power supplied by the drive enclosure 103. In addition, it is determined whether or not the power distribution can be increased for the flash drive 122 that is a bottleneck.

When the optimum power allocation calculation program 202 determines that it is possible to increase the distribution of power consumption to the bottleneck flash drive 122, as described later, the power consumption of the bottleneck flash drive 122 is increased. The operation mode of the flash drive 122 is determined. The operation mode of the flash drive 122 selects a mode ID that relaxes the restriction on power consumption.

On the other hand, the operation mode of the flash drive 122 to be reduced is determined in order to reduce the allocation of power consumption to the flash drive 122 that can reduce power consumption among the other flash drives 122. As the operation mode of the flash drive 122 to be reduced, a mode ID that strengthens the restriction on power consumption is selected.

The optimum power allocation calculation program 202 according to the present embodiment is a “medium load” drive in which Write processing IOPS is equal to or lower than the first threshold and equal to or higher than the second threshold as the flash drive 122 that reduces allocation of power consumption. Are selected for reduction. However, the second threshold value is smaller than the first threshold value.

Note that the present embodiment is not limited to this, and the flash drive 122 that is the target of power consumption reduction may be a drive that does not have a problem in write performance even if power consumption is reduced. For example, the optimal power allocation calculation program 202 changes a drive whose write processing IOPS is equal to or lower than the first threshold among the flash drives 122 whose write load state 2234 is “high load” to “medium load”, The operation mode 2233 can be changed to “2” to reduce the allocated power consumption.

The drive operation mode change notification program 203 is actually used when the optimum power allocation calculation program 202 determines to change the allocation of power consumption for each flash drive 122 or when returning to the state before the change. To change the operation mode.

The cooling performance control program 204 determines whether or not the cooling performance has been changed when it is determined by the optimum power allocation calculation program 202 to change the power consumption allocation for each flash drive 122. In other words, the cooling performance control program 204 controls the power supply and the fan unit 121 to change the rotation speed of the fan when the fan rotation speed needs to be changed or when returning to the state before the change.

The CPU 111 operates as a functional unit that provides a predetermined function by processing according to each program. For example, the CPU 111 functions as a power performance optimization unit by performing processing according to the power performance optimization program 200. The same applies to other programs. Furthermore, the CPU 111 also operates as a function unit that provides each function of a plurality of processes executed by each program. A computer and a computer system are an apparatus and a system including these functional units.

Information such as programs and tables for realizing each function of the controller unit 104 is stored in a storage subsystem, a nonvolatile semiconductor memory, a hard disk drive, a nonvolatile storage device such as an SSD, or a computer reading such as an IC card, an SD card, a DVD, etc. It can be stored on possible non-transitory data storage media.

FIGS. 3A to 3F show that the power performance optimization program 200, the write load monitoring program 201, the optimum power allocation calculation program 202, the drive operation mode change notification program 203, and the cooling performance control program 204 are operated. It is a figure which shows an example of the information used.

FIG. 3A is a diagram illustrating an example of the enclosure power upper limit table 210. The enclosure power upper limit table 210 includes an enclosure ID 2101 for storing identification information of each drive enclosure 103, a chassis power upper limit 2102 set for each drive enclosure 103, and an upper limit value of power supplied to a single flash drive installed in the drive enclosure 103. Is included in one entry.

Although FIG. 1 shows an example in which the drive enclosure 103 is one, a plurality of drive enclosures 103 can be connected to the storage controller 102. For this reason, the drive enclosure 103 is specified by the enclosure ID 2101.

Further, the housing power upper limit 2102 has been degraded due to a failure of the power source and fan unit 121-1 or 121-2 even when a plurality of power sources and fan units 121-1 and 121-2 are constantly supplied with power. In this case, the upper limit value (maximum value) that can be supplied is shown. As in this embodiment, the maximum supply power (maximum as a rating, not a momentary one) when a redundant power supply (two units) is adopted needs to be a power that can be supplied even during degeneration. . For example, in the case of a two-unit redundant configuration, since it is essential to operate even when one unit fails, the power that can be supplied by one surviving unit is the maximum supply power. In addition, in the case of 3 units redundant configuration, if up to 1 unit is tolerated, the maximum supply power is the power that can be supplied by 2 units when one unit fails. In this case, the sum of 2 units is the maximum supply power. It becomes.

The information in the enclosure power upper limit table 210 is set in advance by a management computer or the like (not shown).

