JP3190546B2 - Block address conversion method, rotation type storage subsystem control method, and disk subsystem - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a block address conversion technique and a rotary storage subsystem control technique.
In particular, the present invention relates to a technique that is effective when applied to the control of a rotary storage subsystem that uses a rotary storage medium such as a magnetic disk divided into a plurality of logical volumes.

[0002]

2. Description of the Related Art In recent years, in magnetic disk subsystems, for example, Nikkei BP, published on April 26, 1993, "Nikkei Electronics April 26," P77-P.
103, etc., a small and inexpensive magnetic disk is arranged in an array to provide high-speed, large-capacity,
RAID (Redundant Array of Inex) realizing high reliability
pensive Disks) technology.

In a magnetic disk subsystem of this type, a control device for controlling the entire magnetic disk subsystem divides continuous addresses on a magnetic disk into logical volume units to realize a plurality of logical volume configurations. I have.

In general, information having a common meaning between logical volumes is stored in the same relative position of each logical volume. However, when a magnetic disk is divided into logical volumes, the magnetic disk is divided into logical volumes. Since the data blocks arranged in the ascending order of the addresses are simply divided by the capacity of the logical volume, the information having the same relative position in each logical volume exists at a position separated by the number of data blocks of the logical volume capacity. .

[0005]

In a computer system comprising a plurality of logical volumes, a typical benchmark test for evaluating the performance of a disk subsystem is executed, and the status of a read / write command to a magnetic disk is analyzed. Then, it can be understood that the information at the same relative position in each logical volume of the magnetic disk subsystem (information having a common meaning in each logical volume, etc.) is often accessed over a plurality of logical volumes. ing. This means that when a general job is executed in a computer system comprising a plurality of logical volumes, a read / write command to a magnetic disk does not access one logical volume sequentially, but extends over a plurality of logical volumes. Access to the Internet.

On the other hand, in a conventional magnetic disk subsystem in which one magnetic disk is divided into a plurality of logical volumes and used, a method is employed in which data blocks arranged in ascending order of the magnetic disk are simply divided by the capacity of the logical volume. It is common to do. For this reason, the information at the same relative position of each logical volume exists at a physical position on the magnetic disk at a distance, and when a general job is executed, the job performance is poor because the moving time of the head is long. There was a problem.

An object of the present invention is to divide a physical volume into a plurality of logical volumes and divide each logical volume into a plurality of unit areas to use a block address capable of improving access performance in a storage subsystem. To provide a conversion technology.

Another object of the present invention is to improve the access performance in a rotary storage subsystem which divides a rotary storage medium into a plurality of logical volumes and divides each logical volume into a plurality of unit areas for use. It is an object of the present invention to provide a control technology of a rotary storage subsystem that can perform the control. Another object of the present invention is to provide a disk subsystem which divides a rotary storage medium into a plurality of logical volumes and divides each logical volume into a plurality of unit areas to use the same relative position in each logical volume. An object of the present invention is to realize high-speed access when certain information is accessed for a plurality of logical volumes.

[0009]

According to the present invention, each of a plurality of logical volumes set on a physical volume is divided into certain units, and each logical volume is divided into units at the same relative position. The block address is converted so as to be located at a close position on the physical volume.

When the block address arrangement conversion according to the present invention is applied to a rotary storage subsystem, the following two means are provided.

In the first means, a subsystem control device for controlling the entire rotary storage subsystem recognizes the block address of the rotary storage medium and stores the same relative position information in each logical volume in the rotary storage medium. The address arrangement is converted and accessed so as to be located close to the recording medium. The parameter of how to convert the address arrangement is determined in advance by the subsystem control device according to the use condition of the host before the start of system operation, and the address of the data block added in ascending order on the rotary storage medium Is left as it is.

When the subsystem controller writes and reads data, the address of the data block of the rotary storage medium is changed so that the information having the same relative position in each logical volume is close to the rotary storage medium. Convert.

In the second means, a single rotary storage medium control device converts the address arrangement. The subsystem control device determines the parameter value of the address translation before starting the system operation, according to the usage status of the host.

The subsystem control device controls the number of logical volumes (N; Number), the capacity of logical volumes (C; Capacity), and the logical volume for the rotary storage medium control device that controls a single rotary storage medium. The size (S; Size) of the unit into which the data is divided is transmitted as a parameter, and an instruction for address layout conversion is issued.

