JP5246872B2 - Storage system and storage management method - Google Patents

Description

  The present invention relates to a storage device system (hereinafter also referred to as “storage system”) for storing computer data, and more particularly to a hierarchical storage system in which a plurality of storage devices composed of a plurality of disk devices are arranged in a hierarchy.

  In recent years, there has been a problem of a rapid increase in energy consumption accompanying the enlargement of data centers. Among them, the rate of energy consumed by the storage system is increasing as the amount of data increases rapidly. In particular, a high-performance and large-capacity storage system occupies a large proportion of energy consumption used in applications for online transaction systems and HPC (High Performance Computing) systems. Recently, approximately 20% to 30% of energy consumption in data centers is consumed in storage systems, and there is a report that the rate will increase with the rapid increase in data volume. It is considered to be one of the important issues.

  As means for solving the above-described problem, Patent Documents 1 and 2 disclose techniques for controlling power on / off of a hard disk mounted on a storage system. The method of Patent Document 1 is a method in which power control is performed in units of hard disks in a hard disk group constituting a RAID (Redundant Arrays of Independent Disks), and a product using the method responds to an access request without delay. In order to achieve this, there is a method of providing a plurality of hard disks that are always operating (Non-Patent Document 1).

  The method of Patent Document 2 is a method of turning off the power of a hard disk of a hard disk group that constitutes a RAID that is not accessed or putting it in a power saving state.

  Further, as another means for solving the above problem, Non-Patent Document 2 discloses a technique using a hierarchical storage system. The method uses a high-speed storage device and a low-power, large-capacity storage device in a hierarchical storage system. The generated data is first stored in the high-speed storage device, and the data with low access frequency is stored in the high-speed storage device. By shifting from a storage device to a low-power / large-capacity storage device, the increase in the capacity of a high-speed storage device with high power consumption is suppressed, and the increase in energy consumption as a whole system is suppressed.

  As another means for solving the above problem, Patent Document 3 discloses a technique for controlling the number of revolutions of a hard disk using information held in a database management system in a storage system used for a database. . The method usually involves spinning down hard disks to reduce power consumption, and using a query plan (column of storage device access processing) created by the database management system, which hard disks are accessed in advance. It was a method of knowing and spinning up / down the accessed hard disk.

US Patent Application Publication No. 2004/0054939 JP 2000-293314 A JP 2007-293479 A

[online] [Search on March 16, 2009], Internet <URL: http: // www. copasys. com / pdfs / Revolution200TDataSheet. pdf> ILM and Tiered Storage. Storage Networking Industry Association, 2006.

  As a method for saving power in a storage system, the conventional techniques disclosed in Patent Document 1 and Non-Patent Document 1 are provided with a plurality of hard disks that are always operating in order to respond to access requests without delay. However, if the requested data is not stored on the hard disk, the power-saving hard disk is returned to the operating state before being accessed, so the response penalty is large and it cannot be applied to applications that require high performance. There was a problem.

  Further, since the conventional technique disclosed in Patent Document 2 is passive power control in which the power of the hard disk is turned off when there is no access, it responds to an access request without delay as in the above method. There is a problem that it cannot be applied to applications that require high performance.

  In addition, the conventional technique disclosed in Non-Patent Document 2 has an access request when data is frequently migrated to a low-power / large-capacity storage device in order to suppress the capacity of a high-speed storage device with high power consumption. The proportion of data stored in low-speed, low-power, large-capacity storage increases, and the performance of the entire system deteriorates. There is a trade-off between performance and power saving, and both high performance and power saving are compatible. There was a problem that it was difficult.

In addition, the prior art disclosed in Patent Document 3 is based on information of a query plan (a column of access processing to a storage device), and the database management system determines when and which hard disk the database management system accesses the storage. It can be determined by itself, ie the application. However, there is a problem that it cannot be applied to a batch processing type application in which the execution start time of the process cannot be determined by the application itself, for example, an application such as scientific calculation executed in the HPC system.
Therefore, an object of the present invention is to achieve both high performance and low power consumption even when the execution start time of processing executed on a computer cannot be determined by the application itself.

  In order to solve the above-described problems, an embodiment of the present invention has the following configuration. Specifically, one or more first hard disks (corresponding to reference numeral 42) connected to a plurality of computers to which a first management apparatus (corresponding to reference numeral 18) is connected. The first storage device (corresponding to reference numeral 51) having a volume (corresponding to reference numeral 51) and one or more second hard disk devices (corresponding to reference numeral 43) connected to the first storage device. A second storage apparatus (corresponding to reference numeral 12) having one or more second volumes (corresponding to reference numeral 52), the first storage apparatus, the second storage apparatus, and the first management. A storage system having a second management apparatus (corresponding to reference numeral 19) connected to the apparatus, wherein the first management apparatus is information on jobs (job information) that are sequentially executed on the computer, and is being executed The second management device collects the job information, collects the job queue information (corresponding to reference numeral 24), and collects Analyzing means (corresponding to reference numeral 25) for analyzing the job information and the job queue information, the analyzing means specifying a second volume accessed by the job from the job information, and the job Means for calculating an average waiting time until the execution of each job waiting for execution is started from the queue information.

  The second storage device has a means for controlling the power supply of the second hard disk device, and all the data is stored in the second volume and constitutes the second volume that is not accessed. The second hard disk drive is turned off.

  The second management device powers on the second hard disk device that constitutes the second volume specified by the analyzing means to put it in an operating state, and sets the second volume to the first volume. Means for calculating a threshold time required for copying to the disk, means for comparing the average waiting time with the threshold time, an instruction for controlling the power supply of the second hard disk device, the first storage device, and the first storage device. The second management device has a means for issuing an instruction to copy data between the two storage devices, and when the average waiting time of the job is shorter than the threshold time of the job when the job is submitted Issues an instruction to delay the execution of the job to the first management apparatus at least for the threshold time from the time the job is submitted.

  In addition, the second management device turns on the second hard disk device constituting the second volume accessed by the job to put it in an operating state, and copies the second volume to the first volume. Is issued to the second storage device and the first storage device, and after the copying of the second volume to the first volume is completed, the second volume constituting the second volume is configured. An instruction to turn off the power of the hard disk device is issued to the second storage device.

In addition, when the average waiting time of the job is longer than the threshold time of the job when the job is submitted, the second management device reaches the threshold time of the job at the latest. Immediately before, the second hard disk device constituting the second volume accessed by the job is turned on to be in an operating state, and an instruction to copy the second volume to the first volume is sent to the second volume. Issued to the storage device and the first storage device.
In addition, the problem which this application discloses and the solution method are clarified by the column and drawing of embodiment of invention.

  According to the present invention, even when the execution start time of processing executed on a computer cannot be determined by the application itself, both high performance and low power consumption can be achieved.