FIG. 3B is a diagram illustrating an example of the enclosure cooling capacity table 211. The enclosure cooling capacity table 211 includes an enclosure ID 2111 for storing identification information of each drive enclosure 103, a maximum cooling performance 2112 for storing the maximum value of the cooling capacity set for each drive enclosure 103, and the cooling capacity of the drive enclosure 103. The standard cooling performance 2113 for storing the rated value and the cooling function degeneration cooling performance 2114 for storing the cooling capability at the time of degeneration of the power supply and fan unit 121 when the power supply and fan unit 121 are made redundant are combined into one. Include in entry.

The maximum cooling performance 2112 indicates the cooling performance (W) when the cooling device including the redundant system is operated to the maximum out of the cooling performance of each drive enclosure 103. The standard cooling performance 2113 indicates the cooling performance of the power supply and fan unit 121 in the standard load state. The cooling performance 2114 when the cooling function is reduced indicates the maximum cooling performance when the redundant system is degenerated.

Note that information in the enclosure cooling capacity table 211 is set in advance by a management computer or the like (not shown).

FIG. 3C is a diagram illustrating an example of the enclosure cooling state table 212. The enclosure cooling state table 212 includes an enclosure ID 2121 that stores identification information of each drive enclosure 103, a cooling performance 2122 that indicates the cooling performance of each drive enclosure 103 that is currently connected to the controller unit 104, and the redundant system is degenerated. One entry includes a cooling state degeneration presence / absence 2123 indicating whether or not there is.

The enclosure cooling state table 212 is periodically updated by the controller unit 104 and updated to the latest state.

FIG. 3D is a diagram illustrating an example of the mounted drive information table 220. The mounted drive information table 220 includes an enclosure ID 2201 that stores an identifier of the drive enclosure 103 in which the flash drive 122 is mounted, a drive ID 2202 that stores an identifier for each flash drive 122, and an operation mode that can be changed by the flash drive 122. Mode number 2203 for storing the number, mode 1 power 2204 for storing the power consumption (W) when the operation mode is 1, and mode 1 write performance for storing the write performance (MB / s) when the operation mode is 1. 2205, mode 2 power 2206 for storing power consumption (W) when the operation mode is 2, mode 2 write performance 2207 for storing write performance (MB / s) when the operation mode is 2, and operation not shown Power consumption and writing when mode is 3 Including the performance to one of the entry.

In FIG. 3D, only Write performance is retained, but performance information such as read performance may be retained. The power consumption and write performance fields may be set according to the number of modes 2203.

The information in the mounted drive information table 220 is set in advance by a management computer (not shown) or the like.

FIG. 3E is a diagram illustrating an example of the parity group configuration information table 221. The parity group configuration information table 221 holds the combination information of the flash drive 122 groups constituting the RAID as a combination of a drive ID and a parity group ID.

The parity group configuration information table 221 includes an enclosure ID 2211 that stores an identifier of the drive enclosure 103 in which the flash drive 122 is mounted, a drive ID 2212 that stores an identifier for each flash drive 122, and a parity group to which the flash drive 122 belongs. One entry includes a parity group ID 2213 for storing an identifier and a power allocation change enable / disable flag 2214 for setting whether or not to change the allocation of power consumption.

Also, a drive set as a spare drive so that availability can be maintained by changing the role of the flash drive 122 in which a failure has occurred can be identified as “SPARE” in the parity group ID 2213. Note that the mapping of SPARE with a specific parity group ID can be changed to a value suitable for each storage system.

The power allocation change enable / disable flag 2214 sets whether or not to allow a change in power allocation. If the power allocation change enable / disable flag 2214 is “permitted”, the power consumption change by the storage controller 102 is accepted, and if it is “no”, the power consumption change is rejected.

In this embodiment, an example in which the flash drive 122 is used in RAID 5 or RAID 6 is shown, but the present invention is not limited to this. For example, when one storage area is configured by a plurality of flash drives 122, the flash drives 122 can be managed by a group of storage areas instead of a parity group.

The IO pattern learning table 222 records or learns the data access history for each flash drive 122 or parity group 601 to 605. Specifically, history information to be recorded and information to be learned include access source host 100 information, access destination flash drive 122 information, access destination parity group information, access command information, and access frequency. Information, access data size information, access time information, access cycle information, and the like, but other information, history, and performance information may be recorded.

FIG. 3F is a diagram illustrating an example of the drive operation mode state table 223. The drive operation mode state table 223 holds the operation status of the flash drive 122 currently mounted in each drive enclosure 103.

The drive operation mode state table 223 includes an enclosure ID 2231 that stores an identifier of the drive enclosure 103, a drive ID 2232 that stores an identifier for each flash drive 122, an operation mode 2233 that stores an operation mode for each flash drive 122, and a flash drive. Each entry includes a write load state 2234 for storing a write load state for each 122.

