JP2002023960A - Disk device and computer system provided with plural disk devices - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the load of a host computer and to execute an efficient access by managing file constitution information by a disk device. SOLUTION: A meta data management part 122b for managing meta data defining the constitution of files stored in an HDD 121 is formed in a controller 122 built in a disk device 12. When the positional information of data constituting a file to be accessed is inquired from a file system 110 in a host computer 11, the meta data management part 122b reads out the meta data from the HDD 121, analyzes the read data to specify the position of the data constituting the file and informs the host computer 11 of the positional information. The host computer 11 requests a data access to the disk device 12 on the basis of the positional information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk device accessible from a host computer, and more particularly to a disk device having a function of managing file configuration information (metadata) defining a file configuration and a plurality of such devices. Computer system.

[0002]

2. Description of the Related Art Generally, data stored in a disk device such as a magnetic disk device in response to a request from a host computer is managed by a file system on the host computer side.

For this purpose, the host computer has management data (hereinafter, referred to as metadata) indicating at which address on the disk the data constituting each file is stored. Then, the host computer (the file system), when accessing the file,
The data at the corresponding address on the disk is accessed with reference to this metadata. At this time, the disk device only processes the access to the requested address, but does not sense the file configuration at all.

[0004]

Originally, disk device access from the host computer has an access pattern based on the configuration on the disk device for each file to be actually accessed. However, in the related art, the disk device does not manage the information on the file configuration, so that the pattern cannot be predicted.

For this reason, in the prior art, even if a large-capacity disk cache is provided, data reading and R
AID (Redundant Arrays of Inexpensive Disks)-
In internal data read processing (processing required for parity data generation) associated with writing in the five configuration, effective prefetch cannot be performed, and the effect of the cache is limited.

In a SAN (storage area network) environment recently proposed, data can be transferred between a plurality of storage devices (such as magnetic disk devices or magnetic tape devices) connected to the SAN. Actually, this function cannot be effectively used unless the storage device recognizes the file configuration.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to provide a disk device which can reduce the load on a host computer and efficiently access by managing file configuration information on the disk device side. Is to do.

Another object of the present invention is to provide a disk drive capable of improving the hit rate of a disk cache.

Still another object of the present invention is to provide a computer system capable of realizing file transfer between disk devices such as file copying in an environment in which a plurality of disk devices are connected by an interface such as a storage area network. is there.

Still another object of the present invention is to provide a computer system capable of realizing mirroring between disk devices in an environment in which a plurality of disk devices are connected by an interface such as a storage area network.

[0011]

According to the present invention, there is provided a disk drive connected to a host computer, comprising at least one disk drive for storing information including various files, and a controller for controlling the disk drive. A data access management means for managing access to the disk drive, and file configuration information defining a configuration of a file stored in the disk drive, and a file to be accessed for data access from a host computer. File position information management means for identifying the position of the data constituting the access target file based on the file configuration information and informing the host computer of the position information when the position information of the data constituting the file is inquired. Control Characterized in that a chromatography La.

In a disk device having such a configuration, management of file configuration information (metadata) defining a file configuration, which has conventionally been managed by a file system on the host computer side, is provided in a controller of the disk device. File configuration information management means). Therefore, on the host computer side, when accessing the file held in the disk device,
There is no need to read the file configuration information from the disk device and analyze it to specify the position of the data constituting the file, thereby reducing the load on the host computer. That is, when accessing the file, the host computer only needs to inquire the disk device of the position information of the data constituting the file and access the data of the disk device again based on the obtained position information of the data.

Although one disk device can be physically connected to a plurality of host computers, each host computer has a separate file system, so that file sharing is performed between the host computers. It is said to be an obstacle above. On the other hand, in the present invention, since the management of the file configuration information is performed on the disk device side, the disk device can be shared by a plurality of host computers, that is, data sharing among multiple hosts can be realized.

Further, the present invention applies the above configuration to a disk device having a disk cache, and provides the following functions to the data access management means, that is, the file configuration information management means stores the file to be accessed to the host computer. When the position information of the data to be constituted is notified, a function of reading in advance the data constituting the file to be accessed from the disk drive to the disk cache based on the position information is provided.