It is a figure which shows the structural example of the computer system containing a storage system. It is a figure which shows an example of a function structure of a computer management server and a storage management server. It is a figure which shows the example of a structure of a 1st tier storage apparatus. It is a figure which shows the example of a structure of a 2nd tier storage apparatus. It is a figure which shows the example of a structure of a file server. It is a figure which shows an example of the state of a JOB queue. It is a figure which shows an example of the JOB queue information table 1 which a computer management server manages. It is a figure which shows an example of the JOB queue information table 2 which a storage management server manages. FIG. 5 is a diagram illustrating an example of a correspondence relationship between a file storage directory, a file storage virtual volume, and a file storage first and second volume when the volume is not staged. It is a figure which shows an example of a volume management table. It is a figure which shows an example of a volume use condition management table. FIG. 5 is a diagram illustrating an example of a procedure for determining the staging of a volume between a file server and first and second tier storage apparatuses, and a staging / destaging procedure. FIG. 5 is a diagram illustrating an example of a procedure for determining the staging of a volume between a file server and first and second tier storage apparatuses, and a staging / destaging procedure. FIG. 5 is a diagram illustrating an example of a procedure for determining the staging of a volume between a file server and first and second tier storage apparatuses, and a staging / destaging procedure. It is a figure which shows an example of a computer execution script. It is a figure which shows an example of the correspondence of the file storage directory at the time of volume staging, the file storage virtual volume, and the file storage first and second volumes. FIG. 10 is a diagram showing another example of the determination of the volume staging timing between the file server and the first and second tier storage apparatuses, and the staging / destaging procedure. FIG. 10 is a diagram showing another example of the determination of the volume staging timing between the file server and the first and second tier storage apparatuses, and the staging / destaging procedure. FIG. 10 is a diagram showing another example of the determination of the volume staging timing between the file server and the first and second tier storage apparatuses, and the staging / destaging procedure.

  Hereinafter, modes for carrying out the present invention (referred to as “embodiments”) will be described with reference to the drawings.

≪First embodiment≫
FIG. 1 is a diagram illustrating a configuration example of a computer system including a storage system according to the first embodiment. The computer system 1 includes a storage system 2, an IP switch 16, a computer 14, and a computer management server 18. The storage system 2 includes a file server 13, a first tier storage device (first storage device) 11, a second tier storage device (second storage device) 12, a fiber channel (FC: Fiber Channel) switch 17, And a storage management server 19.

  As illustrated in FIG. 1, the storage system 2 and the computer 14 are connected by connecting the file server 13 and the computer 14 via the IP switch 16. The computer management server 18 and the storage management server 19 are connected to each other via a LAN (Local Area Network) 15. Further, the storage management server 19, the file server 13, the first tier storage apparatus 11, and the second tier storage apparatus 12 are connected to each other via a LAN 15.

  The first tier storage apparatus 11 is directly connected to the file server 13. As the connection interface, it is common to use an interface of a protocol for sending block data, such as Fiber Channel or iSCSI (Internet Small Computer System Interface). Here, there is no problem even if the first tier storage apparatus 11 is connected to the file server 13 via a switch.

  The second tier storage apparatus 12 is connected to the first tier storage apparatus 12 via the FC switch 17. As a connection interface, there is no problem even if an interface of a protocol for sending block data such as iSCSI other than Fiber Channel is used.

The first tier storage apparatus 11 has a file storage first volume (first volume) 51 for storing a file to be input / output processed by the file server 13. The second tier storage apparatus 12 has a file storage second volume (second volume) 52 for storing a file for which the file server 13 performs input / output processing. In addition, the first tier storage apparatus 11 is a virtual that virtually provides a volume that the second tier storage apparatus 12 has as a volume that the first tier storage apparatus 11 provides to the computer 14, that is, as a file storage virtual volume 61. It has a function to convert.
Note that “usr1”, “usr2”,... Shown in the file storage virtual volume 61 and the file storage second volume 52 in FIG. 1 execute file input / output processing by the computer 14. This means a user who uses the volume for this purpose. In other words, each user is assigned a volume that can be used in the first tier storage apparatus 11 and the second tier storage apparatus 12. However, the allocation method can be changed according to the operation of the computer system 1, for example.

  FIG. 3 shows an example of the configuration of the first tier storage apparatus 11. The controller 31 is connected to a channel IF (interface) unit 32 that controls data write / read access from a host device such as the file server 3 and the computer 14, and a plurality of high-speed hard disks (first hard disk devices) 42. A disk IF (interface) 33 for controlling data write / read access to the high-speed hard disk 42; a cache memory 34 for temporarily storing write / read data to the high-speed hard disk 42; and a control memory 38 for storing control data. And a coupling unit 35 for connecting the channel IF unit 32, the disk IF unit 33, and the cache memory 34. The coupling unit 35 is generally composed of one or more switches, but there is no problem even if it is composed of one or more common buses.

  The channel IF unit 32 controls data transfer with the cache memory 34 when receiving data write / read access from the host device, and the disk IF unit 33 writes / reads data to / from the high-speed hard disk 42. Sometimes it controls data transfer to and from the cache memory 34. By such data exchange between the channel IF unit 32 and the disk IF unit 33 via the cache memory 34, data is written / read to / from the high-speed hard disk 42 from the host device. In order to perform such control, the channel IF unit 32 and the disk IF unit 33 have one or more processors (not shown). An internal LAN 37 is connected to this processor. Further, a storage management server 19 outside the first storage device 11 is connected to the internal LAN 37 via the LAN 15.

Here, the configuration of the controller 31 described above is merely an example, and the configuration is not limited to the above. There is no problem if the controller 31 has a function of writing / reading data to / from the high-speed hard disk 42 in response to a data write / read request from the computer 14.
Further, the controller 31 may include a power control unit 36 that controls the power on / off (turning on / off) of the high-speed hard disk 42. In this case, the power control unit 36 is connected to the internal LAN 37.

  The hard disk mounting unit 41A (41) has a hard disk power supply 46A (46) for supplying power to each of the plurality of high-speed hard disks 42. The plurality of high-speed hard disks 42 are grouped into a RAID group (Gr.) 1:44 composed of a plurality of high-speed hard disks 42.

  Here, as the high-speed hard disk 42, a high-speed hard disk having a rotation speed of 10,000 rpm (revolution per minute) or 15,000 rpm and having an FC or SAS (Serial Attached SCSI) interface is generally used. Further, there is no problem even if a solid state memory (SSD) that has recently been installed in a storage apparatus is used. By doing so, the first tier storage apparatus 11 can be made even faster and consume less power than when a high-speed hard disk is used.

  Here, the hard disk power source 46A is connected to each individual high-speed hard disk 42 or RAID Gr. There may be no problem even if one or two (in the case of a redundant configuration) are provided every 1:44.

A power supply control unit 36 in the controller 31 is connected to the hard disk power supply 46A, and controls on / off of the power supply.
Here, there is no problem even if the power supply control unit 36 is not in the controller 31 but in the hard disk mounting unit 41A. Further, there is no problem even if the power control unit 36 is directly connected to the storage management server 19.

  The file storing first volume 51 described in the explanation of FIG. 1 is a RAIDGr. It is formed on a 1:44 area.

  FIG. 4 shows an example of the configuration of the second tier storage apparatus 12. The controller 71 is a computer connection port 76 for connecting a host device such as the first tier storage device 11, a disk connection port 78 for connecting a plurality of large capacity hard disks (second hard disk devices) 43, and writing to the large capacity hard disk 43. / A shared memory 73 for temporarily storing read data and a processor 72. Further, the computer connection port 76, the disk connection port 78, the processor 72, and the shared memory 73 are connected via a coupling unit 74. The coupling unit 74 is generally composed of a switch, but there is no problem even if it is composed of a common bus.