The write load state for each flash drive 122 is high load if the above-mentioned IOPS exceeds the first threshold, medium load if the IOPS is below the first threshold, and exceeds the second threshold, and second If it is below this threshold value, it can be set as a low load. The resolution of the write load does not necessarily need to be three stages of high load, medium load, and low load, and finer resolution can be adopted.

FIG. 4 is a block diagram showing an example of the configuration of the flash drive 122. The flash drive 122 includes a disk interface (DISK I / F in the figure) 400, a drive controller 401, a memory 402, a plurality of flash memory access channels 403, and a plurality of flash memory chips 404.

The disk interface 400 is connected to the controller unit 104 via the enclosure interface 120 and the back-end interface 113. In the present embodiment, the connection between the enclosure interface 120 and the back-end interface 113 is shown as an example of two systems for redundancy.

The drive controller 401 is a type of CPU that processes commands received from the disk interface 400. The drive controller 401 executes the chip control program 420 stored in the memory 402, refers to the data, and executes various processes.

The memory 402 stores programs and data executed by the drive controller 401. In this embodiment, the memory 402 holds an operation mode change program 10, an operation mode table 411, and a chip control program 420.

The memory 402 may be used for storing other control programs and cache data. The memory 402 can be composed of SDRAM (Synchronous Dynamic Random Access Memory), but may be composed of memory elements other than those described above.

The chip control program 420 processes a command received by the disk interface 400 and controls reading and writing to the flash memory chip 404.

Also, the chip control program 420 controls autonomous data movement (for example, garbage collection and leveling of the number of writes) according to the storage status of valid data in the flash memory chip 404 and the number of rewrites.

The flash memory access channel 403 is a path for connecting the drive controller 401 and the flash memory chip 404 and accessing the drive controller 401.

The number of flash memory access channels 403 differs depending on the drive controller 401 to be used, and the larger the number, the more flash memory chips 404 can be connected. Also, by simultaneously activating a plurality of flash memory access channels 403, it becomes possible to access a plurality of flash memory chips 404 at the same time, improving performance while consuming a large amount of power during Write.

The flash memory chip 404 is a NAND flash memory chip connected to the drive controller 401 via the flash memory access channel 403, and stores actual data and the like. A plurality of flash memory chips 404 are connected to each flash memory access channel 403.

The operation mode change program 410 is a program stored in the memory 402 and executed by the drive controller 401. The operation mode change program 410 changes the operation mode of the flash drive 122 in accordance with an instruction from the controller unit 104.

The operation mode of the flash drive 122 is stored in the operation mode table 411. The operation mode table 411 is stored in the memory 402 and is referred to when the operation mode is changed by the operation mode change program 410. Information included in the operation mode table 411 is collected by the controller unit 104 and set in the mounted drive information table 220.

FIG. 5 is a diagram illustrating an example of information held in the operation mode table 411. The operation mode table 411 includes a mode ID 4111 that can be set in the flash drive 122, drive power consumption (W) 4112 in the operation mode, write performance (MB / s) 4113 in the operation mode, and flash memory at the time of writing. Write multiplicity 4114 indicating the number of simultaneous accesses to the chip 404 is held as a combination of mode IDs.

In this embodiment, the maximum value of power consumption is controlled by the write multiplicity 4114 indicating the number of simultaneous accesses to the flash memory chip 404 at the time of writing to the flash drive 122.

When the mode ID = 1, the number of simultaneous accesses to the flash memory chip 404 at the time of writing becomes 256, and the power consumption is maximized and the write performance is maximized. On the other hand, when the mode ID is 3, the number of simultaneous accesses to the flash memory chip 404 at the time of writing is 64, and the power consumption is minimum and the writing performance is minimum.

When mode ID = 2, the number of simultaneous accesses to the flash memory chip 404 at the time of writing is 128, and the power consumption and the writing performance are intermediate values.

In FIG. 5, information other than the write performance is not described, but information other than the above may be retained. Further, the flash drive 122 having a larger number of modes to be held than in FIG. 5 can be configured. Furthermore, the performance information can be configured to include a range such as a minimum value or a maximum value. Further, the above information may be collected and used by the controller unit 104.

Note that the mode ID (operation mode) 4111 of the flash drive 122 has been shown as an example of three stages, but can be changed by increasing or decreasing the write multiplicity classification to be written simultaneously. In the present embodiment, an example in which the power consumption and the write performance are controlled in the operation mode of the flash drive 122 is shown, but the present invention is not limited to this. For example, the power consumption and write performance of the flash drive 122 may be controlled by the write multiplicity of the flash drive 122.

FIG. 6 is a diagram showing the physical structure of the storage system 101 and the structure of the parity group. FIG. 6 shows an example in which the storage controller 102 and the drive enclosure 103 are mounted as different enclosures and connected by cables.