In the disk device having such a configuration, the data prefetch to the disk cache can be performed at the file level at the stage of inquiring the position information of the data constituting the file to be accessed from the host computer. The disk cache hit rate at the time of reading can be improved.

Further, a computer system according to the present invention is a computer system in which a plurality of disk devices of the above configuration are connected by a predetermined interface, and the plurality of disk devices of the above configuration are connected to a host computer by the interface. When a file copy request is given from the host computer to each disk device, the file configuration information management means specifies the position information of the data constituting the copy source file or the copy destination file, and uses the position information. And a file copying means for copying a file within the same disk device or between different disk devices.

In the computer system having such a configuration, functions that cannot be provided by the conventional system, such as copying (file transfer) files independently from the host computer in the same disk device or between different disk devices, are provided. it can.

Further, a computer system according to the present invention is a computer system in which a plurality of disk devices of the above configuration are connected by a predetermined interface, and the plurality of disk devices of the above configuration are connected to a host computer by the interface. In the disk unit, the following means,
That is, when a request to create a mirror file is given from the host computer, the file configuration information management means of the own device specifies the position information of the data constituting the target file and the file of the disk device where the mirror file is created. Mirror file creation means for creating a mirror file on the disk device where the mirror file is created by using the location information to specify the location information of the data constituting the mirror file by the configuration information management means; When a failure occurs at the time of a file access request, disk identification information for designating a disk device holding the mirror data of the requested file and mirror file notifying means for notifying the host computer of the position information of the mirror data are provided. It is characterized by the following.

In the computer system having such a configuration, mirroring between disk devices connected to an interface such as a SAN (storage area network) can be realized, and in particular, file units can be stored on any disk device on the interface. Mirror data can be created. Also, even if a failure occurs in the disk device requested at the time of a file access request from the host computer,
Instead of the location information of the data constituting the requested file, the location information of the mirror data of the file is notified to the host computer, so that the host computer can continue the file access.

[0020]

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

[First Embodiment] FIG. 1 is a block diagram showing a configuration of a computer system to which a disk device according to a first embodiment of the present invention is applied. The computer system of FIG. 1 includes a host computer 11 having a file system 110 and a disk device 12 accessible from the host computer 11 (the file system 110 therein).
It is composed of

The disk device 12 is, for example, a magnetic disk device, and is a SCSI (Small Computer System Interface).
The host computer 11 is connected by an interface 13 represented by e). As this interface, in addition to the above SCSI, FiberChannel
l, IEEE (Institute of Electrical and Electron
ics Engineers) 1394 or the like. In addition, the disk device 12 is
Assuming the case of realizing by AID or the like, an internal I / O bus interface of a computer such as a PCI bus (Peripheral Component Interconnect bus) may be applied.

The disk device 12 serves as an interface between a magnetic disk drive (hereinafter, referred to as HDD) 121 and the host computer 11 and has a
And a controller 122 for controlling the In addition,
Although a plurality of HDDs 121 can be connected to the controller 122, in the configuration of FIG.
An example in which only one D121 is connected is shown. HD
In the D121, a metadata area 121a for storing metadata as file configuration information (file configuration definition information) necessary for managing the configuration of various files stored in the HDD 121 is secured.

The controller 122 includes a data access management unit 122a for managing access to the HDD 121.
And a metadata management unit 122b that manages metadata stored in the HDD 121 (the metadata area 121a therein). Metadata management unit 122b
Is the file system 110 of the host computer 11.
Has a function of reading metadata in response to an inquiry about position information of data constituting a file, obtaining position information of the data, and notifying the file system 110 of the acquired metadata.

Next, the operation of the configuration of FIG. 1 will be described with reference to the operation explanatory diagram of FIG. When the host computer 11 attempts to access a file stored in the disk device 12 (the HDD 121 therein), the file system 110 of the host computer 11
Metadata management unit 1 provided in the controller 122
A file configuration data location specification request is given to the user 22b as an inquiry request for location information of data constituting a file to be accessed (step S1).