  The processor 72 controls data transfer between the computer connection port 76 and the shared memory 73 when receiving data write / read access from the host device, and at the time of data write / read to the large-capacity hard disk 43. The data transfer between the large-capacity hard disk 43 and the shared memory 73 is controlled. By exchanging data between the computer connection port 76 and the large-capacity hard disk 43 via the shared memory 73, data is written / read to / from the large-capacity hard disk 43 from the host device.

  An internal LAN 77 is connected to the processor 72. Further, a storage management server 19 outside the second storage device 12 is connected to the internal LAN 77 via the LAN 15.

Here, the configuration of the controller 71 described above is merely an example, and the configuration is not limited to the above. There is no problem if the controller 71 has a function of writing / reading data to / from the large-capacity hard disk 43 in response to a data write / read request from the computer 14.
Further, the controller 71 may include a power control unit 75 that controls on / off (turning on / off) the power of the large-capacity hard disk 43, and in this case, the power control unit 75 is connected to the internal LAN 37.

  The hard disk mounting unit 41B (41) and the hard disk power supply 46B (46) are the same as the configuration (41A, 46A) of the first tier storage apparatus 11 shown in FIG.

  Here, the large-capacity hard disk 43 has a rotation speed of 7,200 rpm (revolution per minute) or less, a SATA (Serial Advanced Technology Attached) interface, and power consumption per capacity is smaller than that of the high-speed hard disk 42. It is common to use large capacity, low power hard disks. Further, when access does not come, a hard disk having a power saving function such as reducing the number of rotations to reduce power consumption may be used.

  The file storage second volume 52 described in the description of FIG. 1 is a RAID Gr. It is formed on the area of 2:45.

  3 and 4, the general configurations of the first tier storage apparatus 11 and the second tier storage apparatus 12 have been described. However, they are not limited to the configurations described above. The I / O (Input / Output) processing performance required for the first tier storage apparatus 11 may be an apparatus having an I / O processing performance exceeding that of the second tier storage apparatus 12. The specifications required for the second tier storage apparatus 12 may be any apparatus that can realize the capacity required by the computer 14 with a smaller number of hard disks than the first tier storage apparatus 11. In other words, any device that consumes less power per capacity may be used.

  Here, there is no problem even if the first tier storage apparatus 11 and the second tier storage apparatus 12 are configured by one storage apparatus. That is, for example, in the first tier storage apparatus 11, the high-speed hard disk 42 and the large-capacity hard disk 43 are mixedly mounted in the hard disk mounting unit 41A, and RAID Gr. 1:44, RAID Gr. 2:45, and RAID Gr. 1:44 and RAID Gr. A file storage first volume 51 and a file storage second volume 52 may be formed on the 2:45 area, respectively. By doing so, the power consumption of the controller 71 of the second tier storage apparatus can be reduced.

  FIG. 5 shows an example of the configuration of the file server 13. The file server 13 includes an input / output controller 251, an input / output controller 252, a processor 250, and a memory 253. The input / output controller 251 is connected to the IP switch 16 and performs input / output processing of file data. The input / output controller 252 is connected to the first tier storage apparatus 11 and performs writing and reading processing of block data to the first tier storage apparatus 11. In the memory 253, data buffering / caching is performed between the input / output controller 251 and the input / output controller 252. In the processor 250, LINUX (registered trademark) operates as an OS (Operating System), and NFS (Network File system) operates as its file system. This file system performs processing for converting file data accessed from the host server into block data addresses. Management information such as a conversion table necessary for conversion between file data and block data is stored in the memory 253 by the processor 250. Here, the OS is not limited to LINUX, and the file system is not limited to NFS. There is no problem if it has a function of receiving file data from the host server, converting it to block data, and accessing the first tier storage apparatus 11.

FIG. 2 shows functional configurations of the computer management server 18 and the storage management server 19.
The computer management server 18 includes a job management unit 21 that manages jobs (jobs) executed by the computer 14, a user management unit 22 that manages users who request the computer 14 to execute jobs, and job information executed by the computer 14. An information providing unit 23 serving as an interface provided to the storage management server 19 is provided. The job refers to a job that is sequentially executed in a batch processing type application.

  Note that each functional unit such as the JOB management unit 21 and the information analysis unit 25 used when describing the components in the present embodiment may be logically configured by software (program), or a dedicated LSI ( (Large Scale Integration) or the like, or may be realized by a combination of software and hardware. When logically configured, each function unit of the storage management server 19 is stored on the memory 94 (storage unit), and the function is realized by processing performed by the processor 95 (control unit). Is done. Each functional unit of the computer management server 18 is stored on the memory 99, and the function is realized by executing processing by the processor 98 (computer management server control unit).

  The JOB management unit 21 includes an input JOB management unit 201, a JOB scheduler 202, and an end JOB management unit 206. The JOB scheduler 202 has a waiting queue 203 and an execution queue 205.

  The user creates a calculation execution script 234 shown in FIG. 13 and inputs it to the computer management server 18 in order to execute the calculation JOB (JOB) on the computer 14. The input is, for example, a client terminal (not shown) connected to the computer management server 18 directly or via the GUI (Graphical User Interface) or CLI (Command Line Interface) of the computer management server 18. )

  The input calculation execution script 234 is managed by the input JOB management unit 201 and is prepared in the waiting queue 203 in descending order of priority: queue 1: 211, queue 2: 212, queue 3: 213, or queue 4 : Any one of 214. The distribution method, that is, the method of assigning priorities is, for example, the number of CPUs (Central Processing Units) 301 (Number of CPUs) 301 and the maximum calculation time (MAX CPU TIME) 302 described in the calculation execution script 234. , The amount of main memory capacity (Memory Size) 303 to be used, and the like, and the user explicitly specifies the priority order in the calculation execution script 234. Jobs are executed in the order in which they are queued. In addition, the execution order of the jobs in the queues 1 to 4: 211 to 214 is sequentially executed from the queue 1: 211 having a higher priority. After a job with a high priority queue enters execution, if the CPU resource of the computer 14 is free and the next priority job can be executed, that job is also executed on the free CPU in parallel. And execute. The same applies to the job with the next priority. The job being executed is managed in the execution queue 205, and when the job is completed, the management is transferred to the end job management unit 206.

  The user management unit 22 manages users who use the computer 14 from the computer management server 18 or a client terminal connected to the computer management server 18, that is, a user for storing files used for user authentication and calculation by the user. Manage the directory. For this management, for example, a protocol such as NIS (Network Information Service) or LDAP (Lightweight Directory Access Protocol) is used.

  In addition, the information providing unit 23 transmits the calculation execution script 234 of each JOB and the information indicating the execution order of the JOB to the storage management server 19 and information on the user who uses the computer 14 and the user directory used by the user. To do.