Note that the present invention is not limited to this, and the storage controller 102 and the drive enclosure 103 may be mounted as an integral casing. Further, one drive enclosure 103 is mounted as the storage system 101 and has an enclosure ID = 00 as an identification ID, but a plurality of drive enclosures 103 may be connected.

Note that when a plurality of drive enclosures 103 are connected to the storage controller 102, the enclosure controller 102 can be identified by setting the enclosure IDs so that they do not overlap. In the illustrated example, 24 flash drives 122 are mounted on the drive enclosure 103, but it is not always necessary to mount all of them.

In this embodiment, it is assumed that all the flash drives 122 have the same model and have the mode shown in FIG. 5 as the operation mode table 411. Parity groups 601 to 605 are set in each flash drive 122, respectively.

Parity group 1 (parity group ID = 1) 601 is composed of four flash drives 122. Parity group 2 (parity group ID = 2) 602 is composed of eight flash drives 122.

Parity group 3 (parity group ID = 3) 603 includes four flash drives 122. Parity group 4 (parity group ID = 4) 604 includes six flash drives 122.

The spare group (parity group ID = SPARE) 605 is composed of two flash drives 122. The configuration of these parity groups 601 to 605 is also shown as an example, and it is possible to adopt a different configuration or a configuration in which no parity group is set.

In the example shown in the drawing, the connection between the power supply and fan unit on the storage controller 102 side, the host 100, and the host 100 and the front end interface 110 is omitted.

FIG. 7 is a flowchart showing an example of processing performed by the storage controller. Hereinafter, a control method in the storage system 101 as illustrated in FIGS. 1 and 6, that is, processing of the power performance optimization program 200 will be described using flowcharts.

The power performance optimization program 200 and the like shown in this flowchart are loaded into the memory 112 and executed by the CPU 111 of the controller unit 104, but function as a part of the storage control program 150 that controls the entire storage system 101. Also good. This process is repeatedly executed.

First, in the controller unit 104, the storage control program 150 that controls the entire storage system 101 calls the power performance optimization program 200 to start processing (700).

The power performance optimization program 200 detects the state of the power supply and the fan unit 121 mounted in the drive enclosure 103, and determines whether there is an abnormality such as degeneration of the cooling device (720). The power performance optimization program 200 proceeds to step 701 if degeneration has not occurred in the power source and fan unit 121, and proceeds to step 710 if degeneration has occurred.

Note that if the redundant cooling device (air cooling fan) is degenerated, changing the allocation of the power consumption of the flash drive 122 to eliminate the bottleneck is prohibited.

In step 701, the power performance optimization program 200 starts the write load monitoring program 201. The write load monitoring program 201 detects the load (for example, IOPS) of the flash drive 122 mounted in the drive enclosure 103, and the bottleneck (load is detected) when the IOPS exceeds a preset first threshold. (Concentrated) device.

The write load monitoring program 201 detects the load of each flash drive 122 and calculates the load state in three stages (high load, medium load, and low load) from the first threshold and the second threshold as described above. The drive operation mode state table 223 is updated.

The power performance optimization program 200 acquires the detection result of the write load monitoring program 201 and determines whether or not there is a flash drive 122 that is a bottleneck in the write process.

The power performance optimization program 200 proceeds to step 703 if the write processing load is concentrated and there is a flash drive 122 that is a bottleneck, and proceeds to step 710 if there is no bottleneck.

In step 703, the power performance optimization program 200 activates the optimum power allocation calculation program 202.

The optimal power allocation calculation program 202 first reads the drive operation mode state table 223 and identifies the flash drive 122 with a high write load. The optimum power allocation calculation program 202 relaxes the limitation on the power consumption of the flash drive 122 whose write load status 2234 of the drive operation mode status table 223 is “high load” and selects it as an increase target to increase the allocation of power consumption. .

Note that the present invention is not limited to the above, and the parity group to which the flash drive 122 having a high write load belongs may be used as the flash drive 122 whose power consumption is to be increased.

The optimum power allocation calculation program 202 refers to the parity group configuration information table 221, and among the flash drives 122 in which the power allocation change enable / disable flag 2214 is set to “permitted”, the drive that can shift to the operation mode with low power consumption. As described above, the flash drive 122 whose load state 2234 is “medium load” is selected as a power consumption reduction target (power consumption restriction enhancement target).

In the above description, the optimal power allocation calculation program 202 selects the flash drive 122 whose load state 2234 is “medium load”. However, the optimal power allocation calculation program 202 refers to the IO pattern learning table 222, You may make it select as the flash drive 122 which lowers | writes the write performance of the flash drive 122 anticipated to change in the state with low write load, ie, can be changed into the operation mode of low power consumption.