Upon receiving the file configuration data position specification request from the file system 110, the metadata management unit 122b reads the metadata from the metadata area 121a of the HDD 121 (step S2) and analyzes the read metadata. Then, the position of the data constituting the access target file specified by the file configuration data position specification request is specified (step S3). Then, the metadata management unit 122b notifies the requesting file system 110 of the position of the data constituting the specified access target file (Step S4). conventionally,
The operations in steps S2 and S3 are performed in the file system 1
10 himself had gone.

When notified of the position (position information) of the data constituting the file to be accessed from the metadata management unit 122b, the file system 110 controls the controller 122 of the disk device 12 based on (information of) the data position. Issues a data access request to the data access management unit 122a provided in
5).

Then, the data access management unit 122a
The HDD 121 is accessed in response to a data access request from the file system 110, and data is read from the data position specified by the data access request (when the access type is read), or data is written to the specified data position ( (When the access type is write) (step S6).

When the access type is write, the file system 110 sends a request to the metadata management unit 122b.
A metadata update request is issued as needed (step S7). Here, when the metadata needs to be updated, for example, when the data file is increased or decreased,
This is a case where writing that affects the content of metadata (file configuration information) is performed.

The metadata management unit 122b responds to a metadata update request from the file system
The metadata in the metadata area 121a of D121 is updated (step S8). The file system 110
When the metadata is updated by the metadata management unit 122b, the file access is completed.

As described above, in the present embodiment, in order to access a file from the host computer 11 (the file system 110 thereof) to the disk device 12, the metadata management unit 122b in the disk device 12 needs to
A file configuration data position specifying request is issued, and the data access management unit 122a based on the data position notified from the metadata management unit 122b in response to the request.
It is only necessary to issue a data access request to. Therefore, the load on the host computer 11 is reduced as compared with the conventional case where the file system 110 itself reads and analyzes the metadata from the HDD 121 and specifies the position of the data constituting the access target file. You.

[Second Embodiment] FIG. 3 is a block diagram showing a configuration of a computer system to which a disk device according to a second embodiment of the present invention is applied. The computer system shown in FIG. 3 includes a host computer 31 having a file system 310 and a disk device 32 connected to (the file system 310 of) the host computer 31 via an interface 33.

The disk device 32 has the same configuration as the disk device 12 in FIG. 1, that is, includes a HDD 321 in which a metadata area 321a is secured, and a controller 322. The controller 322 has a data access management unit 322a and a metadata management unit 322b. The disk device 32 is different from the disk device 12 in FIG. 1 in that, in addition to the HDD 321 and the controller 322, a cache memory (hereinafter, referred to as a disk cache) for holding a part of data stored in the HDD 321 is provided. Is a point.

A feature of the disk device 32 is that the metadata management unit 322b specifies the position of the data constituting the access target file in response to the file configuration data position specification request from (the file system 310 of) the host computer 31. In addition to the function, it has a function of prefetching data at the data position from the HDD 321 to the disk cache 323.

Next, the operation of the configuration of FIG. 3 will be described with reference to the operation explanatory diagram of FIG. When the host computer 31 attempts to access a file held in the disk device 32 (the HDD 31 therein), a file configuration data position specification request is sent from the file system 310 of the host computer 31 to the controller 322 of the disk device 32. (Step S11). This file configuration data position specification request is received by the metadata management unit 322b in the controller 322.

The metadata management unit 322b reads the metadata from the metadata area 321a of the HDD 321 (similar to the metadata management unit 122b in the first embodiment) (step S12), and analyzes the read metadata. Then, the position of the data constituting the access target file specified by the file configuration data position specification request is specified (step S13). Then, the metadata management unit 322b notifies the file system 310 of the request source of the position of the data constituting the specified access target file (Step S14).

At the same time, based on the position information notified to the file system 310, the metadata management unit 322b forms the data in the HDD 321 indicated by the position information, that is, the access target file specified by the file configuration data position specification request. It requests the data access management unit 322a to prefetch data (access target data) in the HDD 321 into the disk cache 323. Thereby, the data access management unit 322a prefetches the access target data at the position specified by the metadata management unit 322b from the HDD 321 to the disk cache 323 (Step S15). However, if the requested data is already stored in the disk cache 323, this prefetch operation is not necessary.