  The storage management server 19 includes an information collection unit 24 that receives JOB information and JOB queue information executed by the computer 14 from the information providing unit 23 of the computer management server 18, and stores the JOB information and JOB queue information in the storage device (11, 12). The information analysis unit 25 that performs analysis for use in the server, mounts / unmounts the volumes of the first tier storage apparatus 11 and the second tier storage apparatus 12 to the user directory managed by the file server 13 based on the analyzed information A volume management unit 26 that manages staging / destaging of files or volumes between the first tier storage device 11 and the second tier storage device 12, and a user area management unit that manages user directories handled by the file server 13. 27, file server 3. Instruct the volume allocation and volume mounting / unmounting to the first tier storage apparatus 11 and the second tier storage apparatus 12, and the power control unit 36 in the first tier storage apparatus 11 and the second tier storage apparatus 12. , 75 has a storage management unit 28 for instructing power control of the hard disk.

  FIG. 6 shows the state of a queue when one JOB is input in one queue (queue 4: 214). Here, λ is an average input frequency when a JOB is input to the queue, and an inverse of μ (average execution frequency when the JOB input to the queue is executed) is an average JOB when the JOB is executed It represents the execution time (average execution time) Te. The reciprocal of λ represents the average JOB input interval Ti. Tw represents an average waiting time until the execution of the JOB (in this figure, an average waiting time until the execution of the input JOB). The calculation method of the average waiting time is shown below.

  FIG. 7 shows an example of the job queue information table 1:70 stored in the job management unit 21 of the computer management server 18. The job queue information table 1:70 stores a value indicating the current state of each job. A “JOB ID (Identifier)” 701 indicates identification information for identifying a JOB. JOB IDs are assigned in the order in which JOBs are input. A “JOB status” 702 indicates the current status of each JOB. “In progress” indicates that the job is being executed. “Waiting” indicates that the job is waiting to be executed. JOBs waiting to be executed are given a priority indicating the order of execution. Usually, the priority is assigned so as to increase the priority of the JOB input first. However, it is possible for the user to specify the priority, or for the administrator of the computer management server 18 to change the priority based on the calculation conditions and the usage status of the computer 14. Here, the format of the JOB queue information table 1:70 in FIG. 7 is merely an example, and the format is not limited as shown in the figure. It is sufficient that at least the information described above is included.

  FIG. 8 shows an example of the JOB queue information table 2:80 stored in the memory 94 of the storage management server 19. “JOB ID” 801 indicates identification information for identifying a JOB. “JOB status” 802 indicates the current status (queue status) of each JOB. “In progress” indicates that the job is being executed. “Waiting” indicates that the job is waiting to be executed. JOBs waiting to be executed are given a priority indicating the order of execution. “Tw” 803 indicates the average waiting time of each JOB. When “execution” is performed as in JOB1, the waiting time is “0”, so Tw is not stored. “User ID” 804 indicates identification information of a user (user) who executes each JOB. “Target dir” 805 indicates identification information of a directory used by each JOB. “Tth” 806 turns on the storage device (storage device) constituting the logical volume used to execute each JOB and puts it in an operating state, and sets the logical volume or the file to be used stored in the volume to the first. The time required for staging (copying) to the file storage first volume 51 in the tiered storage 11 (referred to as threshold time) is shown. Here, the format of the JOB queue information table 2:80 in FIG. 8 is merely an example, and the format is not limited as shown. There is no problem if at least the information described above is included.

  FIG. 9 shows an example of the relationship between the user directory and the volume before the user starts the calculation. The file server 13 mounts the file storage virtual volume 61 as a file storage directory 81 that is a user directory based on an instruction from the storage management unit 28.

Based on an instruction from the storage management unit 28, the controller 31 in the first tier storage apparatus 11 converts the file storage second volume 52 in the second tier storage apparatus 12 into a file storage in the first tier storage apparatus 11. It is virtualized as a virtual volume 61 and managed by the first tier storage apparatus 11. By doing so, the first tier storage apparatus 11 can collectively manage the volumes of the second tier storage apparatus 12, thereby simplifying volume management.
Here, the file storage second volume 52 may be directly mounted as the file storage directory 81.

  An example of the relationship between the user directory, the file storage virtual volume 61, and the file storage second volume 52 is shown in the directory tree 101. Here, a separate file storage second volume 52 is assigned to each of the directories dir0 and dir1 under usr0 and the directories dir0, dir1 and dir2 under usr1.

  Here, the correspondence (mapping) between the user directory (file storage directory 81) and the file storage second volume 52 is managed by the user area management unit 27 in the storage management unit 19. This correspondence is created or changed when a user issues a request from a client terminal connected to the storage management server 19.

  All files handled by the user (including files input / output from the computer 14) are stored in the file storage directory 81, that is, the file storage second volume 52 in the second tier storage apparatus 12. Further, normally, the file storage second volume 52 is unmounted and the file storage second volume 52 is configured while it is not accessed by the computer 14 or the user or when the file / volume staging / destaging is not performed. The large capacity hard disk 43 to be turned off is turned off. By doing so, it is possible to reduce the power consumption of the entire storage system 2.

  Here, without turning off the power of the large-capacity hard disk 43, spin down (rotating the disk of the hard disk device below a predetermined rotation speed (rotation speed stored in the memory)), spin-off (removal of the hard disk device) The disk rotation may be stopped) or the power saving mode may be set. By doing this, the amount of power consumption to be reduced is reduced, but it is possible to shorten the time to start up the large-capacity hard disk 43 (time until input / output is enabled) before access.

  The directory tree 101 is managed for each user by, for example, the volume management table 100 shown in FIG. The volume management table 100 is stored in the memory 94 of the storage management server 19. “User ID” 1001 indicates identification information of each User. “Dir ID” 1002 indicates directory identification information for each user. “LU ID” 1003 indicates identification information of a logical unit (logical volume) corresponding to each dir. “RAID ID” 1004 indicates identification information of a RAID group constituting each LU. The RAID group is composed of a plurality of storage devices. Here, the format of the volume management table in FIG. 10 is merely an example, and the format is not limited as shown. There is no problem if at least the information described above is included.

  FIG. 11 shows the volume usage status management table 110. The volume usage status management table 110 indicates the usage status of each logical volume in the first tier storage apparatus 11. The volume usage status management table 110 is stored in the memory 94 of the storage management server 19. “LU ID” 1101 indicates identification information of each logical volume of the first tier storage apparatus 11. “Usage status” 1102 indicates whether each logical volume of the first tier storage apparatus 11 is used by a User (computer). “Size” 1103 indicates the capacity of each logical volume. Here, the format of the volume usage status management table 110 in FIG. 11 is merely an example, and the format is not limited as shown. It is sufficient that at least the information described above is included.

  FIG. 12 (generic name of FIG. 12A to FIG. 12C) shows a file staging / destaging procedure in the storage system of this embodiment. This procedure is executed in parallel for each queue in the waiting queue 203 (queue 1: 211 to queue 4: 214). By the processor 95 being the processing subject, functions by the respective functional units (24, 25, etc.) are realized, and the above procedure is repeated periodically.

  First, in step 401, the information collection unit 24 of the storage management server 19 periodically receives information indicating the execution order of all jobs in the waiting queue 203 from the information providing unit 23 of the computer management server 18 and the end job management unit. The end JOB information (JOB queue information) in 206 is acquired. When acquiring, for example, a command for monitoring the JOB scheduler 202 of the computer management server 18 is transmitted to the computer management server 18, and the computer management server 18 receives the information as a response to the command.