Further, the optimum power allocation calculation program 202 indicates that the number of flash drives 122 that can be changed to the low power consumption operation mode is the sum of the power consumption of the flash drives 122 to be increased in power consumption and the consumption of other flash drives 122. The total sum of the powers only needs to be equal to or less than the chassis power upper limit 2102 of the enclosure power upper limit table 210.

Based on the selection result, the optimum power allocation calculation program 202 sets the flash drive 122 having a high write load to mode ID (4111) = 1 and sets the maximum write performance 2205 with the maximum power consumption. On the other hand, the optimal power allocation calculation program 202 sets the flash drive 122 that can be changed to the low power consumption operation mode to mode ID = 3, and sets the minimum write performance with the minimum power consumption.

Also, the optimum power allocation calculation program 202 calculates the total power value of the power consumption of all the flash drives 122 when the flash drives 122 other than the increase target and the reduction target are operated with mode ID = 2.

Next, in step 704, the optimum power allocation calculation program 202 calculates the total power value, the value of the enclosure power upper limit 2102 obtained from the enclosure power upper limit table 210, and the value of the maximum cooling performance 2112 of the enclosure cooling capacity table 211. Each is compared to determine whether to change the power allocation.

The optimum power allocation calculation program 202 proceeds to step 705 in order to change the power allocation when the total power value is lower than both the value of the chassis power upper limit 2102 and the value of the maximum cooling performance 2112. On the other hand, if the total power value exceeds either the value of the chassis power upper limit 2102 or the value of the maximum cooling performance 2112, the optimum power allocation calculation program 202 prohibits the power allocation change and returns to step 720. Repeat the above process.

Note that the optimum power allocation calculation program 202 does not use an operation mode in which the drive power consumption in the operation mode table 411 is larger than the single drive supply power upper limit 2103 in the enclosure power upper limit table 210.

In step 705, the power performance optimization program 200 activates the drive operation mode change notification program 203 in order to change the operation mode of each flash drive 122.

The drive operation mode change notification program 203 refers to the drive operation mode state table 223 and changes the operation mode to mode ID = 3 for the flash drive 122 to be reduced to be changed to the operation mode with low power consumption. I do. The drive operation mode change notification program 203 updates the operation mode 2233 of the drive operation mode state table 223 for the flash drive 122 whose mode ID has been changed.

Next, the power performance optimization program 200 starts the cooling performance control program 204. In step 706, the cooling performance control program 204 determines whether the cooling performance needs to be enhanced. The cooling performance control program 204 refers to the cooling performance 2122 of the enclosure cooling state table 212 and strengthens the cooling performance 2122 if the total power value calculated by the optimum power allocation calculation program 202 exceeds the cooling performance 2122. Proceed to step 707. On the other hand, if the total power value is equal to or lower than the cooling performance 2122, the cooling performance control program 204 maintains the cooling performance 2122 and proceeds to step 708.

In step 707, the cooling performance control program 204 increases the rotational speed of the cooling fan mounted on the power supply and fan unit 121. After the cooling fan rotation speed change is completed, the power performance optimization program 200 starts the drive operation mode change notification program 203 again. If the designed cooling performance exceeds all the modes of each flash drive 122, the cooling performance control program 204 and its reference information can be omitted.

The drive operation mode change notification program 203 restarted refers to the drive operation mode state table 223 and changes the operation mode of the flash drive 122 to be increased among the flash drives 122 having a high write load to mode ID = 1. Then, a mode change operation is performed (708). The drive operation mode change notification program 203 updates the operation mode 2233 of the drive operation mode state table 223 for the flash drive 122 whose mode ID has been changed.

The power (operation mode) allocated to the flash drive 122 by the series of processing in steps 701 to 708 is changed according to the write load status, and the write performance is changed by changing the operation mode of the flash drive 122 that has become a bottleneck. And the overall performance of the storage system 101 is improved.

Further, the write load monitoring program 201 monitors whether the write load is concentrated on the specific flash drive 122 (702). If the load is not concentrated, the process proceeds to step 710.

In step 710, the power performance optimization program 200 determines whether or not the power allocation of the flash drive 122 has been changed. If it has been changed, the process proceeds to step 711 to start processing for returning the operation mode.

On the other hand, if it is determined in step 710 that the power allocation has not been changed, the process returns to step 720 and the above processing is repeated. That is, when the cooling device is degenerated and the power consumption of the flash drive 122 is not changed, the power performance optimization program 200 prohibits the power consumption allocation change.

In step 711, first, the drive operation mode change notification program 203 refers to the drive operation mode state table 223 and sets the mode ID = 2 to the flash drive 122 whose operation mode is set to mode ID = 1. Make a change.

