JP5206103B2 - Storage device, storage device control system, storage device control method, and program - Google Patents

Description

  The present invention relates to a storage apparatus, a storage apparatus control system, a storage apparatus control method, and a program.

  In recent years, interest in power saving has increased due to global environmental problems and increasing global power demand. Especially in large data centers, power costs are a major factor in spending. Reducing power consumption not only reduces the power of the apparatus, but also reduces the cost of cooling equipment. Storage devices are loaded with a large number of hard disks, and the power is said to account for about 30% of the entire data center. Against this background, power saving is becoming a major factor in storage devices as well as conventional performance, reliability, and cost.

  As for power saving of storage devices, MAID (Massive Array of Inactive Disks) is known. Moreover, the proposal regarding power saving is disclosed by patent documents 1-3.

  The disk array device of Patent Document 1 is a disk array device that is connected to a host device and transmits / receives information to / from the host device. A function of associating a magnetic disk device having a first power saving mode to be suppressed and a second power saving mode to reduce power consumption by controlling the rotation of the spindle, and a set of magnetic disk devices and access from a host device And a control unit having a function of counting a time from an access from a host device to the next access, and a function of specifying one magnetic disk device and executing a power saving mode of the magnetic disk device. The magnetic disk device is set to the second power saving mode when there is no access to the set for a predetermined time.

  That is, Patent Document 1 discloses a method of having a plurality of power saving modes, putting a disk in a power saving mode if there is no access for a predetermined time, and diagnosing when a certain time or a specified time has passed in order to maintain reliability. ing.

  According to this disk array device, the power consumption of the magnetic disk device mounted in the disk array device can be suppressed, and the magnetic disk device that is not accessed by the host device is targeted. It is said that energy saving is possible while suppressing performance degradation.

The storage device system of Patent Document 2 is a storage device system connected to a computer, and the storage device system has a plurality of logical units each composed of a disk device, and each logical unit from the computer is converted into a logical unit. An instruction to turn on / off the power of the corresponding disk device is received, and based on the instruction, the power of the disk device corresponding to the logical unit is turned on / off independently of the disk device corresponding to the other logical unit. To do.
That is, Patent Document 2 discloses a method for controlling a power saving mode from a host computer.
According to this storage device system, only the disk device that is reading / writing data is operated, and the disk device that is not reading / writing data is stopped, thereby extending the operation period of the disk array. The power consumption of the disk array can be reduced.
JP 2000-293314 A JP 2005-157710 A

  However, in the technique described in Patent Document 1, when there is randomness in access to the storage, the condition that there is no access for a predetermined time does not readily hold for the entire disk, and the power saving mode is not established. In addition, if the predetermined time is shortened to facilitate the transition to the power saving mode, access is often made immediately after entering the power saving mode, which not only greatly deteriorates the performance, but also frequently switches the power saving mode. Conversely, power consumption increases.

  In the technique described in Patent Document 2, a program running on the host computer must be compatible, and there is a limitation in versatility.

  As a general method, there is a method in which unaccessed data is moved to an archive storage and a pointer is placed at the original location (hierarchical storage management). Since the archive storage contains data that is not accessed, it can be put into power saving mode.

  However, this method requires complicated policy control such as extracting unaccessed data, rearranging the data in the entire storage, and determining which data to migrate. Moreover, since it is necessary to clearly separate the storage for access and the storage for archive, there is a drawback in terms of flexibility. For example, in an environment where most data is accessed, there is a lot of free space in the archive storage, but there is not enough storage for access. In that case, it is only necessary to be able to easily change from archive to access, but changing the setting requires a number of procedures. The same applies to the change from access to archive.

  Therefore, an object of the present invention is to provide a storage device, a storage device control system, a storage device control method, and a program that can enhance the effect of the power saving mode and at the same time flexibly change the storage to be set in the power saving mode.

  The device of the present invention is a storage device comprising a plurality of disk devices, a management storage area, and each disk device and a control device that controls the storage area, wherein the control device is written In addition to performing additional writing in which data is written in order from the address with the smallest address, by appending not only updated data but also read data, unaccessed data has the smaller address of the disk device. The disk device that collects the unaccessed data is set in a power saving mode.