This means also keeps the address of the data block added in ascending order on the rotary storage medium. In the rotary storage medium control device that has received the parameters, the block addresses are arranged so that the units (S) having the same relative position in each logical volume are grouped into one group and exist close to the rotary storage medium. Have the logic to convert

[0016]

When the information at the same relative position in each logical volume is accessed by a plurality of logical volumes by each of these means, the physical movement distance of the head that moves on the rotary storage medium and accesses an arbitrary area. Can be shortened, and high-speed access becomes possible.

In general, information having the same meaning in each logical volume is recorded at the same relative position of each logical volume. Even in a process of updating a plurality of logical volumes, the same logical volume has the same information. The information stored in the relative position is often updated. By incorporating logic for converting the arrangement of data block addresses of a rotary storage medium so that information having the same relative position in each logical volume is close to the rotary storage medium, information can be transferred across a plurality of logical volumes. In the access processing, the physical movement distance of the head is reduced, and the processing performance is improved.

Thus, it is possible to surely improve the general job performance of accessing each information existing at the same relative position in each logical volume.

[0019]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a conceptual diagram showing an example of the configuration of a rotary storage subsystem in which a block address conversion method according to an embodiment of the present invention is implemented.
FIG. 1 is a conceptual diagram showing an example of a configuration of a conventional rotary storage subsystem. In the present embodiment, a case where the present invention is applied to a magnetic disk subsystem using a magnetic disk as a medium will be described as an example of a rotary storage subsystem.

In FIG. 1, 10 is a host computer, 20 is a disk subsystem, 21 is a subsystem control device for controlling the entire disk subsystem, 22 is a disk control device for controlling a single disk, and 23 is a magnetic disk drive , 30 are interface cables for connecting the host computer and the subsystem controller, 4
Reference numeral 0 denotes an interface cable for connecting the subsystem controller and the disk controller, and reference numeral 50 denotes an interface cable for connecting the controller, the disk controller, and the magnetic disk drive. Reference numeral 24 denotes an address transition table showing an example of how an address is converted based on a conversion parameter.

In the case of this embodiment, as an example, the physical storage area of the magnetic disk drive 23 is logically divided into three logical volumes 0 to 2 and managed, and each logical volume further includes: It is divided into five zones 0 to 4, and within each zone, ten data blocks are allocated. Therefore, in the above logical volume, 150
There are data blocks.

Each of the 150 data blocks in all the logical volumes is assigned a logical block address LBA (0 to 149) for uniquely identifying each data block.

On the other hand, the magnetic disk area in the magnetic disk drive 23 is divided into 150 data blocks of the same size as the data blocks in the logical volume.
BA is given. This physical block address PB
A is set in ascending or descending order in a seek direction of a head (not shown) with respect to a magnetic disk as a storage medium and in a rotating direction of the magnetic disk. Data blocks having adjacent physical block addresses PBA are physically located on the magnetic disk. Are also next to each other.

In the case of this embodiment, the subsystem control device 2
1 is provided with an address conversion logic 21a, performs a conversion operation as exemplified in the address transition table 24 in FIG. 1, and responds to an input / output request of data issued from the host computer 10 at a higher level. Issue a data input / output instruction to the subordinate disk controller 22.

First, FIG. 7 shows a drawback of the conventional system in which a physical volume is simply divided into a plurality of logical volumes. Parts corresponding to those in the configuration of the present embodiment are denoted by the same reference numerals.

The magnetic disk drive 23 divides the recording surface equally in units of data blocks, and each data block is provided with a data block address in ascending address order. The subsystem controller 21 specifies a data block address to the disk controller 22 and
Reads and writes data.

Assume that the total number of data blocks of the magnetic disk drive 23 is 150, the size of the logical volume is 50, and one magnetic disk drive is 3
It is assumed that the volume is equally divided into two logical volumes. That is, logical volume 0 has data block addresses 0 to 4
9. Logical volume 1 has data block addresses 50 to
99, logical volume 2 has data block address 10
0 to 149, respectively, and the subsystem controller 21 is conscious of the relationship between the data block address of the magnetic disk drive 23 and the logical volume.

Now, assuming that an access request from the host computer 10 accesses one data block at the head of each logical volume, the subsystem controller 21 sends a data block address 0, 50 and 100 access requests will be made. The disk control device 22 specifies the specified 3
The head is moved to access one data block, but each data block at the same position in each logical volume is almost the same as the number of data blocks that make up each logical volume on the actual magnetic disk. Since they are physically separated from each other, it is inevitable that the moving distance of the head, that is, the time required for moving the head becomes longer, and the access performance deteriorates.

On the other hand, in the present embodiment, the problem of the prior art as illustrated in FIG. 7 described above is solved by the subsystem controller 21 converting the address arrangement as follows.

That is, the subsystem control device 21
The address of the read / write request for each of the logical volumes 0 to 2 received from the host computer 10 is converted into the actual address of the magnetic disk drive 23. At this time, the block address arrangement conversion is performed. Number of logical volumes (N), logical volume capacity (C), size of logical volume division unit (S), which are conversion parameters for block address arrangement conversion
Are defined and held by the subsystem control device 21.

Now, it is assumed that an access request for one data block at the same relative position of each logical volume shown below has been issued from the host computer 10 to the subsystem controller 21.

Data block address 0 (logical block address L) which is information of relative No 0 of logical volume 0
BA = 0) Data block address 50 (logical block address LBA = 5) which is information of relative No0 of logical volume 1
0) Data block address 100 (logical block address LBA = 1) which is information of relative No0 of logical volume 2
00) The subsystem controller 21 performs address conversion on the address of the access request received from the host computer 10 based on the above-described conversion parameters as follows.

Data block address 0 (LBA =
0) → conversion → data block address 0 on disk (physical block address PBA = 0) data block address 50 (LBA = 50) →
Conversion → Data block address 10 on disk (physical block address PBA = 10) Data block address 100 (LBA = 100)
→ Conversion → Data block address 20 on disk
(Physical block address PBA = 20) This conversion can be calculated by the following equations (1) to (3) based on the above-described N, C, and S conversion parameters. That is, the logical volume No (integer) is LV,
Zone No. (relative No. in logical volume: integer) is Z
Then, LV = LBA / C... Remainder α (1) Z = α / S... Remainder β (2) PBA = Z × S × N + S × LV + β (3) The corresponding physical block address PBA can be uniquely determined. For example, if LBA = 100, LV =
2, α = 0, Z = 0, β = 0, and PBA = 0 × 10
× 3 + 10 × 2 + 0 = 20.

Thus, the individual logical volumes 0 to 2
Have the same relative position, LBA = 0, LBA = 5
0, LBA = 100, the above three access requests whose logical block addresses are separated from each other are PBA = 0, PBA = 10,
Since the data blocks are arranged at physically close positions as in PBA = 20, the data block is accessed at a short distance, so that the moving distance of the head at the time of access is shortened, and the required access time is shortened. The access performance is improved.

(Embodiment 2) FIG. 2 is a conceptual diagram showing an example of the configuration of a rotary storage subsystem in which a block address conversion method according to another embodiment of the present invention is implemented. FIG. 4 is a flowchart showing an example of the operation.

In the case of the second embodiment, the disk controller 22 under the subsystem controller 21 is provided with the address conversion logic 22a to execute the block address layout conversion. Is different.

The subsystem controller 21 instructs the disk controller 22 on the address conversion parameters. The conversion parameters are the number of logical volumes (N), the capacity of the logical volume (C), and the size (S) of the unit for dividing the logical volume. The conversion parameter is instructed, for example, by the subsystem controller 21 and the disk controller 22 by SCSI (Small Computer System Interface).
ace), it can be executed as part of a procedure such as a drive parameter setting command as illustrated in FIG.

The disk control device 22 that has received the read / write request from the subsystem control device 21 converts the data block address of the read / write request into the data block address based on the above-described equations (1) to (3) based on the conversion parameter. Performs block address layout conversion by control logic,
Access the magnetic disk drive 23. That is,
As illustrated in FIG. 4, the disk control device 22 reads from the subsystem control device 21 according to the LBA instruction.
When the / WRITE request is received (step 101), an operation for converting LBA to PBA is performed by a method described later (step 102), and further, the PBA is converted into the cylinder address CC and the head address H in the magnetic disk drive 23.
H, SEE which converts the head to a hardware address such as a sector address SS (step 103) and moves the head to a target cylinder indicated by the cylinder address CC.
K is started (step 104), and after the SEEK is completed, the target track indicated by the head address HH in the cylinder is selected, and READ / WRITE is executed on the area indicated by the sector address SS in the track (step 104). Step 106).

Now, it is assumed that in the step 101, the subsystem controller 21 requests the disk controller 22 to access one data block at the same relative position of each logical volume as shown below.

Data block address 40 (LBA = 40) which is information of relative No. 4 of logical volume 0 Data block address 90 (LBA = 90) which is information of relative No. 4 of logical volume 1 Certain data block address 140 (LBA = 140) In step 102, the disk controller 22 performs an address conversion on the address of the access request received from the subsystem controller 21 based on the conversion parameter as follows.

The data block address 40 (LBA =
40) → Conversion → Data block address 120 on disk (PBA = 120) Data block address 90 (LBA = 90) →
Conversion → Data block address 130 on disk
(PBA = 130) Data block address 140 (LBA = 140)
→ Conversion → Data block address 140 on disk
(PBA = 140) For example, when LBA = 40, LV = 0, α = 40,
Z = 4, β = 0, and PBA = 4 × 10 × 3 + 10 ×
By 0 + 0 = 120, it is converted to PBA = 120.

The physical block address PBA is CC: H
H: Since the hardware addresses such as SS are assigned in ascending or descending order so as to be continuous, a data block group having a close physical block address PBA is located at a physically close position on the magnetic disk. When accessing the data block group, the physical movement distance of the head is short. Therefore, as in this embodiment,
Data blocks located at the same relative position of a plurality of logical volumes and having different logical block addresses LBA such as 40, 90, 140, etc.
By converting to the adjacent physical block address PBA as described above, the required access time such as SEEK and rotation wait is shortened, and the access performance is surely improved.

(Embodiment 3) FIG. 5 is a conceptual diagram showing an example of the configuration of a rotary storage subsystem in which a method for controlling a rotary storage subsystem according to still another embodiment of the present invention is implemented.

In this embodiment, the host computer 210
And a plurality of magnetic disk drives 230 constituting a disk array, with a RAID controller 220 interposed therebetween. In other words, the RAID controller 220 distributes data transmitted to and received from the host computer 210 in a plurality of magnetic disk drives 230 and stores the data, thereby improving access performance by parallel data transfer and parity data from the data. RAI for Improving Data Reliability by Generating and Storing Data
In addition to configuring the D system, the operation for improving the access performance by converting the logical block address LBA into the physical block address PBA is performed.

That is, each magnetic disk drive 2
A plurality of logical volumes 0 to 3 are set in the logical volume 30, and the RAID controller 220 determines that the data blocks stored in the same position relatively in each logical volume are:
An operation of converting the logical block address LBA of the data block into a physical block address PBA is performed so that the logical block address is physically located in the magnetic disk as a medium. As the conversion algorithm, the technique exemplified in the first embodiment can be used.

In the case of the third embodiment, the access performance when a logical volume is set in a plurality of magnetic disk drives 230 constituting a disk array can be improved, and the performance of the entire RAID system can be improved.

(Embodiment 4) FIG. 6 is a conceptual diagram showing an example of the configuration of a rotary storage subsystem according to still another embodiment of the present invention. In the case of the fourth embodiment, a plurality of heads for recording / reproducing data to / from the rotary storage medium 300 so as to correspond to each of a plurality of logical volumes 0 to 3 set in the rotary storage medium 300. 301-3
04, and these heads 301 to 304 are supported by a common actuator 305 so that the rotary storage medium 3
In the radial direction of 00, they are simultaneously displaced in the same direction.

With such a configuration, the rotary storage medium 3
Thus, it is possible to simultaneously access the same relative position in each of the plurality of logical volumes 0 to 3 set above, and to simultaneously access the same relative position of each of the plurality of logical volumes 0 to 3. It is possible to improve the access performance in the case of taking an appropriate access form.

Although the invention made by the inventor has been specifically described based on the embodiments, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the invention. Needless to say.

For example, the rotary storage subsystem is not limited to the one using a magnetic disk as a medium, but can be widely applied to a rotary storage device such as an optical disk, a magneto-optical disk, etc., which requires a head seek operation at the time of access. .

[0052]

According to the block address conversion method of the present invention, the physical volume is divided into a plurality of logical volumes, and the access performance in a storage subsystem that divides each logical volume into a plurality of unit areas is used. The effect of being able to improve is obtained.

Further, according to the control method of the rotary storage subsystem of the present invention, the rotary storage medium is divided into a plurality of logical volumes, and each logical volume is divided into a plurality of unit areas for use. The effect is obtained that the access performance in the storage subsystem can be improved. According to the disk subsystem of the present invention, the rotary storage medium is divided into a plurality of logical volumes,
In a disk subsystem in which an individual logical volume is divided into a plurality of unit areas and used, it is possible to realize high-speed access when accessing information at the same relative position in a plurality of logical volumes in each logical volume. Is obtained.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing an example of a configuration of a rotary storage subsystem in which a block address conversion method according to an embodiment of the present invention is implemented.

FIG. 2 is a conceptual diagram showing an example of a configuration of a rotary storage subsystem in which a block address conversion method according to another embodiment of the present invention is implemented.

FIG. 3 is a flowchart illustrating an example of an operation of a rotary storage subsystem in which a block address conversion method according to another embodiment of the present invention is performed.

FIG. 4 is a flowchart illustrating an example of an operation of a rotary storage subsystem in which a block address conversion method according to another embodiment of the present invention is performed.

FIG. 5 is a conceptual diagram showing an example of a configuration of a rotary storage subsystem in which a method of controlling a rotary storage subsystem according to still another embodiment of the present invention is implemented.

FIG. 6 is a conceptual diagram showing an example of a configuration of a rotary storage subsystem according to still another embodiment of the present invention.

FIG. 7 is a conceptual diagram showing an example of a configuration of a conventional rotary storage subsystem.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Host computer, 20 ... Disk subsystem, 21 ... Subsystem control device, 21a ... Address conversion logic, 22 ... Disk control device, 22a ... Address conversion logic, 23 ... Magnetic disk drive, 24 ... Address transition table, 30, 40, 50: Interface cable, 210: Host computer, 220: RAID
Controller, 230: magnetic disk drive, 300
... Rotary storage medium, 301-304 ... Head, 305 ...
Actuator, LBA ... logical block address, PB
A: Physical block address.

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Naoya Takahashi 2880 Kozu, Kozuhara-shi, Kanagawa Pref.Hitachi, Ltd. Storage System Division (72) Inventor Akira Kurano 2880 Kozu, Kozu, Odawara-shi, Kanagawa Hitachi Storage System Co., Ltd. Within the Business Division (72) Inventor Akira Yamamoto 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Hitachi, Ltd.System Development Laboratory (72) Inventor Atsushi Takayasu 2880 Kozu, Kofu, Odawara City, Kanagawa Prefecture Storage System Business Division, Hitachi, Ltd. (72) Inventor Minoru Yoshida 2880 Kokuzu, Odawara-shi, Kanagawa Prefecture Storage Systems Division, Hitachi, Ltd. (56) References JP-A-2-304614 (JP, A) JP-A-58-76958 (JP, A) JP-A-2-227873 (JP, A) (58) Fields investigated ( Int.Cl. ⁷ , DB name) G06F 3/06

Claims (6)

(57) [Claims] 1. When a physical volume is divided into a plurality of logical volumes and used, the unit areas are arranged such that unit areas at the same relative position in each of the logical volumes are closely arranged on the physical volume. Converting a logical block address indicating the storage position of the data block belonging to the logical volume in the logical volume and a physical block address indicating the storage position in the physical volume. 2. A method for controlling a rotary storage subsystem, in which a rotary storage medium is divided into a plurality of logical volumes and controlled, wherein the logical volume is divided into unit areas of a desired length, and each logical volume is divided into logical volumes. Wherein the arrangement of the addresses of the data blocks belonging to the unit area is converted so that the unit areas having the same relative position exist close to each other on the rotary storage medium. Control method. 3. The method of controlling a rotary storage subsystem according to claim 2, wherein said rotary storage subsystem comprises:
A first control device provided in a rotary storage device including the rotary storage medium, and a plurality of the rotary storage devices and the host device interposed between the plurality of rotary storage devices and the host device; And a second control device for controlling the transfer of information between the first control device and the second control device. For the control device of
The first control device provides a conversion parameter including whether or not the address layout conversion is performed, the number of the logical volumes, the capacity of each logical volume, and the size of the unit area for dividing the logical volume. And a second method of performing the address arrangement conversion, wherein the control method of the rotary storage subsystem is performed. 4. A disk subsystem comprising: a disk drive; and a disk control device for controlling the disk drive, wherein the disk drive is divided into a plurality of logical volumes, and the disk control device When converting the block address of the volume to the block address of the disk drive, the same
The unit area at one relative position is
A disk subsystem that allocates addresses to closely adjacent locations . 5. A disk subsystem comprising a disk drive, a disk controller for controlling the disk drive, and a subsystem controller for controlling the entire system, wherein the disk drive is divided into a plurality of logical volumes. are, the subsystem controller when converting the block address of the logical volume block address of the disk drive, put in each logical volume
Unit area at the same relative position on the disk drive
A disk subsystem that allocates addresses to physically close locations . 6. A disk subsystem comprising: a disk drive divided into a plurality of logical volumes; and a disk control device for controlling the disk drive, wherein a unit of the same relative position in the plurality of logical volumes is provided.
The area is the physical block address of the disk drive
A disk subsystem arranged in close proximity to each other .