The drive operation mode change notification program 203 updates the operation mode 2233 of the drive operation mode state table 223 for the flash drive 122 whose mode ID has been changed.

Next, the power performance optimization program 200 starts the cooling performance control program 204.

In step 712, the cooling performance control program 204 refers to the enclosure cooling state table 212 and determines whether or not the cooling performance has been enhanced (changed). If the cooling performance is strengthened (changed), the process proceeds to step 713 and the cooling performance control program 204 changes the rotation speed of the cooling fan mounted on the power supply and fan unit 121 to the normal state.

The cooling performance control program 204 updates the cooling performance 2122 of the enclosure cooling state table 212 to a value corresponding to the rotational speed by reducing the rotational speed. The cooling performance corresponding to the rotation speed of the cooling fan is set in advance.

On the other hand, if the cooling performance is not enhanced (changed), the process proceeds to step 714.

Thereafter, the power performance optimization program 200 activates the drive operation mode change notification program 203 again, refers to the drive operation mode state table 223, and for the flash drive 122 whose operation mode is set to mode ID = 3. Then, the mode ID is changed to 2 (714). The drive operation mode change notification program 203 updates the operation mode 2233 of the drive operation mode state table 223 for the flash drive 122 whose mode ID has been changed.

From the state in which the power allocation of each flash drive 122 has been changed by the processing in steps 710 to 714 above, the bottleneck is resolved and the power allocation is restored, and the power allocation to each flash drive 122 is returned uniformly. It is.

Also, when an abnormality such as degeneration is detected in the power supply and fan unit 121 mounted on the drive enclosure 103 by the power performance optimization program 200, the power allocation state is optimized by performing the processing from 710 to 714. Then, control is performed so that the power supply and fan unit 121 can operate even in a degenerated state.

On the other hand, when the cooling performance is sufficiently high, it is possible to omit the control when the power source and the fan unit 121 are degenerated as shown in step 720.

Furthermore, the control in step 720 can be expanded, and for example, when the power supply and fan unit 121 is degenerated, all the flash drives 122 can be changed to a mode with less power consumption. When such expansion is performed, the maximum cooling performance of the power supply and fan unit 121 can be lowered, and the cost of the storage system 101 can be reduced by adopting a lower priced power supply and fan unit 121.

Note that the termination condition of the power performance optimization program 200 shown in this flowchart is not explicitly shown, but it is assumed that the termination is performed according to the termination condition of the program that controls the entire storage system.

An example in which power optimization is performed according to the flowchart of the present invention shown in FIG. 7 will be described with reference to FIGS. 8A, 8B, 9A, and 9B.

FIG. 8A is a diagram illustrating an example of the enclosure cooling state table 212 before the power allocation is changed. FIG. 8B is a diagram illustrating an example of the drive operation mode state table 223 before the power allocation is changed.

FIG. 9A is a diagram showing an example of the enclosure cooling state table 212 after the power allocation is changed. FIG. 8B is a diagram showing an example of the drive operation mode state table 223 after the power allocation is changed.

8A and 8B show a state in which all flash drives 122 are operating in mode ID = 2 in the storage system 101 including the storage controller 102 and the drive enclosure 103 shown in FIG. In the illustrated example, the write load is concentrated on the parity group 2 (602), resulting in a high load and a bottleneck.

In the illustrated example, the parity group 1 (601), the parity group 3 (603), and the spare group (605) are in a low load state, and the parity group 4 (604) is in a medium load state.

FIG. 9A and FIG. 9B show the result of performing the power reallocation by the power performance optimization program 200 on the storage system 101 in this state.

The parity group 2 (602) in which the write load is concentrated is changed to mode ID (operation mode 2233) = 1 to increase power consumption, and the parity group 1 (601), parity group 3 (603), and spare The group (605) is changed to mode ID = 3 to reduce power consumption, and the parity group 4 (604) remains mode ID = 2. Also, the cooling performance of the drive enclosure 103 is enhanced from 360 W to 375 W.

As described above, by applying the present invention, it is possible to optimize the power and performance against the difference in write load between the parity groups 601 to 605.

Also, the write load from the host 100 as shown in FIGS. 6, 8A, 8B, 9A, and 9B is predicted using the IO pattern learning table 222, and the power allocation is recalculated (703). In addition to the above, it can also be applied to the write processing generated by the storage controller 102 itself. A specific example is shown using FIG.

FIG. 10 is a diagram showing the physical structure of the storage system 101 and the configuration of the parity group when the flash drive 122 fails.

In the storage system 101 composed of the storage controller 102 and the drive enclosure 103 similar to those in FIG. 6, all the flash drives 122 are assumed to have the same model and have the mode shown in FIG.