  The system of the present invention includes a storage device having a plurality of disk devices, a storage area for management, a control device for controlling each of the disk devices and the storage area, and reading / writing data to the disk device. A storage device control system comprising a computer to perform and a network connecting the storage device and the computer, wherein the control device performs additional writing in which written data is written in order from a small address In addition, by adding not only the updated data but also the read data, the unaccessed data is set to the smaller address of the disk device, and the disk device collecting the unaccessed data is provided. It is characterized by being in power saving mode.

  The control method of the present invention is a control method for a storage device that controls a plurality of disk devices and a management storage area, in addition to performing additional writing in which written data is written in order from the address with the smallest address. Thus, by adding not only the updated data but also the read data, the unaccessed data is set to the smaller address of the disk device, and the disk device collecting the unaccessed data is set in the power saving mode. It is characterized by doing.

  The program of the present invention is a program for causing a computer to execute processing for controlling a plurality of disk devices and a management storage area, and the written data is written to the computer in order from the address having the smallest address. A disk that collects the unaccessed data by making the inaccessible data harden to the smaller address of the disk device by executing the additional recording process and the process of additionally writing the read data as well as the updated data It is characterized by putting the device into a power saving mode.

  According to the present invention, since only the data that is not truly accessed is collected, the effect of the power saving mode is enhanced, and at the same time, the storage in the power saving mode can be flexibly changed. Can be realized.

  In the present invention, write-once data management is used. In the existing write-once type, only updated data is added at the end, but in the present invention, the accessed data is also copied to the last disk. By doing so, the earlier disk contains only data that is not accessed, so that the power saving mode can be set.

  An embodiment of a storage device according to the present invention is a storage device comprising a plurality of disk devices, a management storage area, and a control device that controls each disk device and storage area. In addition to performing additional writing in which written data is written in order starting from the address with the smallest address, by adding not only updated data but also read data, unaccessed data is addressed to the disk device. The disk device that collects data that is not accessed is put into a power saving mode.

  According to the above configuration, by collecting only data that is not truly accessed, the effect of the power saving mode is enhanced, and at the same time, the storage to be in the power saving mode can be flexibly changed. Can be realized.

  In another embodiment of the storage apparatus according to the present invention, in addition to the above configuration, the control apparatus compares the last access time of the read data with the current time, and appends it when it is longer than the predetermined time. It is characterized by doing.

  In another embodiment of the storage apparatus according to the present invention, in addition to the above configuration, the control apparatus additionally writes when the length of data from the read data to the last valid data is longer than a predetermined length. It is characterized by that.

  In another embodiment of the storage apparatus according to the present invention, in addition to the above configuration, the control apparatus may have a length of data from the read data to the last valid data longer than a predetermined ratio of all valid data. It is characterized by adding to the case.

  In another embodiment of the storage device according to the present invention, in addition to the above configuration, when the data added by reading is overwritten, the control device adds the address added by reading instead of adding data. It is characterized by overwriting.

  An embodiment of a storage device control system according to the present invention includes a storage device having a plurality of disk devices, a storage area for management, a control device that controls each disk device and the storage area, and a disk device A storage device control system including a computer that reads / writes data and a network that connects the storage device and the computer, and the control device writes the written data in order from the address with the smallest address In addition to performing additional recording, not only updated data but also read data is additionally recorded so that unaccessed data is consolidated in the smaller address of the disk device, and the disk device that collects unaccessed data The power saving mode is set.

  According to the above configuration, by collecting only data that is not truly accessed, the effect of the power saving mode is enhanced, and at the same time, the storage to be in the power saving mode can be flexibly changed. Can be realized.

  An embodiment of a storage device control method according to the present invention is a storage device control method for controlling a plurality of disk devices and a storage area for management, in which written data is written in order from the address with the smallest address. In addition to adding data to be written, in addition to updated data, data that has been read is additionally written so that unaccessed data is consolidated in the smaller address of the disk device, and unaccessed data is collected. The disk device is set in a power saving mode.