In step 402, the job queue information received last time described in the job queue information table 2:80 and the current state of each job described in the job queue information table 1:70 received in step 401 are displayed for each job ID. In comparison with the above, it is checked whether or not the state of the queue has changed for one corresponding queue (queue 1: 211 to queue: 214) in the waiting queue 203. Here, the state of the queue has changed means that a new job has been entered, a job in the wait queue has been moved to execution, a job in the wait queue has been canceled, or a job has been executed. I mean at least. More specifically, this means that the value of the job status 702 and the value of the job status 802 of the record are different.
Here, the JOB ID of the job being executed (JOB1 in the case of FIG. 7 or FIG. 8) is managed in the execution queue 205. Further, the JOB ID of the JOB that has been executed is managed by the ending job management unit 206.

  If the queue state has not changed (No in step 402), the process returns to step 401 and waits for the next queue state acquisition. If the queue status has changed (Yes in step 402), the job ID and job status columns (801, 802) in the job queue information table 2:80 are replaced with the acquired contents, and the process proceeds to step 403.

  In step 403, it is checked whether or not the change in the queue state is the end of job execution (completion). If the change in the queue state is the end of job execution (Yes in step 403), the process proceeds to step 418. If the change in the queue state is not the end of job execution (No in step 403), the process proceeds to step 404. Here, it can be confirmed that the change of the queue state is the end of job execution by the corresponding job in the execution queue 205 having moved to the job end management unit 206.

Next, in step 404, if the change in queue status is JOB input, the job execution calculation script 234 of the corresponding queue in the waiting queue 203 is acquired and analyzed. Before explaining the analysis, FIG. 13 shows an example of a calculation execution script 234 describing information on an execution job (or calculation) executed by the computer 14. Since a plurality of users input the calculation execution script 234, FIG. 13 shows an image with a plurality of calculation execution scripts 234. Scheduling of mount / unmount of volumes of the first tier storage device 11 and the second tier storage device 12 to a user directory (file storage directory 81) managed by the file server 13, and staging of files between the user directories / In order to manage the scheduling of destaging and the user directory handled by the file server 13, the calculation execution script 234 includes at least information about a directory 300 for storing calculation parameters and calculation execution results.
Further, it includes at least the number of CPUs to be used (Number of CPUs) 301, the length of maximum calculation time (MAX CPU TIME) 302, and the main memory capacity to be used (Memory Size) 303. Prioritization of JOBs is performed based on these pieces of information, and the JOBs are distributed to a plurality of queues according to the priorities.
Here, the format of the calculation execution script of FIG. 13 is merely an example, and the format is not limited as shown in the figure. There is no problem if at least the information described above is included.

  In the information analysis unit 25, the directory name (directory information 300) of the input / output file is extracted from the calculation execution script 234 of each job, and the column (805) of the target dir (directory) in the job queue information table 2:80. Enter the directory name in the corresponding JOB ID.

  Furthermore, the average value and variance of job submission intervals and the average value and variance of job execution times are calculated from the job queue information of the corresponding queue for all cases, not only when the change in queue status is job submission. . These average values and variances are used as statistical information every time the status of the corresponding queue is acquired, and the time interval from the time when the previous JOB was input until the next JOB is input, and the previous JOB is executed. By collecting the time interval from the time when the job is executed to the next time, it can be obtained from the collected values.

  Next, in step 405, an average waiting time Tw from λ and μ to the execution of each job shown in FIG. 6 is calculated using an expression derived from the queue theory. Then, the value is input to the corresponding JOB ID in the Tw column (803) of the JOB queue information table 2:80. Here, by using the dispersion values of λ and μ, the average waiting time can be calculated more accurately than when the average value is used.

Further, when a job is input, the directory for file storage in which the corresponding file is stored by tracing the directory tree 101 shown in FIG. 9 from the directory of the file used for calculation by the corresponding job extracted in step 405. The second volume 52 is specified by the volume management unit 26.
Then, a threshold time Tth is calculated. As described above, the threshold time is as follows: “The RAID Gr2: 45 large-capacity hard disk 43 configuring the specified volume is turned on and brought into operation, and the volume or a file to be used stored in the volume “Time required for staging (copying) to the first volume 51 for file storage in the one-tier storage 11”. Tth can be obtained from the size of the file / volume to be staged and the data transfer rate between the first tier storage apparatus 11 and the second tier storage apparatus 12. After calculating Tth, the value is input to the corresponding JOB ID in the Tth column (806) of the JOB queue information table 2:80.
Here, the RAID Gr2: 45 constituting the specified volume is specified from the RAID ID (1004) of the volume management table: 100.

  In step 406, it is checked whether or not the change in the queue state is JOB input. If the change in the queue state is not JOB input (No in step 406), the process proceeds to step 409. If the change in the queue state is JOB input (Yes in step 406), the process proceeds to step 407. Here, it can be confirmed that the change in the queue state is JOB input because a new JOB (JOB ID) exists at the end of the corresponding queue.

  In step 407, using the JOB queue information table 2:80 shown in FIG. 8, the average waiting time (Tw) until execution of the input JOB is compared with the threshold time (Tth). If Tw is long ( In step 407, the process proceeds to step 401). If Tw is equal to or less than Tth (Yes in step 407), the process proceeds to step 408.

  In step 408, the information analysis unit 25 of the storage management server 19 notifies the job management unit 21 of the computer management server 18 so as to wait for execution of the input JOB for at least the threshold time Tth. When Tw is equal to or less than Tth, the file / directory accessed during execution of the JOB (in other words, the second volume 52 for storing files to be accessed) is changed before the corresponding JOB starts execution. This is because staging to the one-tier storage apparatus 11 cannot be completed. In this case, since there is no file / directory accessed to the first tier storage apparatus 11, an input / output error from the computer 14 to the first tier storage apparatus 11 occurs. Alternatively, since input / output is directly performed from the file storage second volume 52 of the second tier storage apparatus 12, the input / output performance is degraded. In order to prevent such a situation from occurring, processing for delaying the execution of JOB is performed for at least the threshold time Tth from the time the job is submitted.

  Next, in step 409, the average waiting time (Tw) until job execution is compared with the threshold time (Tth) plus a certain time (α) for all the queues in the corresponding queue. For comparison, for example, the JOB queue information table 2:80 is used. If Tw is longer than the time obtained by adding α to Tth (No in step 409), the process proceeds to step 401. On the other hand, if Tw is equal to or less than the time obtained by adding α to Tth (Yes in step 409), the process proceeds to step 410. When there is a JOB that satisfies this condition, it means that the execution of the corresponding JOB may be started at least after a time when α is added to Tth. Therefore, it is necessary to perform file or volume staging processing in step 410 and subsequent steps.

  Here, since Tth may vary depending on the operating status of the system, a margin of α is added. If this margin is too large, the files or volumes used by the JOB will be staged from the second tier storage device 12 to the first tier storage device 11 for many JOBs. Will increase. As a result, the number of high-speed hard disks with large power consumption increases, so the power saving effect is reduced. Therefore, α is set to about 10% or less of Tth, for example.