Parity groups 601 to 605 are set in the flash drive 122. In the example shown in the figure, it is assumed that one unit belonging to parity group ID = 1 constructed as RAID 5 is a failed flash drive 1010.

The controller unit 104 sets a parity group including three flash drives 122 with parity group ID = 1 as parity group 1 (degenerate) 601A. Then, the controller unit 104 sets the flash drive 122 assigned as an alternative drive from the flash drive 122 with the parity group ID = SPARE instead of the failed flash drive 1010 as the parity group 1 (active spare) 601B.

For the flash drives 122 other than the above, each parity group 2 (parity group ID = 2) 602 is configured with eight flash drives 122. Parity group 3 (parity group ID = 3) 603 is composed of four flash drives 122.

Parity group 4 (parity group ID = 4) 604 is composed of six flash drives 122. The spare group (parity group ID = SPARE) 605 is composed of one flash drive 122.

FIG. 11 shows the state of the flash drive 122 in FIG. FIG. 11 is a diagram illustrating an example of a drive operation mode state table when a failure occurs in the flash drive.

Parity group 1 (degenerate) 601A is in a state where redundancy by RAID 5 is lost, and since one flash drive 122 of parity group 1 (active spare) 601B is added, the RAID configuration of storage controller 102 A rebuild is performed.

Specifically, the data is read from the three flash drives 122 belonging to the parity group 1 (degenerate) 601A, the parity calculation is performed, and the calculation result is written to the flash drive 122 of the parity group 1 (active spare) 601B. .

At this time, Read is concentrated on the flash drive 122 of the parity group 1 (degenerate) 601A, while Write is concentrated on the flash drive 122 of the parity group 1 (active spare) 601B.

Even in such a situation, by applying the power performance optimization program 200 including the processing shown in FIG. 7 of the present invention, the power of the storage system 101 can be optimized and the performance can be improved.

FIG. 12 is a diagram showing an example of a drive operation mode state table during reconstruction of a RAID configuration. As shown in FIG. 12, the operation mode of the flash drive 122 belonging to the parity group 1 (degeneration) 601A is changed to mode ID = 3, and the operation mode of the flash drive 122 belonging to the parity group 1 (active spare) 601B is changed to the mode ID. By changing to = 1, the rebuilding time of the RAID configuration can be greatly shortened.

In the parity group configuration information table 221, by having a plurality of power allocation change enable / disable flags, for example, in the case of data access from the host 100 and in the case of access such as RAID configuration reconstruction by the storage controller 102, respectively Implementations that change the range of power reallocation are also possible.

As described above, in this embodiment, when the storage controller 102 detects the flash drive 122 in which writing is concentrated, the allocation of power consumption of the flash drive 122 with a low writing load is reduced. The storage controller 102 can improve the performance of the entire storage system 101 within the range of power that can be supplied by increasing the power by relaxing the restriction on the power consumption of the flash drive 122 where writing is concentrated. .

Note that the flash drive 122 in which writing is concentrated has a write load exceeding the first threshold value and is subject to power consumption restriction relaxation, and the flash drive 122 with a low writing load has a power consumption that can allow a reduction in power consumption. It is a target for power reduction (target for tightening restrictions).

Further, when the allocation of power consumption to the flash drive 122 is increased, the cooling performance can be increased, so that the flash drive 122 can be prevented from overheating and the performance of the flash drive 122 can be stabilized. it can.

<Summary>
The selection of the flash drive 122 whose power consumption is to be increased and the selection of the flash drive 122 whose power consumption is to be reduced are not limited to the above embodiment. For example, the operation mode may be set to “3” with all but the flash drive 122 whose power consumption is to be increased as the power consumption reduction target, depending on the balance between the performance required for the storage system 101 and the power consumption. Can be changed as appropriate.

In the above conventional example, the flash drive 122 employing the NAND flash memory as the nonvolatile semiconductor memory is shown as the nonvolatile semiconductor drive. However, the present invention is not limited to this example. Therefore, the present invention can be applied to a nonvolatile semiconductor drive in which power consumption increases with the increase of.

In addition, this invention is not limited to the above-mentioned Example, Various modifications are included. For example, the above-described embodiments are described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. In addition, any of the additions, deletions, or substitutions of other configurations can be applied to a part of the configuration of each embodiment, either alone or in combination.