As described above, the data at the position specified in response to the file configuration data position request from the file system 310 is transferred from the HDD 321 to the disk cache 32.
The reason for prefetching to No. 3 is as follows. First, the fact that the host computer 31 has issued the file configuration data position specifying request can be predicted that the possibility that the host computer 31 will generate a data access request to the data position specified according to the request is extremely high. That is, the data access pattern on the host computer 31 side can be predicted on the disk device 32 side. Therefore, the data at the position specified in response to the file configuration data position specification request is transferred from the HDD 321 to the disk cache 3.
This is because if the data is prefetched to the storage location 23, the data access request for the specified position can be processed at high speed.

On the other hand, when the position (position information) of the data constituting the file to be accessed is notified from the metadata management unit 322b, the file system 310 determines the controller of the disk device 32 based on (information of) the data position. A data access request is issued to the H.322 (step S16). This request is received by the data access management unit 322a in the disk device 32.

Then, the data access management unit 322a
Data access is performed in response to a data access request from the file system 310 (step S17). The data at the data position specified by this data access request is
The data always exists on the disk cache 323 in the prefetch operation in step S15. Therefore, when the access type specified by the data access request is read, the data access request is
3, data access to the HDD 321 does not occur, and high-speed data access to the disk cache 323 is performed.

When the access type is write, data access (write access) to the disk cache 323 is performed irrespective of hit / miss hit. However, when the disk device 32 adopts the RAID-5 configuration, parity can be generated based on data in the disk cache 323 at the time of a hit, whereas parity must be generated based on data at the HDD 321 at the time of a mishit. That is, access to the HDD 321 occurs. Therefore, in the RAID-5 configuration, the above prefetch operation is effective even when the access type is write.

As described above, in the present embodiment, the metadata (file configuration information) is managed by (the metadata management unit 322b of) the disk device 32, and the file configuration data location from the host computer 31 (the file system 310) is managed. By processing the specific request in the disk device 32 (the metadata management unit 322b thereof), the disk device 32 can predict the data access pattern of the host computer 31 almost certainly. Therefore, by performing prefetch from the HDD 321 to the disk cache 323 according to this prediction, the efficiency of the disk cache 323 can be significantly improved.

Third Embodiment FIG. 5 is a block diagram showing a configuration of a computer system to which a disk device according to a third embodiment of the present invention is applied. The computer system in FIG. 5 includes a host computer 51 and a plurality of, for example, two disk devices A and B accessible from the host computer 51. Here, the disk device A and the disk device B are connected to the host computer 51 by an interface capable of connecting a plurality of disk devices, for example, a SAN (storage area network) 53.

The host computer 51 corresponds to the host computer 11 (31) in FIG. 1 (FIG. 3), and includes a disk device A, which has a file system (not shown).
B corresponds to the disk device 12 (32) in FIG. 1 (FIG. 3), and includes HDDs 521A and 521B and a controller (not shown) having a data access management unit and a metadata management unit. That is, the disk devices A and B have a function of managing metadata (file configuration information).
Note that the disk devices A and B may include a disk cache.

Next, a file copy operation between disk devices in the configuration of FIG. 5 will be described. The host computer 51 is, for example, a disk device A (with an HDD 52
When a file copy is to be performed from 1A) to the disk device B (the HDD 521B therein), a file copy request is sent to the disk device A via the SAN 53 (step S21). This file copy request includes information (file name, path, etc.) specifying the target file (copy source file), information specifying the copy destination disk device (B) (disk device identification information), and the copy destination file. Contains information to specify (file name, path, etc.).

Disk device A (metadata management unit therein)
Analyzes the metadata in response to a file copy request from the host computer 51, and specifies the data position of the data block group 54 constituting the designated file (copy source file) (step S22). Then, the disk device A sends information on the data position of the data block group 54 and information specifying the copy destination file included in the file copy request from the host computer 51 to the copy destination disk device B via the SAN 53. Request a file copy to the disk device B.