  According to the above configuration, by collecting only data that is not truly accessed, the effect of the power saving mode is enhanced, and at the same time, the storage to be in the power saving mode can be flexibly changed. Can be realized.

  In another embodiment of the storage apparatus control method according to the present invention, in addition to the above configuration, the last access time of the read data is compared with the current time, and added when it is longer than the predetermined time. It is characterized by that.

  In another embodiment of the storage apparatus control method according to the present invention, in addition to the above configuration, additional writing is performed when the length of data from the read data to the last valid data is longer than a predetermined length. It is characterized by.

  In another embodiment of the storage apparatus control method according to the present invention, in addition to the above configuration, the length of data from the read data to the last valid data is longer than a predetermined ratio of all valid data It is characterized by being added to.

  In another embodiment of the storage apparatus control method according to the present invention, in addition to the above configuration, when data added by reading is overwritten, the address added by reading instead of appending is added. On the other hand, it is overwritten.

<Program and storage medium>
The storage apparatus of the present invention described above is realized by a program that causes a computer to execute processing. Examples of the computer include general-purpose computers such as personal computers and workstations, but the present invention is not limited to this.

  An embodiment of a program according to the present invention is a program for causing a computer to execute a process for controlling a plurality of disk devices and a management storage area. By executing the appending process that writes in order, the process that appends not only the updated data but also the read data, the data that is not accessed is fixed to the smaller address of the disk device, and the data that is not accessed is The collected disk device is set in a power saving mode.

  According to the above configuration, by collecting only data that is not truly accessed, the effect of the power saving mode is enhanced, and at the same time, the storage to be in the power saving mode can be flexibly changed. Can be realized.

  In another embodiment of the program according to the present invention, in addition to the above configuration, when the computer compares the last access time of the read data with the current time, the result of the comparison is longer than a predetermined time It is characterized in that a process for appending is executed.

  In another embodiment of the program according to the present invention, in addition to the above configuration, the computer determines whether the length of the data from the read data to the last valid data is longer than a predetermined length. Processing is characterized in that additional processing is executed when the length of data from the read data to the last valid data is longer than a predetermined length.

  In another embodiment of the program according to the present invention, in addition to the above configuration, the computer can determine whether the length of data from the read data to the last valid data is longer than a predetermined ratio of all valid data. A process for determining whether or not the length of the data from the read data to the last valid data is longer than a predetermined ratio of all valid data.

  In another embodiment of the program according to the present invention, in addition to the above-described configuration, the computer determines whether or not the data added by reading is overwritten, and the data added by reading is overwritten. In this case, it is characterized in that the address added is overwritten by being read instead of being additionally written.

  Thus, the storage apparatus and the storage apparatus control system of the present invention can be realized anywhere as long as there is a computer environment in which the program can be executed.

Such a program may be stored in a computer-readable storage medium.
Here, examples of the storage medium include computer-readable storage media such as a CD-ROM (Compact Disc Read Only Memory), a flexible disk (FD), a CD-R (CD Recordable), and a DVD (Digital Versatile Disk). Semiconductor memories such as HDD (Hard Disc Drive), flash memory, RAM (Random Access Memory), ROM (Read Only Memory), and FeRAM (ferroelectric memory).

  The above-described embodiment shows an example of a preferred embodiment of the present invention, and the present invention is not limited thereto, and various modifications can be made without departing from the scope of the invention. is there.

Next, an embodiment of a storage apparatus according to the present invention will be described with reference to the drawings.
FIG. 2 is a conceptual diagram showing an embodiment of the storage apparatus control system according to the present invention.
The system shown in FIG. 1 has at least one storage node 1, and the storage node 1 has a plurality of disk devices 2. The storage node 1 is connected to the computer 3 via the network 4, and the computer 3 designates an address via the network 4 and performs READ / WRITE (read / write) to the storage node 1. The address may be a block address or a file path name or ID.