  In this way, even if the average waiting time Tw is larger than the threshold time Tth, the processing after step 410 is suspended until the average waiting time Tw becomes substantially the same value as the threshold time Tth. Is done. Therefore, staging is executed at a much earlier stage than when the job is executed, and waste of power due to the use of the capacity of the first tier storage apparatus 11 for a long time can be saved.

  Next, at step 410, the target RAID Gr. The storage management unit 28 instructs the second tier storage apparatus 12 to turn on (ON) the large-capacity hard disk 43 of 2:45. The target RAID group can be determined using the volume management table: 100 shown in FIG.

  In step 411, the process waits until the power supply of the large-capacity hard disk 43 is completed and the large-capacity hard disk 43 becomes operational.

  When the operating state is reached (Yes in step 411), in step 412, the storage management unit 28 turns on the file server 13 to turn on the file server 13 and sets the second volume 52 for storing files to the file storing directory 81. Instruct to mount. Thereafter, the second tier storage device 52 accessed from the storage management unit 28 to the second tier storage device 12 and the first tier storage device 11, or the file to be accessed stored in the volume is first An instruction is given to stage (copy) to the file storage first volume 51 of the hierarchical storage apparatus 11.

  Here, the first file storage volume 51 has a plurality of volumes LU0 to LUn, and it is necessary to determine which volume of the volumes to copy. It selects from those volumes that are unused and whose size is greater than or equal to the size of the volume or file being copied. The usage status and size of each volume of the file storage volume 51 are managed by the volume usage status management table: 110 shown in FIG. 11, and the storage management unit 28 indicates the unused volume that satisfies the size condition in this table. Select from 110.

In step 413, wait until staging is completed.
When the staging is completed (Yes in step 413), the corresponding file storage second volume 52 (accessed volume) mounted in the file storage directory 81 is unmounted in step 414, and the staging is completed. The storage management unit 28 instructs the file server 13 to remount (switch) the directory in the first volume 51.

  FIG. 14 shows the state of the directory and volume after staging is completed and mount switching is performed. This figure shows a case where the file storage second volume 52 accessed from JOB is LU00. The file stored in LU00 is staged (copied: moved) 110 to LU0 which is unused and has a size capable of storing the file. After the mount switching, the mapping between VLU00 (LU00) and the directory usr0 / dir0 disappears temporarily.

In step 415, the process waits until the mount switching is completed.
When the mount switching is completed (Yes in step 415), information (staging completion information) that staging is completed (data preparation is completed) is sent from the storage management unit 28 of the storage management server 19 to the computer management server 18 in step 416. The job management unit 21 is notified. Based on the staging completion information transmitted from the storage management server 19, the computer management server 18 checks whether or not staging of files necessary for execution of each JOB has been completed before starting execution of each JOB. If the staging is completed, the job is executed. If the staging is not completed, the job is executed after the staging is completed. That is, the execution of the job is delayed until the notification is received. By doing so, the execution of JOB is started before the completion of staging, and it becomes possible to prevent an input / output error or input / output from the second tier storage apparatus 12 with low performance.

  Next, at step 417, the RAID Gr. The storage management unit 28 instructs the second tier storage apparatus 12 to turn off (OFF) the large-capacity hard disk 23 configuring 2:45. After the instruction, the process returns to step 401.

  Next, the processing after step 418 will be described. These processing are processing performed at the end of JOB execution (Yes in step 403). In step 418, the target RAID Gr. Constituting the file storage second volume 52 in which the file / directory accessed from the completed JOB was originally stored / mounted. The storage management unit 28 instructs the second tier storage apparatus 12 to turn on (ON) the large-capacity hard disk 43 of 2:45.

  In step 419, the process waits until the power supply of the large-capacity hard disk 43 is completed and the large-capacity hard disk 43 becomes operational.

  When the operation state is reached (Yes in Step 419), the storage management unit 28 has accessed the first tier storage apparatus 11 and the second tier storage apparatus 12 from the corresponding JOB in the first tier storage apparatus 11 in Step 420. The file storage first volume 51 or the accessed file stored in the volume is instructed to be destaged (copied) to the original file storage second volume 51. Note that, after completion of destaging, the files stored in the file storage first volume 51 may be deleted regardless of the timing.

In step 421, the process waits until the destaging is completed.
When the destaging is completed (Yes in Step 421), the corresponding file storage first volume 51 (the accessed volume) mounted in the file storage directory 81 is unmounted in Step 422, and the destaging is completed. The storage management unit 28 instructs the file server 13 to remount (switch) the directory to the file storage second volume 52.

In step 423, the process waits until the mount switching is completed.
When the mount switching is completed (Yes in Step 423), the storage management unit 28 causes the file server 13 to unmount the file storage second volume 52 that has been destaged from the file storage directory 81 in Step 424. And instruct.

  Next, in step 425, the storage management unit 28 instructs the second tier storage apparatus 12 to turn off (turn off) the large-capacity hard disk 23 that constitutes the file storage second volume 52 that has been destaged. Instruct. After the instruction, the process returns to step 401.

  According to the present embodiment, when a job to be executed by the computer 14 is necessary, a necessary file can be staged on the first tier storage apparatus 11, so that the high performance of the first tier storage apparatus 11 is utilized to the computer 14. It is possible to speed up the input / output of files. Further, the large-capacity hard disk 43 of the second tier storage apparatus 12 can be turned off except when accessed. Furthermore, the capacity of the first tier storage apparatus 11 can be minimized. Therefore, the power consumption of the storage system 2 can be reduced. Therefore, it is possible to provide a high-speed and large-capacity tiered storage system that minimizes performance degradation and enables low power consumption for batch processing type applications that require high performance.

<< Second Embodiment >>
Next, a second embodiment of the present invention will be described.
FIG. 15 (generic name of FIGS. 15A to 15C) shows a staging / destaging procedure of the second embodiment. The procedure shown in FIG. 15 is the same as the procedure shown in FIG. 12 except for the following points (Steps 501 to 525 in FIG. 15 are the same as those shown in FIG. 12, except for the following points. Is the same).
The difference is that the number of JOBs waiting in the queue is used instead of the average waiting time in order to determine the execution delay of JOB and the start timing of staging. Hereinafter, the different points will be described.

  Therefore, in step 504, the sum of the number of JOBs waiting in the corresponding queue (the number of waiting JOBs) and the number of jobs being executed in the queue (the number of jobs being executed) (corresponding to k in FIG. 6). ). Further, the reciprocal of the average job execution frequency μ (see FIG. 6) is taken to calculate the average job execution time Te (average value of the job execution time).

In step 505, the threshold number (kth) of k for determining the delay of execution of JOB and the start timing of staging is calculated by queuing theory. kth is the probability (failure probability) that staging of the accessed file / volume cannot be completed before the start of JOB execution when the average JOB execution time and threshold time Tth obtained in step 504 and the threshold number are kth. And calculate using formulas derived from queuing theory. The upper limit value of the failure probability is designated in advance (stored in the memory 94). The threshold number (kth) of k is determined so as to be equal to or less than the upper limit value.
Then, the execution delay of JOB and the start timing of staging are determined in steps 507 and 509, respectively.

  Specifically, in step 507, using the job queue information table 80 shown in FIG. 8, the number k of jobs waiting in the corresponding queue and the sum k of the number of jobs being executed in the queue, The threshold numbers (kth) are compared (kth can be calculated from Te, Ti, and Tth by queuing theory). If k is large (No in step 507), the process proceeds to step 501. If k is equal to or less than kth (Yes in step 507), the process proceeds to step 508.

  In step 509, for all the queues in the corresponding queue, the sum k of the number of JOBs waiting in the corresponding queue and the number of executing jobs in the queue is compared with the threshold number (kth). . For comparison, for example, the JOB queue information table 2:80 is used. If k is larger than kth (No in step 509), the process proceeds to step 501. If k is equal to or less than kth (Yes in step 509), the process proceeds to step 510. If there is a JOB that satisfies this condition, it means that the execution of the corresponding JOB may be started at least at the kthth.

  If the processing is performed in this way, even if the sum k of the number of JOBs waiting in the corresponding queue and the number of jobs being executed in the queue is larger than the threshold number kth, the result is as described above. Until the sum k becomes substantially the same value as the threshold number kth, the processing after step 510 is suspended. Therefore, staging is executed at a much earlier stage than when the job is executed, and waste of power due to the use of the capacity of the first tier storage apparatus 11 for a long time can be saved.

  In this embodiment, it is possible to suppress the probability (failure probability) that the staging cannot be completed before the start of JOB execution to a value that is specified in advance by the administrator.

≪Others≫
Each of the above-described embodiments is suitable for carrying out the present invention, but the form of implementation is not limited to these, and various modifications can be made without departing from the scope of the present invention. It is.

  For example, the storage management server 19 may have a system configuration that also functions as the computer management server 18. Specifically, the storage management server 19 has functional units such as a JOB management unit 21, a user management unit 22, and an information providing unit 23, and the processor 95 executes processing so as to realize the functions of the functional units. May be.

  In addition, specific configurations of hardware, software, flowcharts, and the like can be appropriately changed without departing from the spirit of the present invention.

1 Computer System 2 Storage System 11 First Tier Storage Device (First Storage Device)
12 Second tier storage device (second storage device)
13 File server 14 Computer 15 LAN
16 IP switch 17 FC switch 18 Computer management server 19 Storage management server 21 JOB management unit 22 User management unit 23 Information provision unit 24 Information collection unit 25 Information analysis unit 26 Volume management unit 27 User area management unit 28 Storage management unit 42 High-speed hard disk (First hard disk device)
43 Large-capacity hard disk (second hard disk device)
51 First volume for file storage (first volume)
52 Second volume for storing files (second volume)
61 Virtual volume 94 for storing files Memory (storage unit)
95 Processor (control unit)
98 processor (control unit for computer management server)
99 memory

Claims (20)

A first storage device including one or more first hard disk devices constituting one or more first volumes;
A second storage device including one or more second hard disk devices constituting one or more second volumes;
A storage management server that manages the first volume and the second volume in which files accessed from the computer when executing a job sequentially executed on the computer;
In a storage system that is connected so that
The storage management server
At least a storage unit that stores, for each job, identification information of the second volume in which a job queue state and a file accessed from the computer when the job is executed are stored;
A control in which the job is to calculate the average waiting time on the basis of the average execution frequency when the jobs submitted to the average loading interval and the queue when it is submitted to the queue is executed,
The second hard disk device is put in an operating state to such an extent that the file input / output processing can be performed, and the time required for moving the file stored in the second volume to the first volume is thresholded. Control to calculate as value time,
The average waiting time is compared with the threshold time, and when the average waiting time is equal to or less than the threshold time, the execution of the job is delayed by at least the threshold time from the time when the job is submitted;
Before the job is executed, the second hard disk device that constitutes the second volume in which the file is stored is put in an operating state to the extent that the input / output processing of the file can be performed, and the file is moved to the first volume. And a control unit that executes the job on the computer while the file is stored in the first volume and the file is stored in the first volume. Wherein,
After pre-SL movement is completed, the storage system according to claim 1, characterized in that to perform the control to the second hard disk device in a non-operating state. The controller is
When the average waiting time is larger than the threshold time, the second hard disk constituting the second volume in which the file is stored until the average waiting time becomes substantially the same as the threshold time at the latest. Control the apparatus to be in an operating state to the extent that input / output processing of the file is possible, and to move the file to the first volume;
2. The storage system according to claim 1, wherein after the movement is completed, control is performed to place the second hard disk device in a non-operating state. The controller is
When it is confirmed that the execution of the job has been completed with reference to the queue status, the file input / output processing is performed on the second hard disk device that constitutes the second volume in which the file is originally stored. Control to move the file from the first volume to the second volume,
After the move is completed, the storage system according to any of claims 1 to 3, characterized by executing a control for the second hard disk device in a non-operating state, the. A computer management server that manages jobs executed sequentially on the computer is connected to be communicable,
The computer management server is
From the storage management server, until it receives a notification the movement of the file is complete, claims 1 to 3, characterized in that it comprises a computer management server control unit for executing control to delay the execution of the job The storage system according to any one of the above. All files that are not accessed from the computer are stored in the second volume,
The non-operating state of the second hard disk device constituting the second volume is that the power supply of the second hard disk device is turned off or the disk of the second hard disk device is stopped. The storage system according to any one of claims 1 to 4, wherein the storage system is in a state of being rotated at a rotation speed equal to or lower than a predetermined rotation speed. A file server that is communicably connected via the computer and the first storage device is communicably connected,
The first volume and the second volume are mounted on a directory managed by the file server,
The storage system according to any one of claims 1 to 4, wherein the storage unit stores the directory for each job. A first storage device including one or more first hard disk devices constituting one or more first volumes;
A second storage device including one or more second hard disk devices constituting one or more second volumes;
A storage management server that manages the first volume and the second volume in which files accessed from the computer when executing a job sequentially executed on the computer;
In a storage system that is connected so that
The storage management server
At least a storage unit that stores, for each job, identification information of the second volume in which a job queue state and a file accessed from the computer when the job is executed are stored;
Referring to the storage unit, the sum of the number of jobs whose queue status is waiting and the number of jobs being executed is calculated, and the job is determined based on the average execution frequency when the jobs submitted to the queue are executed. Control for calculating the average execution time when executed,
The second hard disk device is put in an operating state to such an extent that the file input / output processing can be performed, and the time required for moving the file stored in the second volume to the first volume is thresholded. Control to calculate as value time,
The probability that the movement cannot be finished, and the failure probability calculated based on the average execution time, the average insertion interval, and the threshold time is less than or equal to a value stored in the storage unit. Control to determine the threshold number;
A control that compares the sum with the threshold number, and delays execution of the job for at least the threshold time from when the job is submitted when the sum is less than or equal to the threshold number;
Before the job is executed, the second hard disk device that constitutes the second volume in which the file is stored is put in an operating state to the extent that the input / output processing of the file can be performed, and the file is moved to the first volume. And a control unit that executes the job on the computer while the file is stored in the first volume and the file is stored in the first volume. Wherein,
After pre-SL movement is completed, the storage system according to claim 8, characterized in that to perform the control to the second hard disk device in a non-operating state. The controller is
When the sum is larger than the threshold number, the second hard disk device that constitutes the second volume in which the file is stored until the sum becomes substantially the same as the threshold number at the latest. Control to move the file to the first volume, and move the file to the first volume so that file input / output processing is possible;
The storage system according to claim 8, wherein after the movement is completed, control is performed to place the second hard disk device in a non-operating state. A first tier storage apparatus including one or more high-speed hard disks constituting one or more file storage first volumes;
A second tier storage apparatus including one or more large-capacity hard disks constituting one or more file storage second volumes;
A storage management server for managing the file storage first volume and the file storage second volume in which files accessed from the computer when executing a job sequentially executed on the computer;
In a storage system that is connected so that
The storage management server
At least a job storing a job JOB state and identification information of the second volume for storing files in which a file accessed from the computer when executing the job is stored for each job;
A control in which the job is to calculate the average waiting time on the basis of the average execution frequency when the jobs submitted to the average loading interval and the queue when it is submitted to the queue is executed,
Time required for putting the large-capacity hard disk in an operating state to the extent that the file input / output processing is possible and staging the file stored in the file storage second volume to the file storage first volume a control for calculating a threshold value time,
The average waiting time and the threshold time are compared, and based on the comparison result, the job execution timing and the large-capacity hard disk constituting the file storage second volume in which the file is stored are operating. Or a control for adjusting the timing of making the non-operating state and executing the job on the computer in a state where the file is stored in the first volume for file storage,
The processor is
When the average waiting time is equal to or less than the threshold time, a control for delaying execution of the job by at least the threshold time from the time when the job is submitted;
The large-capacity hard disk that constitutes the second file storage volume in which the file is stored is brought into an operating state to the extent that the input / output processing of the file can be performed, and the file is staged on the first file storage volume Control to
After the staging is completed, control to put the large-capacity hard disk in a non-operating state, and
When the average waiting time is larger than the threshold time, the large capacity constituting the second volume for storing files in which the file is stored until the average waiting time becomes approximately the same as the threshold time at the latest Control the hard disk to be in an operating state to the extent that input / output processing of the file is possible, and staging the file to the first volume for storing the file;
After the staging is completed, control to put the large-capacity hard disk in a non-operating state, and
When it is confirmed that the execution of the job has been completed with reference to the JOB state, the input / output processing of the file is performed on the large-capacity hard disk constituting the second volume for file storage in which the file was originally stored. Control to bring the file into the operating state to the extent that it is possible and destage the file from the file storage first volume to the file storage second volume;
After the destaging is completed, a control for executing the large capacity hard disk in a non-operating state is executed. A first storage device including one or more first hard disk devices constituting one or more first volumes;
A second storage device including one or more second hard disk devices constituting one or more second volumes;
A storage management server that manages the first volume and the second volume in which files accessed from the computer when executing a job sequentially executed on the computer;
In a storage management method in a storage system that is communicably connected,
The storage unit of the storage management server is
At least the job queue status and the identification information of the second volume in which the file accessed from the computer when the job is executed are stored for each job,
The control unit of the storage management server
And processing the job to calculate the average waiting time on the basis of the average execution frequency when the jobs submitted to the average loading interval and the queue when it is submitted to the queue is executed,
The second hard disk device is put in an operating state to such an extent that the file input / output processing can be performed, and the time required for moving the file stored in the second volume to the first volume is thresholded. Processing to calculate the value time;
Comparing the average waiting time with the threshold time, and when the average waiting time is equal to or less than the threshold time, a process of delaying execution of the job by at least the threshold time from the time when the job is submitted,
Before the job is executed, the second hard disk device that constitutes the second volume in which the file is stored is put in an operating state to the extent that the input / output processing of the file can be performed, and the file is moved to the first volume. A storage management method comprising: moving to one volume and executing the job on the computer in a state where the file is stored in the first volume. Wherein,
After pre-SL movement is completed, the storage management method according to claim 12, characterized in that performing the processing of the second hard disk device in a non-operating state. The controller is
When the average waiting time is larger than the threshold time, the second hard disk constituting the second volume in which the file is stored until the average waiting time becomes substantially the same as the threshold time at the latest. A process for moving the file to the first volume by putting the device in an operating state to the extent that input / output processing of the file is possible; and
The storage management method according to claim 12, wherein after the movement is completed, a process of setting the second hard disk device to a non-operating state is executed. The controller is
When it is confirmed that the execution of the job has been completed with reference to the queue status, the file input / output processing is performed on the second hard disk device that constitutes the second volume in which the file is originally stored. A process for moving the file from the first volume to the second volume,
The storage management method according to any one of claims 12 to 14, wherein after the movement is completed, a process of setting the second hard disk device to a non-operating state is executed. A computer management server that manages jobs executed sequentially on the computer is connected to be communicable,
The computer management server control unit of the computer management server includes:
The storage management according to any one of claims 12 to 14, wherein processing for delaying execution of the job is executed until a notification of completion of the movement of the file is received from the storage management server. Method. All files that are not accessed from the computer are stored in the second volume,
The non-operating state of the second hard disk device constituting the second volume is that the power supply of the second hard disk device is turned off or the disk of the second hard disk device is stopped. The storage management method according to any one of claims 12 to 15, wherein the storage management method is in a state of being rotated at a rotation speed equal to or lower than a predetermined rotation speed. A first storage device including one or more first hard disk devices constituting one or more first volumes;
A second storage device including one or more second hard disk devices constituting one or more second volumes;
A storage management server that manages the first volume and the second volume in which files accessed from the computer when executing a job sequentially executed on the computer;
In a storage management method in a storage system that is communicably connected,
The storage unit of the storage management server is
At least the job queue status and the identification information of the second volume in which the file accessed from the computer when the job is executed are stored for each job,
The control unit of the storage management server
Referring to the storage unit, the sum of the number of jobs whose queue status is waiting and the number of jobs being executed is calculated, and the job is executed based on the average execution frequency when the job submitted to the queue is executed. Processing to calculate the average execution time when
The second hard disk device is put in an operating state to such an extent that the file input / output processing can be performed, and the time required for moving the file stored in the second volume to the first volume is thresholded. Processing to calculate the value time;
The probability that the movement cannot be finished, and the failure probability calculated based on the average execution time, the average insertion interval, and the threshold time is less than or equal to a value stored in the storage unit. Processing to determine the threshold number;
Comparing the sum with the threshold number, and when the sum is less than or equal to the threshold number, a process of delaying execution of the job for at least the threshold time from when the job was submitted,
Before the job is executed, the second hard disk device that constitutes the second volume in which the file is stored is put in an operating state to the extent that the input / output processing of the file can be performed, and the file is moved to the first volume. A storage management method comprising: moving to one volume and executing the job on the computer in a state where the file is stored in the first volume. Wherein,
After pre-SL movement is completed, the storage management method according to claim 18, characterized in that performing the processing of the second hard disk device in a non-operating state. The controller is
When the sum is larger than the threshold number, the second hard disk device that constitutes the second volume in which the file is stored until the sum becomes substantially the same as the threshold number at the latest. A process for moving the file to the first volume, and an operation state that enables file input / output processing; and
The storage management method according to claim 18, further comprising: executing a process of bringing the second hard disk device into a non-operating state after the movement is completed.

Images