Then, (the data access management unit of) the disk device B sends a data read request to the disk device A in accordance with the data position information sent from the disk device A, so that the data constituting the copy source file is transmitted from the disk device A. The block group 54 is read, and the designated copy destination file (obtained by analyzing the metadata by the metadata management unit in the disk device B) is written to the position of the data to be composed. Thereby, the file copy from the disk device A specified by the host computer 51 to the disk device B is realized (step S23). In addition,
The disk device A moves to the copy destination disk device B,
A request is made to specify the position of the data to constitute the copy destination file, and further, based on the information on the data position specified by the disk device B, the data block group 54
Requesting the writing of the file can also realize the file copy from the disk device A to the disk device B.

As described above, in this embodiment, since both the copy source disk device A and the copy destination disk device B manage the file configuration information, the host computer 51 sends the copy source disk device A to the copy source disk device A. It is only necessary to send a file copy request including information specifying a target file (copy source file), information specifying a copy destination disk device, and information specifying a copy destination file. As a result, the load on the host computer 51, which has conventionally occurred with file copying, can be reduced. Note that the file copy may be performed in the same disk device.

Next, creation of a mirror file and access to a mirror file in the configuration of FIG. 5 will be described with reference to FIG. The host computer 51 uses a file write request when, for example, it is desired to create mirror data for a file on the disk device A (the HDD 521A therein) in another disk device (here, the disk device B) in file units. This file write request (mirror file creation request) includes information (file name, path, etc.) specifying the target file, information specifying the disk device (B) to which the mirror file is to be written (disk device identification information), and mirror information. It contains information that specifies a file (file name, path, etc.).

Now, it is assumed that a file write request has been sent from the host computer 51 to the disk device A (step S31). In this case, the disk device A (the metadata management unit therein) reads out and analyzes the metadata from the HDD 521A in response to the file write request from the host computer 51, thereby obtaining the data of the data block group 54 constituting the target file. The position is specified (step S32).

Then, the disk device A writes the data block group 54 to the HDD 521B of the disk device B in cooperation with the disk device B in the same procedure as in the case of the file copy described above, so that the The corresponding file-level mirror data 55 (mirror file) is created (step S33). At this time, the disk device A (the metadata management unit therein) holds the specification information of the disk device B where the mirror file is created and the specification information of the mirror file in association with the specification information of the target file.

Here, in order to simplify the explanation,
The mirror data of the file stored in the disk device A (the HDD 521A) is transferred to the disk device B (the HDD 5
21B), it is also possible to create mirror data (mirror file) of the file in the disk device B when writing the file to the disk device A.

Thereafter, it is assumed that the host computer 51 has given the disk device A a file configuration data position specification request for inquiring the data position of the file to be accessed for file access (step S41).

Disk device A (internal metadata management unit)
Receives the file configuration data position specification request from the host computer 51, reads out and analyzes the metadata from the HDD 521A, and specifies the data position of the data block group 54 configuring the access target file (step S42). If any failure has occurred on the HDD 521A holding the target data (step S43), the disk device A (the metadata management unit therein) holds the target data in association with the access target file. File configuration data position identification for inquiring the disk device B of the position of the data constituting the mirror file based on the disk device identification information specifying the disk device (B) where the mirror file is created and the information specifying the mirror file Send request. Thereby, the disk device A obtains information on the position of the data constituting the mirror file (that is, the position of the mirror data on the disk device B) from the disk device B, and sends it to the host computer 51 together with the information specifying the disk device B. Notify (step S44).

Upon receiving the notification from the disk device A, the host computer 51 performs actual data access to the disk device B indicated by the notified disk device identification information based on the notified mirror data position information. Perform (Step S45).

As described above, in the present embodiment, the SAN
Mirroring between disk devices on the SAN 53 can be realized, and mirror data can be created on any disk device on the SAN 53 in file units. This allows
A highly available and flexible system configuration in which a plurality of disk devices cooperate can be realized.

The present invention is not limited to the above embodiments, but can be variously modified at the stage of implementation without departing from the scope of the invention. Further, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some components are deleted from all the components shown in the embodiment, at least one of the problems described in the column of the problem to be solved by the invention can be solved, and the problem described in the column of the effect of the invention can be solved. In the case where at least one of the effects described above is obtained, a configuration from which this component is deleted can be extracted as an invention.