In addition, each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. In addition, each of the above-described configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be stored in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

Also, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

Claims

A storage device including a plurality of nonvolatile semiconductor drives;
A storage system including a processor and a memory to control the storage device,
The processor measures the write load of the nonvolatile semiconductor drive,
When the processor detects a nonvolatile semiconductor drive in which the write load exceeds a predetermined threshold, the processor selects the nonvolatile semiconductor drive as a power consumption restriction relaxation target,
The processor selects a non-volatile semiconductor drive that is allowed to reduce power consumption from non-volatile semiconductor drives other than the power consumption restriction relaxation target, as a power consumption restriction strengthening target,
The processor increases the power consumption of the non-volatile semiconductor drive subject to the restriction of power consumption after reducing the power consumption of the non-volatile semiconductor drive subject to the restriction of power consumption restriction. .
The storage system according to claim 1,
The storage device
A cooling device for cooling the nonvolatile semiconductor drive,
The processor is
A storage system that enhances the cooling performance of the cooling device when increasing the power consumption of a nonvolatile semiconductor drive subject to relaxation of the power consumption restriction.
The storage system according to claim 1,
The processor is
After changing the power consumption of the non-volatile semiconductor drive subject to restriction restriction and restriction relaxation, the nonvolatile semiconductor drive measures the write load of the non-volatile semiconductor drive, and the write load exceeds a predetermined threshold If there is no power consumption, the storage system is characterized in that the power consumption of the non-volatile semiconductor drive subject to the restriction of power consumption restriction and the non-volatile semiconductor drive subject to the restriction restriction of power consumption is restored.
The storage system according to claim 3,
The storage device
A cooling device for cooling the nonvolatile semiconductor drive,
The processor is
When the power consumption of the non-volatile semiconductor drive subject to the restriction of power consumption restriction and the non-volatile semiconductor drive subject to the restriction restriction of power consumption is restored, the cooling performance of the cooling device is restored. Storage system.
The storage system according to claim 1,
The storage device
A redundant cooling device for cooling the nonvolatile semiconductor drive;
The processor is
When degeneration occurs in the redundant cooling device, the change in power consumption is prohibited even when a nonvolatile semiconductor drive having a write load exceeding a predetermined threshold is detected. system.
The storage system according to claim 1,
The memory is
Storing a power allocation change enable / disable flag set in advance as to whether or not to allow change of the power consumption for each nonvolatile semiconductor drive;
The processor is a non-volatile semiconductor drive that is allowed to reduce power consumption from a non-volatile semiconductor drive that is allowed to change the power consumption among non-volatile semiconductor drives that are not targeted for relaxation of the power consumption restriction. A storage system characterized in that a selectable semiconductor drive is selected as a target for enforcing a restriction on power consumption.
The storage system according to claim 1,
The nonvolatile semiconductor drive is
A storage system, wherein the power consumption is limited in an operation mode corresponding to the multiplicity of writing set in the nonvolatile semiconductor drive.
The storage system according to claim 1,
The processor is
A storage system, wherein write IOPS for the nonvolatile semiconductor drive is used as a write load of the nonvolatile semiconductor drive.
The storage system according to claim 1,
The processor is
The power consumption is reduced so that the sum of the power consumption of the nonvolatile semiconductor drive subject to relaxation of the power consumption restriction and the power consumption of the nonvolatile semiconductor drive subject to restriction of power consumption is less than or equal to the power that can be supplied by the power source. A storage system, wherein a non-volatile semiconductor drive to be subjected to power restriction is selected.
A storage device control method for controlling a storage device including a plurality of nonvolatile semiconductor drives in a control device including a processor and a memory,
The processor includes a first step of measuring a write load of the nonvolatile semiconductor drive;
The processor, when detecting a nonvolatile semiconductor drive in which the write load exceeds a predetermined threshold, selecting the nonvolatile semiconductor drive as a power consumption restriction relaxation target; and
The processor selects a non-volatile semiconductor drive that is allowed to reduce power consumption among non-volatile semiconductor drives other than the power consumption restriction relaxation target as a power consumption restriction strengthening target;
A fourth step of increasing the power consumption of the non-volatile semiconductor drive targeted for relaxation of the restriction of power consumption after reducing the power consumption of the non-volatile semiconductor drive targeted for strengthening of the restriction of power consumption;
A method for controlling a storage apparatus.
A storage control device for controlling a storage device including a plurality of nonvolatile semiconductor drives, including a processor and a memory,
The processor measures the write load of the nonvolatile semiconductor drive,
When the processor detects a nonvolatile semiconductor drive in which the write load exceeds a predetermined threshold, the processor selects the nonvolatile semiconductor drive as a power consumption restriction relaxation target,
The processor selects a non-volatile semiconductor drive that is allowed to reduce power consumption from non-volatile semiconductor drives other than the power consumption restriction relaxation target, as a power consumption restriction strengthening target,
The storage control is characterized in that the processor increases the power consumption of the non-volatile semiconductor drive targeted for relaxation of the power consumption restriction after reducing the power consumption of the non-volatile semiconductor drive targeted for strengthening the restriction of power consumption. apparatus.