An example of the internal configuration of the storage node 1 as a storage apparatus according to the present invention is shown in FIG.
The storage node 1 shown in FIG. 1 includes a disk device 2, a management area 5, and a CPU (Central Processing Unit) 9 that controls the disk device 2 and the management area 5.

  Real addresses are assigned to a plurality of disk devices 2 in series. In FIG. 1, addresses 0 to 499 are assigned to the first disk device 2, and real addresses are assigned to the second disk device from the next 500 addresses. The storage node 1 has a management area 5 separately from data storage.

  The management area 5 is a management storage area and is usually stored in the disk device 2 so that it is not lost even when the power is turned off. Since it is necessary to access at high speed during operation, it is possible to prevent a performance bottleneck by caching in memory. The management area 5 includes an address conversion map 6, a real address state map 7, and a tail pointer 8. These will be described below.

  Write-once data management is a method in which updated data is always written at the end of valid data. Since the address that is visible to the outside of the storage node 1 is different from the internal address, it is necessary to convert the address used with the computer 3 into an address within the storage node 1. The map for this is the address conversion map 6.

  When this address conversion map 6 is used and the computer 3 overwrites the data, the data is converted into an append to the end in the storage node 1. The tail pointer 8 (Tail Pointer) records the tail end. In addition, in the write-once type, data is added at the end by update, so the original data is unnecessary (Garbage). The real address state map 7 records the state of whether the data of the real address is in use (Used) or not needed (Garbage).

FIG. 3 is a diagram showing an example of the operation of the storage node 1 when the data at the virtual address 100 in the storage apparatus shown in FIG. 1 is updated.
According to the address conversion map 6, the virtual address 100 is converted into a real address 20. From this state, the computer 3 updates the data at the virtual address 100. In the storage node 1, since the data has been updated, it is added at the end. With reference to the tail pointer 8, the tail pointer 8 is added to the last real address 150, and the tail pointer 8 is changed to 160.

  In the address translation map 6, the real address of the virtual address 100 is changed from 20 to 150. In the real address state map 7, the real address 20 is changed from Used (used) to Garbage (unnecessary data), and the newly added real address 150 is changed from Unused (unused) to Used (used). Is done.

As described above, normal write-once data management adds only updated (WRITE) data to the end, but in this embodiment, all read (READ) data is also added to the end. To do.
In other words, in this embodiment, in addition to performing additional writing in which written data is written in order from a small address, not only updated data but also read data is additionally accessed, so that access is not possible. It is characterized in that the data is set to the smaller one of the addresses of the disk device, and the disk device collecting the data that is not accessed is set in the power saving mode.
According to this method, not only updated data but also accessed data are gathered behind the disk device 2. In this embodiment, since the storage node 1 has a plurality of disk devices 2 and the addresses are serially connected internally, the data to be accessed is gathered in the disk device 2 with the address behind, and the address is the disk with the front address. The device 2 collects only the data whose access time is old and the data whose copy is behind because of the access. Therefore, the disk device 2 with the front address can be set in the power saving mode.

  According to the method of this embodiment, it is possible to create a situation in which data with a low possibility of being accessed is gathered in the disk device 2 whose address is ahead, regardless of how the access is random. Data that has not been accessed for a sufficiently long time is unlikely to be accessed in the future, and it is unlikely that it will be accessed immediately as a power saving mode.

  Further, unlike the conventional hierarchical storage management, this embodiment can be flexibly operated without fixing the disk device 2 to be in the power saving mode. If there is a lot of data that is not accessed, a large number of disk devices 2 may be set to the power saving mode, and if the whole is accessed uniformly, the number of disk devices 2 to be set in the power saving mode may be reduced. The number of power saving modes can be determined based on when the last data in the disk device 2 was accessed, and can be easily determined without collecting and sorting the access times of all data. I can judge. The access time is stored in the real address state map 7, and the access time of the data whose real address is the last and used can be referred to.

  The disk device 2 with the front address includes not only unaccessed data but also unnecessary data (Garbage). When there is a lot of unnecessary data, capacity efficiency deteriorates. This problem can generally be solved by an operation called garbage collection.

FIG. 4 is a diagram showing the garbage collection operation of the storage apparatus according to the present invention.
As shown in FIG. 4, the disk device A (101) and the disk device B (102) contain mixed data of Used and Garbage. Of these, only Used data is collected and newly copied to the disk device C (103). When this operation is completed, the disk device A (101) and the disk device B (102) become new free disks, and the capacity efficiency can be increased.

  This method has an overhead that a process of adding information at the end of each access occurs. A device for reducing the overhead as much as possible will be described.

  First, when data is read when accessed, it is determined by using a certain standard, not by appending in all cases. The criterion may be the elapsed time from the last access, but it is more efficient to use the difference between the data address and the last address (tail pointer 8) or the ratio to the whole.

  In other words, even if a considerable amount of time has passed since the last access, there is no need to add additional storage if there is no active access in the first place. Moreover, even if the elapsed time is short, if there is active access and the data is old as viewed from the whole, it is based on the idea that it should be added at the end. In order to make this determination, the access time is stored in the real address state map 7. This method can reduce the overhead of additional recording processing.

  Further, when the read data is updated, the update may be overwritten on the data added at the time of reading, instead of newly adding the update at the end. The determination of overwriting can use the elapsed time or the difference from the tail pointer 8 in the same manner as described above. By overwriting, unnecessary data does not increase and the load of garbage collection can be reduced.

FIG. 5 is a diagram showing an example of a flowchart of the storage apparatus control method according to the present invention.
The flowchart shown in the figure is a flowchart when a READ request to the virtual address X is received. First, conversion from a virtual address to a real address is performed using the address conversion map 6. Thereby, the real address Y is obtained (S1).

Next, the real address state map 7 is drawn with the real address Y to obtain the access time Ty (S2).
The real address Y and the access time Ty obtained by these processes are compared with the real address Z and the access time Tz of the tail pointer (S3).

Various methods are conceivable for the comparison formula, but here, the ratio of the real addresses to the whole and the age of the access time are used. Y / Z <0.7 means that the real address is smaller than 0.7 times the entire address. In other words, it means that about 30% of the data has been accessed since this last address was accessed. (To be exact, there may be duplicate data in 30%. There are few possibilities). Tz-Ty> 1week means that the terminal has not been accessed for more than one week. If & of these two conditional expressions is taken, it can be determined whether the data is somewhat old.
When the conditional expression is satisfied (S4 / Y), the data is appended at the end. The data is copied at the end, and the address translation map 6, the real address state map 7, and the tail pointer 8 are updated (S5).

When the condition is not satisfied (S4 / N), the data is not added. Only the access time of the real address state map 7 is updated (S6).
In the case of a WRITE request, if additional writing is unconditional, the process of S5 is performed without the determination of S4. However, even in the case of WRITE, it may not be added based on the determination in S4. In the case of WRITE, there is another advantage that generation management can be performed if unconditional appending is performed.

It is a figure which shows an example of the internal structure of the storage node 1 as a storage apparatus concerning this invention. It is a conceptual diagram which shows one Example of the storage apparatus control system which concerns on this invention. FIG. 3 is a diagram illustrating an example of an operation of the storage node 1 when data at a virtual address 100 is updated in the storage device illustrated in FIG. 1. It is the figure which showed the operation | movement of the garbage collection of the storage apparatus concerning this invention. It is a figure which shows an example of the flowchart of the control method of the storage apparatus which concerns on this invention.

Explanation of symbols

1 Storage node (storage device)
2 Disk device 3 Computer 5 Management area 6 Address translation map 7 Real address status map 8 Tail pointer 9 CPU

Claims (8)

A plurality of disk devices to which addresses are continuously assigned ; and
The written-out write-free data to the plurality of disk devices, in addition to performing additional recording to and writes in order from lowest address, by also additional data read from said plurality of disk devices, are not accessed A controller that sets the data to a smaller one of the addresses of the plurality of disk devices, and sets the disk device collecting the unaccessed data to a power saving mode ;
With
The control device is configured such that the address of the read data is smaller than a predetermined ratio of the last address added and the time added at the end and the read time of the read data are The storage device that additionally writes the read data when the difference is longer than a predetermined time . A plurality of disk devices to which addresses are continuously assigned; and
In addition to performing additional writing in which data to be written to the plurality of disk devices is written in order from an address having a smaller address, by additionally writing data read from the plurality of disk devices, data that is not accessed can be A controller that sets the address of the plurality of disk devices to a smaller one, and sets the disk devices that collect the unaccessed data to a power saving mode; and
With
Wherein the control device, if the read data is updated, overwriting the updated data, the data recordable by being Desa read
Storage devices.   A plurality of disk devices to which addresses are continuously assigned; and
  In addition to performing additional writing in which data to be written to the plurality of disk devices is written in order from an address having a smaller address, by additionally writing data read from the plurality of disk devices, data that is not accessed can be A controller that sets the address of the plurality of disk devices to a smaller one, and sets the disk devices that collect the unaccessed data to a power saving mode; and
A storage device having
  A computer for reading / writing data to / from the disk device;
  A network connecting the storage device and the computer;
With
  The control device is configured such that the address of the read data is smaller than a predetermined ratio of the last address added and the time added at the end and the read time of the read data are If the difference is longer than the predetermined time, append the read data
Storage device control system.   A plurality of disk devices to which addresses are continuously assigned; and
  In addition to performing additional writing in which data to be written to the plurality of disk devices is written in order from an address having a smaller address, by additionally writing data read from the plurality of disk devices, data that is not accessed can be A controller that sets the address of the plurality of disk devices to a smaller one, and sets the disk devices that collect the unaccessed data to a power saving mode; and
A storage device having
  A computer for reading / writing data to / from the disk device;
  A network connecting the storage device and the computer;
With
  When the read data is updated, the control device overwrites the updated data by reading the updated data.
Storage device control system.   By additionally writing the data read from the plurality of disk devices in addition to performing the additional writing of sequentially writing the data to be written to the plurality of disk devices assigned addresses in order from the address having the smallest address. A method for controlling a storage device, in which unaccessed data is consolidated to a smaller address of the plurality of disk devices, and the disk devices that collect the unaccessed data are placed in a power saving mode,
  The address of the read data is smaller than a predetermined ratio of the last address added, and the difference between the last time added and the read time of the read data is greater than a predetermined time. Append the read data if it is long
Storage device control method.   By additionally writing the data read from the plurality of disk devices in addition to performing the additional writing of sequentially writing the data to be written to the plurality of disk devices assigned addresses in order from the address having the smallest address. A method for controlling a storage device, in which unaccessed data is consolidated to a smaller address of the plurality of disk devices, and the disk devices that collect the unaccessed data are placed in a power saving mode,
  When the read data is updated, the updated data is overwritten on the added data by being read.
Storage device control method.   On the computer,
  By additionally writing the data read from the plurality of disk devices in addition to performing the additional writing of sequentially writing the data to be written to the plurality of disk devices assigned addresses in order from the address having the smallest address. , A program for causing non-accessed data to be consolidated to a smaller address of the plurality of disk devices, and for executing a process of putting the disk devices collecting the non-accessed data into a power saving mode,
  The address of the read data is smaller than a predetermined ratio of the last address added, and the difference between the last time added and the read time of the read data is greater than a predetermined time. If it is long, the process of appending the read data is executed.
program.   On the computer,
  By additionally writing the data read from the plurality of disk devices in addition to performing the additional writing of sequentially writing the data to be written to the plurality of disk devices assigned addresses in order from the address having the smallest address. , A program for causing non-accessed data to be consolidated to a smaller address of the plurality of disk devices, and for executing a process of putting the disk devices collecting the non-accessed data into a power saving mode,
  When the read data is updated, a process of overwriting the updated data with the added data by executing the read data is executed.
program.

Images