[0058]

As described above in detail, according to the present invention, the file structure information (metadata) defining the file structure is managed on the disk device side, so that the host computer can access the file. In this case, it is only necessary to inquire the disk device about the location information of the data constituting the file and to access the data of the disk device again based on the obtained location information of the data, thereby reducing the load on the host computer and improving efficiency. Can be accessed.

According to the present invention, the data prefetch to the disk cache can be performed at the file level at the stage of inquiring the position information of the data constituting the file to be accessed from the host computer. Can improve the disk cache hit rate and provide a disk device that can be accessed at high speed.

Further, according to the present invention, a disk device (same disk device,
Alternatively, file copying can be performed between the different disk devices without burdening the host computer.

Further, according to the present invention, a mirror file can be created between arbitrary disk devices connected to an interface such as a SAN, and even if a failure occurs in any one of the disk devices, a mirror file of the host computer can be created. File access can be continued.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a computer system to which a disk device according to a first embodiment of the present invention is applied.

FIG. 2 is an exemplary view for explaining an operation at the time of file access in the first embodiment.

FIG. 3 is an exemplary block diagram showing a configuration of a computer system to which a disk device according to a second embodiment of the present invention is applied.

FIG. 4 is an exemplary view for explaining an operation at the time of file access in the second embodiment.

FIG. 5 is a block diagram showing a configuration of a computer system to which a disk device according to a third embodiment of the present invention is applied.

FIG. 6 is an exemplary view for explaining creation of a mirror file and access to a mirror file in the third embodiment.

[Explanation of symbols]

11, 31, 51 ... host computer 12, 32, A, B ... disk device 13, 33 ... interface 53 ... SAN (storage area network, interface) 121, 321, 521A, 521B ... HDD 122, 322, 522A, 522B ... Controllers 122a, 322a: Data access management unit 122b, 322b: Metadata management unit (file configuration information management unit) 323: Disk cache

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 12/16 320 G06F 12/16 320L 13/10 340 13/10 340A F-term (Reference) 5B005 JJ13 KK02 MM11 NN22 5B014 EB04 FA03 5B018 GA04 HA04 KA14 MA12 QA01 5B065 BA01 CA11 CC03 CE11 CE22 EA31 EA33 ZA15

Claims (4)

[Claims] At least one storage device for storing information including various files in a disk device used by being connected to a host computer.
Two disk drives, a controller that controls the disk drives, a data access management unit that performs access management for the disk drives, and manages file configuration information that defines the configuration of files stored in the disk drives. From the host computer, when inquired about the position information of the data constituting the access target file for data access, specify the position of the data constituting the access target file based on the file configuration information, A disk device comprising: a controller having file configuration information management means for notifying position information to the host computer. 2. The system according to claim 1, further comprising a disk cache for holding a part of data stored in said disk drive, wherein said data access management means configures said file to be accessed to said host computer by said file configuration information management means. 2. The disk device according to claim 1, wherein, when the position information of the data to be transmitted is notified, data constituting the access target file is read from the disk drive to the disk cache in advance based on the position information. 3. The plurality of disk devices according to claim 1 or 2, which are accessible from a host computer, and the plurality of disk devices are connected, and the plurality of disk devices are connected to the host computer. Interface, the disk device, when a file copy request is given from the host computer, the file configuration information management means to specify the position information of the data constituting the copy source file or the copy destination file, A computer system comprising a file copy unit for copying a file within the same disk device or between different disk devices by using the position information. 4. The plurality of disk devices according to claim 1 or 2, which are accessible from a host computer, and the plurality of disk devices are connected, and the plurality of disk devices are connected to the host computer. The disk device, when receiving a mirror file creation request from the host computer, causes the file configuration information management means of the disk device to specify position information of data constituting the target file. At the same time, the position information of the data constituting the mirror file is specified by the file configuration information management means of the disk device where the mirror file is created, and the mirror information is specified on the disk device where the mirror file is created by using the position information. Mirror file creation means for creating files, and Mirror file notifying means for notifying the host computer of disk identification information designating a disk device holding the mirror data of the requested file and position information of the mirror data when a failure occurs at the time of a file access request from the host computer A computer system comprising: