JP2000322369A - Disk device and computer system using the device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform exclusive management of the logical units of a disk device to a host computer without using a command for exclusive management by converting a logical unit number designated by the host computer into a logic unit number held by the disk device. SOLUTION: This disk device 103 is a hard disk drive and include a logical unit number converting part. The drive 103 internally has logical units 101a to 101d. The logical unit number converting part converts a logical unit number indicated by a host computer into numbers (drive LUN) 0 to 3 for managing the logical unit 101 in the drive 103. The host computer sets and registers the correspondences between the numbers attached internally to the units 101a to 101d by the host computer and numbers for the drive 103 internally managing the units 101a to 101d to/in a logical unit number conversion table 102.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control method for a logical unit of a disk drive, and a computer system comprising the disk drive and one or more host computers.

[0002]

2. Description of the Related Art One or a plurality of disk devices having a plurality of logical units therein and a plurality of host computers are connected to a Fi
In a computer system that is connected by a high-speed interface such as bre Channel and multiple host computers share one or multiple disk devices, ATAs that are individually connected to each host computer are considered in terms of high speed and price. (AT Att
achment) and SCSI interface disk devices,
There is a high demand to store information unique to each host computer such as an OS in a disk device having a high-speed interface such as Fiber Channel and shared by a plurality of host computers. To store information unique to each host computer in a disk device shared by multiple host computers,
It is essential to exclusively manage a logical unit storing information unique to each host computer as a logical unit dedicated to the host computer.

When a plurality of host computers and one or more disk devices are connected by Fiber Channel, the physical interface is Fiber Channel and the logical interface is usually SCSI.

In SCSI, as commands for exclusively managing logical units, a RESERVE / RELEASE command and a PERSI
STENT RESEERVE IN / OUT command is specified,
These SCSI commands have the following problems in managing exclusive logical units exclusively for each host computer with respect to other host computers.

In the RESERVE / RELEASE command, when the power of the disk device is turned off or the disk device is reset, the RESERVE state of the logical unit is released.
Therefore, when the power is turned on, exclusive management of a dedicated logical unit cannot be performed for each host computer.

[0006] The PERSISTENT RESERVE IN / OUT command does not release the PERSISTENT RESERVE state even if the power of the disk device is turned off or the disk device is reset. However, another host computer holding the reservation key is used. Since the PERSISTENT RESERVE status can be changed, exclusive logical units cannot be completely exclusively managed for each host computer.

In order to exclusively manage a dedicated logical unit for each host computer, a mechanism for exclusively managing the logical unit with respect to the host computer is required inside the disk device, instead of exclusive management using a SCSI command.

[0008] A conventional control method of a logical unit of a disk drive is described in, for example, "SCSI Bus System"
-44181) "" and "Block address conversion method and rotation type storage subsystem control method (Japanese Patent Laid-Open No. 9-62452)"
It is described in. When the above system is configured using conventional disk units, exclusive management of logical units can be performed by using the SCSI RESERVE / RELEASE command or PERSISTENT RE
There was no method other than using the SERVE IN / OUT command, and the exclusive logical unit dedicated to the host computer could not be completely managed.

In the conventional disk device and host computer, the logical disk number assigned by the disk device to manage the logical disk internally and the logical disk number assigned by the host computer to manage the logical disk internally are one. Therefore, if the host computer specifies the same logical unit number, the same logical unit inside the corresponding disk device is accessed regardless of which host computer accesses.

Further, when the host computer and the disk device are connected by the above interface, when the power is turned on, the host computer specifies a logical unit number for each disk device in order from 0 and issues a SCSI INQUIRY command. Transmit and recognize the disk unit and the logical unit inside the disk unit. Normally, the host computer sends the INQUIRY commands in order from the logical unit number 0 for one disk device until it detects that there is no logical unit inside the specified disk device in response to the INQUIRY command. For this reason, the logical units are numbered consecutively and managed inside the host computer.

In the above computer system, it is desirable to store shared applications and data used in the computer system in a disk device having a high-speed interface such as Fiber Channel and shared by a plurality of host computers, from the viewpoint of maintainability and cost. The demand is high. A conventional disk device usually has a logical unit of number 0 therein.

[0012]

SUMMARY OF THE INVENTION An object of the present invention is to provide a computer system in which one or more disk devices are shared by a plurality of host computers, by using a logical unit of the disk device to execute a SCSI command for exclusive management. An object of the present invention is to propose a mechanism in a disk device that exclusively manages a host computer without using it.

[0013]

The disk device includes a hard disk drive and a disk array.

The hard disk drive comprises a disk control unit and a disk drive.

The disk controller comprises a host interface controller, a host manager, a logical unit manager, an address converter, a command processor, a drive cache, a drive cache controller, a drive controller, and a data bus.

The disk array is a mechanism for realizing high performance and high reliability of the disk system. In a disk array, a physical hard disk drive appears as one hard disk drive to a host computer for speeding up. On the other hand, for higher reliability, if a failure occurs in a hard disk drive storing data, redundant data for performing data recovery is stored in another hard disk drive.

The disk array comprises one or a plurality of disk array controllers and a plurality of hard disk drives. The interface between the disk array controller and multiple hard disk drives is Fiber Ch
Annel, SCSI, etc. are used. The disk array controller includes a host interface controller, a host manager, a logical unit manager, an address converter, a command processor, a drive interface controller, a drive cache, and a drive cache controller.

In order to achieve the object of the present invention, a plurality of host computers and one or a plurality of disk devices having a plurality of logical units therein are provided with a physical interface of Fibr.
e Channel and logical interface are SCSI and FC-AL. In the computer system connected by FC-AL topology, the logical unit management part of the disk unit and the logical unit number assigned internally by the host computer and the logical unit assigned inside the disk unit. A logical unit correspondence storage unit for storing correspondence of unit numbers, and a logical unit number for converting a logical unit number designated by a SCSI command by a corresponding host computer into a logical unit number inside the disk device using the above logical unit correspondence storage unit A conversion unit is newly provided. Further, a logical unit correspondence setting unit for setting and storing the correspondence between the logical unit number internally assigned by the host computer and the logical unit number internally assigned by the disk device to the logical unit correspondence storage unit is provided in the logical unit management unit. .

The function and operation of the disk control unit of the hard disk drive will be described. The host interface control unit controls FC-AL signals and protocols, controls link with the host computer, manages addresses on the loop of the hard disk drive, receives SCSI commands,
Sends data flow control and response to SCSI command. The host management unit
World Wide N which is a unique address of bre Channel Port
ame (WWN) and the corresponding host computer and disk unit
Stores and manages unique communication environment information between Fiber Channel Ports.

The logical unit management unit manages the physical position and capacity of each logical unit of the hard disk drive. Also, the logical unit is exclusively managed for the host computer. In order to perform the exclusive control, the logical unit number conversion unit converts the logical unit number specified by the SCSI command into a logical unit number inside the hard disk drive.

The address conversion unit converts a logical block address (LBA) specified by a SCSI command into a physical position of a disk drive.

The data bus is an internal bus for transferring data between the host interface and the disk drive.

The command processing section interprets and controls execution of the SCSI command received from the host computer.

The drive cache is a temporary memory for absorbing a difference in data transfer speed between the host interface and the disk drive, and is managed by a cache management unit.

The drive control unit manages the format of the disk drive and controls writing and reading of data to and from the disk drive.

Next, the function and operation of the disk array controller will be described. Parts other than the address conversion section and the drive interface section have the same functions as the respective sections of the disk control section of the hard disk drive. Here, the address conversion unit and the drive interface control unit will be described.

The address conversion unit converts and generates a SCSI command request transmitted by the host computer into a SCSI command to a hard disk drive in the disk array.

The drive interface control unit controls an interface between the disk array controller and the hard disk drive in the disk array, transmits a command converted and generated by the address conversion unit to the hard disk drive, and transmits a command between the disk array controller and the hard disk drive. Control data flow.

In the logical unit number correspondence setting section newly provided in the present invention, the number assigned inside the disk device to the dedicated logical unit for each host computer is added to the logical unit number assigned inside the corresponding host computer by one. A function is set to correspond to the pair and set in the logical unit correspondence storage unit. On the other hand, a number assigned inside a disk device to a logical unit shared by a plurality of host computers is associated with a logical unit number assigned internally by each of the plurality of host computers, and a function of setting the number in a logical unit correspondence storage unit is provided. .

The logical unit number converter converts the WWN unique to each Fiber Channel Port of the host computer and the logical unit number specified for each command by the host computer using the above-mentioned logical unit correspondence storage unit inside the disk device. Convert to the assigned logical unit number. Therefore, the logical unit management unit using the logical unit number conversion unit
A dedicated logical unit can be exclusively managed for another host computer for each logical unit host calculation.
When the logical unit correspondence storage unit, the logical unit correspondence setting unit, and the logical unit number conversion unit of the present invention are used, exclusive management of a logical unit storing information unique to each host computer is performed without using a SCSI command for exclusive management. be able to.

[0031]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to FIGS. This embodiment is an example in which the present invention is applied to a hard disk drive. FIG. 3 is a system configuration diagram of a computer system including a host computer and the disk device of the present invention. Two host computers 301a,
301b and disk unit 103 have physical interface Fiber C
hannel, logical interface is connected by SCSI.
For connection, a loop topology called FC-AL is used. The SCSI command is transmitted from one of the host computers 301a and 301b to the disk device 103 as an FC-AL frame. Between the host computers 301a and 301b and the disk device 103, two Loop232a and Loop232b
Exists. The host computer 301a has a WWN (World Wide Na
me) are WWN302a and WWN302c, two Fiber Channel Po
has rt. WWN302a-302d is each FibreChannel P
This is a unique value for identifying ort. Has WWN302a
Fiber ChannelPort is connected to Loop232a and WWN30
The Fiber Channel Port having 2c is connected to Loop232b. The host computer 301b has two Fiber Channel Ports whose WWNs are WWN 302b and WWN 302d. Fiber Channel Port with WWN302b is connected to Loop232a,
The Fiber Channel Port having WWN 302d is connected to Loop232b.

In this embodiment, the disk device 103 is a hard disk drive. This hard disk drive 103 has two independent Fiber Channel Ports 211a and 211b,
ibreChannel Port 211a is Loop232a, Fiber Channel Por
t 211b is connected to Loop232b.

The hard disk drive 103 includes the hard disk control unit 20 and the disk drive 250. FIG. 2 is a configuration diagram of the disk control unit according to the first embodiment of this invention.

The disk control unit 201 has a Fiber Channel Con
troller 202, CPU 203, ROM 204, RAM 205, drive cache 206, Disk Formatter 207, drive control circuit 20
8, a data bus 209 and a control bus 210.

The Fiber Channel Controller 202 is an FC-AL
This is hardware that controls the signals and protocols of the host, and performs Link control and frame transmission / reception with the host computer.

The CPU 203 uses the control bus 210 to
The disk control unit 201 is read out and executed, and controls the components of the disk control unit 201 to control the disk control unit 201 in an integrated manner. The CPU 203 stores the temporary control information during the execution of the program in the RAM 205 and refers to it. In this embodiment, the ROM 204 is a 512 KB flash memory, and the RAM 205 is a 256 KB DRAM.

The drive cache 206 is a 2 MB DRAM,
This is a temporary buffer for absorbing the difference between the data transfer speed of the FC-AL (Fibre Channel Arbitrated Loop) and the disk drive 103.

Disk Formatter 207 is a disk drive 2
Manage the physical format that stores 50 data,
It is hardware for controlling the reading and writing of data of the disk drive 250, and inputs and outputs drive read / write data 212 and a read / write control signal 213.

The data bus 209 is a Fiber Channel Control
ler202 and drive cache 206, Disk Formatter207
And data transfer between the drive cache 206.

The drive control circuit 208 is hardware for controlling an actuator for positioning the head of the disk drive 250 and a spindle motor for performing constant rotation control of the medium disk. The drive control circuit 208 includes a head positioning control signal and a spindle for rotating the medium disk. Input / output drive control signals such as motor control signals.

Next, the program stored in the ROM 204 and
The RAM 205 will be described.

The programs stored in the ROM 204 are a real-time OS 213, an FC-AL control unit 214, a disk control circuit 2
15, LU management unit 217, command queue management unit 219, cache management unit 218, command processing unit 220, initialization program 22
Consists of nine.

The RAM 205 includes a system information area 222, a disk control information area 223, an FC-AL control information area 224, a host management information area 225, an LU management information area 226, and a command queue 22.
Including 8.

The real-time OS 213 manages task scheduling, resources such as memory, interrupts, and timers. The system information area 222 is a real-time OS2
An area 13 is used to store task control information, a program stack, and the like.

The FC-AL control unit 214 has a Fiber Channel Control
ller202 is controlled by the Fiber Channel Por of the host computer.
Link between t and Fiber Channel Port of disk unit 103
This driver performs control, manages the address of the hard disk drive 103 on Loop232, controls data flow between the Fiber Channel Controller 202 and the drive cache 206, receives SCSI commands, and transmits responses to SCSI commands. The FC-AL control information area 224 is
l Address on WWN302e and Loop232a of Port 211a, Fi
This is an area for storing the address on the WWN 302f and the Loop 232b of the bre Channel Port 211b.

The disk control circuit 215 is a driver for accessing data of the disk drive 250.
This is a program for controlling the Formatter 207 and controlling the data flow between the disk drive 250 and the drive cache 206. The disk control control information area 223 includes information specific to the disk drive 250, such as information on the number of disks of the disk drive, the number of heads, the data format of the disk drive 250, defective area information, replacement area and replacement information, and information on error recovery. Is stored.

The host management unit 216 includes a host computer 301a,
301b, Fiber Channel Port 211a, 211b and disk unit 103
WWNs 302a-302d of the Fiber Channel Ports of the host computers 301a and 301b
The program is stored in the host management information area 225 in association with any of the above and managed.

The LU management unit 217 has a hard disk drive 1
03 This program manages the internal logical units. LU
The management unit 217 manages the physical position and capacity of each logical unit in the hard disk drive 103. Also,
The logical units inside the hard disk drive 103 are exclusively managed for the host computers 301a and 301b. Although details will be described later, the LU management unit 217 includes a logical unit number conversion unit 230,
A logical unit correspondence setting unit 231 is included. LU management information area 226
Is an area used by the LU management unit 217, and stores the physical position and capacity of each logical unit inside the disk drive. Further, although details will be described later, it includes a logical unit number conversion table 102.

The command queue management unit 219 is a program that stores the SCSI commands received from the host computers 301a and 301b in the command queue 228 and dequeues the SCSI commands that have transmitted the response. Also, the execution order of the stored SCSI commands is scheduled.

The cache management unit 218 determines the start logical address (LB) of the data stored in the drive cache 206.
A), data length (Block Length), drive cache 20
Cache management information area 227 by associating the addresses in 6
And manages the data in the drive cache 206 and the free area of the drive cache 206. Also, LBA, Block L specified by SCSI command
From the length, it is determined whether the data is stored in the drive cache 206.

The address conversion unit 220 is a program for converting the LBA specified by the SCSI command into physical addresses of the cylinder number, head number, and sector number of the disk drive 250. Also, Block Le specified by SCSI command
Convert ngth to the number of sectors in disk drive 250. At the time of conversion, information unique to the disk drive 250 stored in the disk control information area 223 is referred to.

The command processing unit 221 includes a host computer 301a,
This is a program that interprets and executes the SCSI command received from 301b. The command processing unit 221 calls and operates the FC-AL control unit 214, the disk control circuit 215, and the cache management unit 218 according to the interpreted command to execute the SCSI command.

The initialization program 229 is executed when the power is turned on or when the system is reset. The initialization program 229 initializes the entire disk controller 201 including the initialization of the RAM 205.

As described above, the disk control unit 201
It comprises a host interface control unit, a host management unit, a logical unit management unit, an address conversion unit, a command processing unit, a drive cache, a drive cache control unit, a drive control unit, and a data bus.

In this embodiment, the host interface controller is controlled by the Fiber Channel Controller 202, the hardware of the CPU 203, the program of the FC-AL controller 214 stored in the ROM 204, and the FC-AL control information area 224 of the RAM 205. The unit is the CPU 203 and the program of the host management unit 216 and the host management information area 225, and the logical unit management unit is the CPU 203.
And LU management unit 217 program and host management information area 2
In 25, the address conversion unit is the program of the CPU 203 and the address conversion unit 220 and the disk control information area 223, and the command processing unit is the CPU 203 and the program in the ROM 204 and the RAM 205.
The drive cache control unit is the CPU 203, the cache management unit 218, and the cache management information area 227, and the drive control unit is the Disk Formatter 207, the drive control circuit 208, and the CP.
It consists of the hardware of U203, the program of the disk control circuit 215, and the disk control information area 223.

Also, the above-described logical unit correspondence storage unit is
In the logical unit number conversion table 102 in the RAM 205, the setting unit for the logical unit corresponding to the SCSI Mode Select command and the CPU
In the program of the logical unit correspondence setting unit 231 and the logical unit number conversion unit, the logical unit number conversion unit is replaced by the CPU 203 and the logical unit number conversion unit.
It consisted of 230 programs.

FIG. 1 is a diagram showing the logical unit configuration of the disk device 103 and the table format of the logical unit number conversion table 102 according to the first embodiment of the present invention.

In this embodiment, the disk device 103 has four logical units 101a, 101b, 101c and 101d inside.

In this embodiment, the logical units 101a and 101c
Is a logical unit that stores information unique to the host computer 301a, a logical unit 101b is a logical unit that stores information unique to the host computer 301b, and a logical unit 101d is a logical unit shared by the host computers 301a and 301b. A case will be described as an example.

The host computer 301a internally has 0 for the logical unit 101a, 1 for the logical unit 101b, and 101 for the logical unit 101d.
Access by numbering 2. The host computer 301b
Internally, the logical unit 101c is numbered 0 and the logical unit 101d is numbered 1 for access.

On the other hand, the disk device 103 internally manages the logical units 101a to 101d by numbering 0, 1, 2, and 3 sequentially from the logical unit 101a.

The logical unit number conversion table 102 stores the WWN 235a-2 of the Fiber Channel Port of the host computers 301a and 301b.
Host computer WWN 104 for storing 35c, host computer LUN 105 for storing numbers for the host computers 301a and 301b to internally manage logical units, disk device
103 has a drive LUN 106 area for storing a number for internally managing the logical units 101a to 101d.

The correspondence between the numbers internally assigned to the logical units 101a-101d by the host computers 301a and 301b and the numbers for the disk units 103 to internally manage the logical units 101a-101d is set in the logical unit number conversion table 102. Although the registration process will be described later, the WWN of the Fiber Channel Port
Is the WWN 302a, the logical unit number 0 of the host computer 301a
Is associated with the logical unit number 0 inside the disk device 103, the logical unit number 1 in the host computer 301a is associated with the logical unit number 2 inside the disk device 103, and the WWN of the Fiber Channel Port is WWN 235b. The logical unit number 0 of the host computer 301b is associated with the logical unit number 1 inside the disk device 103, and is stored in the logical unit number conversion table 102. The logical unit number 2 of the host computer 301a and the logical unit number 1 of the host computer 301b are associated with the logical unit number 3 inside the disk device 103 and stored in the logical unit number conversion table 102.

The logical unit number converter 230 receives the WWN 104 of the Fiber Channel Port and the host LUN of the corresponding host computer 301a, 301b, and outputs a drive LUN 106 or a LUN error. FIG. 4 shows that the logical unit number conversion unit 230 changes the logical unit number (host LUN) designated by the host computers 301a and 301b for each SCSI command into a number (drive LUN) for managing the logical unit inside the disk device 103. It is a flowchart of a process of converting. When called, the logical unit number conversion unit 230 executes the following processing in order.

(1) Identification of Host Computers 301a and 301b In steps 41 and 42, it is confirmed whether the input WWN is stored in the host computer WWN 104 area of the logical unit number conversion table 102, and the host computers 301a and 301b are checked. Identify. If it is stored, it is determined that it can be identified, and the process of (2) described later is executed. If not, a LUN error is output in step 44, and the process returns to the caller. For example, WWNs WWN235a, WWN235b, WWN235c, and WWN235d are stored in the host computer WWN104 area.
The host computers 301a and 301b having the ibre Channel Port can be identified and execute the process (2). If a WWN other than the above is input, a LUN error is output and the process returns to the caller.

(2) Conversion of the input host computer LUN The host computer in which the input host computer LUN is associated with the corresponding WWN in the logical unit number conversion table 102
Check if it is stored in the LUN105 area. If it is stored, the drive LUN 106 corresponding to the host computer LUN is output in step 43, and the process returns to the calling source. If the host computer LUN is not stored in the host computer LUN 105 area of the logical unit number conversion table 102, a LUN error is output and the process returns to the caller. For example, if the entered WWN is W
If it is WN302a and the host computer LUN is 0, it outputs drive LUN0 and returns to the caller.

The input WWN is WWN 302a and the host computer
If the LUN is 4, output a LUN error and return to the caller.

Next, the disk device 103 is connected to the host computer 301.
a, the flow of processing for executing the SCSI command received from 301b, in which the LU management unit 217
The process of exclusively managing 101d will be described.

FIG. 5 shows that the disk device 103 is connected to the host computer 30.
CPU 20 when executing SCSI command received from 1a, 301b
FIG. 3 shows the flow of a program in the ROM 204.

In FC-AL, N-Por
t Login and Process Login are performed, and the host computer 301a, 3
01b Fiber Channel Port and Disk Unit 103 Fiber Ch
It is necessary to communicate information about the communication environment between the annel ports and configure an image pair. N-Port Lo
gin, Process Login is a Link Service defined by the Fiber Channel standard.
Used for communication environment setting processing between nnel Ports. At this time, the host computer 301 and the Fiber Channel
l The WWN 304 of Port is notified to each other. Disk unit 1
03 stores the information of this communication environment and Image Pair in the host management information area 225.

In this embodiment, the host management information area 225
Has an area in which management information on 16 different host computers 301 can be stored. Each area stores an initiator number 502, a Loop 501, a host computer WWN 104, an S_ID 503, and transfer control information 504.

Initiator number 502 is N-Port Login, P
When configuring the Imagae Pair by performing rocess login, the 0 attached to the corresponding host computer 301 in the disk device 103
Numbers from 15 to 15. Loop501 is the corresponding host computer 3
01a and 301b are values indicating the connected Loop. Fiber Channel Port of the corresponding host computer 301a, 301b is Loop232a
Is stored if it is connected to Loop232b, and 1 if it is connected to Loop232b. S_ID503 is the corresponding host computer 301a, 301b
Of the Fiber Channel Port on the FC-AL.
The transfer control information 504 is control information necessary for data transfer such as the maximum frame length that can be received by the corresponding host computers 301a and 301b.

Upon receiving SCSI commands from the host computers 301a and 301b, the Fiber Channel Controller 202
Causes a SCSI command reception interrupt 508 to occur.

The received SCSI command is stored in the command queue.
Stored in 228. In the present embodiment, the command queue 228 is
It has an area that can store 128 SCSI commands. Each area has a SCSI command CDB (Command Discripter Blo
ck) 507 is stored in association with the initiator number 502, the tag information 506, and the host LUN 105. The tag information 506 is a tag type and a 2-byte tag number specified for each SCSI command.

When a SCSI command reception interrupt 508 occurs, F
The C-AL control unit 214 is called. The FC-AL control unit 214 stores the received SCSI command CDB 507 in the corresponding host computer 30.
1a, 301b initiator number 502, tag information 506, host L
It is stored in the command queue 228 together with the UN 105, and returns to the calling source.

When a SCSI command is stored in the command queue 228, an internal event of the program occurs, and the real-time OS 213 calls the command queue management unit 219. The command queue management unit 219 refers to the tag information 506, schedules the SCSI commands stored in the command queue 228, selects the SCSI command to be executed, outputs the corresponding command queue number 509, and returns to the caller . The command queue number 509 is a number relatively indicating the position where the corresponding SCSI command is stored in the command queue 228. When the command queue number 509 is output, an internal event of the program occurs and the real-time OS 213
Calls the LU management unit 217.

The LU management unit 217 exclusively manages a logical unit by the following two methods.

(1) Exclusive management of a logical unit by a mechanism inside the disk device 103 The logical unit number conversion unit 230 is called, and the host LUN 105 associated with the SCSI command at the position indicated by the command queue number 509 in the command queue 228. Drive LUN1
Attempt to convert to 06. The LU management unit 217 provides the logical unit number conversion unit 230 with the host computer WWN 104 and the host LUN 10
Pass 5 as input. At this time, the LU management unit 217 refers to the command queue 228 and the host management information area 225, and acquires the host computer WWN 104 corresponding to the initiator number 502 associated with the SCSI command at the position indicated by the command queue number 509. I do. If the output of the logical unit number conversion unit 230 is a LUN error, the LU management unit 217 outputs a LUN error and returns to the call source.

(2) Exclusive Management of Logical Unit by SCSI Command If the output of the logical unit number conversion unit 230 is the drive LUN 106, the Reserve state of the drive LUN 106 and the Persi
From the sitent Reserve state, it is checked whether the corresponding host computers 301a and 301b have the right to access the corresponding drive LUN. As a result of the confirmation, if the corresponding host computers 301a and 301b have the right to access the corresponding drive LUN 106, the LU management unit 217 outputs the corresponding drive LUN 106 and returns to the call source. The corresponding host computer 301a, 301b is the corresponding drive LUN10
If you do not have permission to access 6, Reservation Conflic
Output t and return to caller.

If the output of the LU management unit 217 is a LUN error,
In the case of eation Conflict, the command processing unit 221 is called, and the host computer is connected via another FC-AL control unit 214.
An error response of the SCSI command is transmitted to 301a and 301b to terminate the execution of the SCSI command and return to the caller. LU
If the output of the management unit 217 is the drive LUN 106, the command processing unit 221 calls the address conversion unit 220 if the corresponding SCSI command is a command to access the disk drive 250 such as READ, WRITE, SEEK, and The LBA in the CDB 507 is converted into a physical address of the cylinder number, head number, and sector number of the disk drive 250. Also, set the Block Length in CDB507 to the disk drive
Convert to 250 sectors. Further, the command processing unit 221
Calls the other FC-AL control unit 214 and the disk control circuit 215, and the host computers 301a and 301b and the disk drive 250
Data transfer between the devices and transmission of a response to the corresponding SCSI command are performed, the execution of the SCSI command is terminated, and the process returns to the caller.

When the execution of the SCSI command ends, an internal event of the program occurs, and the real-time OS 213 calls the command queue management unit 219. The command queue management unit 219 dequeues the executed SCSI command.

Next, the processing in which the host computers 301a and 301b set and store the correspondence between the number assigned to the logical unit internally and the number assigned to the logical unit inside the disk device 103 in the logical unit number correspondence storage unit 230. explain. The host computers 301a and 301b notify the disk device 103 of the numbers assigned to the logical units unique to the host computers 301a and 301b and the numbers assigned to the logical units shared with the other host computers 301a and 301b by the SCSI Mode Select command. .

FIG. 6 shows the host computer using the Mode Select command.
The page format of a page parameter for notifying the numbers internally assigned to the logical units by 301a and 301b is shown. The parameter page consists of 6-bit page code 601, 1-byte page length 602, number of ports 607, 1-byte dedicated logical unit 603, 1-byte shared logical unit 604, 8-byte dedicated logical unit number 605, 8 It is composed of byte shared logical unit numbers 606, and parameters are arranged in the format of FIG. Dedicated logical unit number 605 is 8
There are only 603 dedicated logical units in byte units.
The shared logical unit number 606 has an area corresponding to the value stored in the shared logical unit number area 604 in units of 8 bytes.
The page code 601 is a code representing the corresponding page of the Mode Select command, and is 0x23 in this example. Page length
602 indicates the length of the corresponding page in bytes. The number of dedicated logical units 603 is the number of logical units that store information unique to the host computers 301a and 301b. Number of shared logical units
Reference numeral 604 denotes the number of logical units shared by a plurality of host computers 301a and 301b. The dedicated logical unit number 605 is a number internally assigned by the host computers 301a and 301b to a logical unit that stores information unique to the host computer. The shared logical unit number 606 is assigned to a plurality of host computers 301a, 301
This number is internally assigned to the logical unit shared by 1b.

FIG. 7 shows that the host computer 301a is a WWN 302a Po
The corresponding page parameter transmitted from rt to the disk device 103 by the Mode Select command is shown. The page length 602 is 0x26. In the present embodiment, the host computer 301a creates two logical units for storing unique information, so the number of dedicated logical units 603 is 0x02, and the two logical units have 0, 1 inside the host computer 301a. And numbered,
The dedicated logical unit number 605a is 0x00, and the dedicated logical unit number 605b is 0x01. In addition, a plurality of host computers 301a, 301
Since one logical unit to be shared is created in 01b, the number of shared logical units 604 is 0x01. Shared logical unit number 6
06 is 0x02.

When using the disk device 103 for the first time, the host computer 301a transmits this page parameter of the Mode Select command to the disk device 103 for each Port.

The operation of the command processing unit 221 (FIG. 5) and the logical unit number correspondence setting unit 231 (FIG. 2) when the disk device 103 receives the Mode Select command will be described. Mo
The processing from the reception of the de Select command to the storage in the command queue 228 is performed by the CPU 2 when executing the above SCSI command.
03 is the same as the program flow in the ROM 204. When the Mode Select command is stored in the command queue 228, an internal event of the program occurs and the real-time OS 21
3 calls the command processing unit 221. Command processing unit 221
Calls the FC-AL control unit 214 and receives the corresponding page parameter. Next, the command processing unit 221 calls the logical unit number correspondence setting unit 231. The logical unit number correspondence setting unit 231 includes, among the internal logical units 101a to 101d,
The dedicated logical unit number 605 of the page parameter is associated with the number of the logical unit 101a-101d that cannot be accessed from any of the host computers 301a and 301b, and is set in the logical unit number conversion table 102. In this example, since the WWN of the host computer 301a is transmitting a Mode Select command from the Fiber Channel Port whose WWN is 302a, the WWN 302a, 0x00 of the dedicated logical unit number 605a, and the number 0x00 assigned to the logical unit inside the disk device 103 are added. The logical unit number conversion table 102 is set in association with the logical unit number conversion table 102.

When the power is turned on, the host computer 301a
In order to recognize the disk device and the logical unit for each Fiber Channel Port, a SCSI Inquiry command is transmitted to the disk device 103 in order from the logical unit number 0. The disk device 103 determines that the specified logical unit number 0 to 2 has a logical unit.
The Inquiry command is reported in response to the Inquiry command. For the Inquiry command in which the logical unit number 3 or more is specified, the absence of the corresponding logical unit is reported in the Inquiry command response. The host computer 301b has two Fiber Chans
For each nel Port, for disk device 103, in order from 0
Send the Inquiry command by specifying the logical unit number. For the specified logical unit numbers 0 and 1, the disk device 103 checks that the logical unit exists.
Report in response to quiry command. For an Inquriry command in which a logical unit number 2 or more is specified, the absence of the corresponding logical unit is reported in a response to the Inquiry command.

In the first embodiment of the present invention, the logical unit correspondence setting unit 231 uses the logical unit shared by the plurality of host computers 301a and 301b and the information unique to the host computers 301a and 301b with high access frequency. Can be assigned to a high-speed logical unit in the disk device 103, and a plurality of host computers 301a, 301
The request for the logical unit shared by b and the logical unit storing information unique to the frequently accessed host computers 301a and 301b can be executed at high speed. The high-speed logical unit in the disk device 103 is
For example, it is a logical unit whose physical position is on the outer periphery of a hard disk drive.

A second embodiment of the present invention will be described with reference to FIG. 3 and FIGS. The second embodiment is an example in which the present invention is applied to a disk array. The disk device 103 shown in FIG. 3 is a disk array in this embodiment. FIG.
FIG. 11 is a diagram showing an internal configuration of a disk array 103 according to a second embodiment of the present invention. In this embodiment, the disk array
103 includes a disk array controller 1001 and seven hard disk drives 1002a, b, c, d, e, f, and g. The physical interface of the disk array controller 1001 and the hard disk drives 1002a-1002g is connected by Fiber Channel, and the logical interface is connected by SCSI. For connection, a loop topology called FC-AL is used. Four FC-AL Loo between disk array controller 1001 and hard disk drives 1002a-1002g
It has p107a, b, c, d. Disk array controller 100
1 is Fiber C for connecting to Loop232a and Loop232b
It has hannelPort 211a and 211b. Four other Fi
Has bre Channel Port. Those Fiber Channel Port
Are connected to different Loops 1007a, b, c, d one by one.
The hard disk drives 1002a, b, and c are connected to Loop 1007a and Loop 1007b, and the hard disk drives 1002d, e, f, and g are connected to Loop 1007c and Loop 1007d.

In this embodiment, the hard disk drive is
An example will be described in which there is a RAID5 RAID group 1003a using five units and a RAID1 RAID group 1003b using two hard disk drives. RAID group 1003a
Is the hard disk drive 1002a, 1002b, 1002c, 1002
d and 1002e, and the RAID group 1003b is composed of R hard disks 1002f and 1002g. Generally, the RAID group 1003a with more hard disk drives 1002a-1002e operating in parallel is better than the RAID group 1003b.
It can be accessed faster than. Also, disk device
103 has four logical units 1004a, b, c, d inside. The logical unit 1004a and the logical unit 1004b exist in the RAID group 1003a, and the logical unit 1004c and the logical unit 1004d exist in the RAID group 1003b. Logical unit
It is assumed that 1004b and logical unit 1004c have the same capacity.

In this embodiment, an example will be described in which the disk device 103 manages the logical units 1004a, b, c, and d by assigning numbers 0, 1, 2, and 3, respectively.

The disk controller 20 according to the first embodiment of the present invention
1 and the host computers 301a and 301b and the disk array controller 1001 of the present embodiment are different from each other in that the disk array controller 1001 uses two Fiber Channel controllers instead of the drive control circuit 208 that directly controls the drives.
It has a troller 1005a and a Fiber Channel Controller 1005b, and controls the hard disk drives 1002a-1002g via the Fiber Channel interface, and the disk array controller 1001 and the hard disk drives 1002a-1002.
This is to perform data transfer between g. The difference between the programs executed by the CPU 203 is that the hard disk drive 1002a-10
Fiber Channel Controller 1005a, 100 that directly controls 02g
The ROM 204 has a hard disk drive 1002a-1002g and a drive I / F control unit 1006 for controlling the data flow between the disk array controller 1001 and the hard disk drive 1002a-1002g. The RAM 205 has a RAID group management information storage area 1008. The RAID group management information storage area 1008 stores information indicating the RAID configuration such as RAID1 and RAID5 for each of the RAID groups 1003a and 1003b, the number of hard disk drives 1002a to 1002e in the RAID group 1003a, the hard disk drives 1002f in the RAID group 1003b, Information for identifying 1002g and the minimum unit for reading and writing data with the hard disk drives 1002a-1002g are stored. The address conversion unit 1008 uses the information stored in the RAID group management information area 1008 to transmit the SCS transmitted by the host computers 103a and 103b.
Request I command for hard disk drive 1002a-1002g
Is converted to a command to generate.

As in the first embodiment, the disk array controller 1001 has an LU management unit 217 in the ROM 204, and has a logical unit number conversion unit 230 and a logical unit correspondence setting unit 231 inside the LU management unit 217. . Further, similarly to the first embodiment, an LU management information area 226 is provided in the RAM 205, and a logical unit number conversion table 102 is provided inside the LU management information area 226. The operations of the LU management unit 217, the logical unit number conversion unit 230, and the logical unit correspondence setting unit 231 are the same as in the first embodiment.

FIG. 9 shows the values stored in the logical unit number conversion table 102 in the second embodiment.

In this embodiment, the logical unit 1004a is a logical unit for storing information unique to the host computer 301a, the logical unit 1004c is a logical unit for storing information unique to the host computer 301b, and the logical unit 1004d.
Is a logical unit shared by the host computers 301a and 301b, and the logical unit 1004b is a logical unit shared by the host computers 301a and 301b.
The case where the logical unit is inaccessible from either of b will be described as an example.

The host computer 301a has a logical unit 1004a
0 and logical unit 1004d1 are numbered 1 and accessed. The host computer 301b has a logical unit number 1004c
0 and logical unit 1004d are numbered 1 and accessed.

On the other hand, in the disk device 103, numbers 0, 1, 2, and 3 are assigned in order from the logical unit 1004a and managed.

The logical unit number 0 of the disk device 103 is associated with the logical unit number 0 of the host computer 301a whose WWN of the Fiber Channel Port is the WWN 302a.
The logical unit number 0 in the disk device 103 is associated with the logical unit number 0 of the host computer 301b whose WWN of the Channel Port is the WWN 302b and stored in the logical unit number conversion table 102. The logical unit number 1 of the host computer 301a and the logical unit number 1 of the host computer 301b are associated with the logical unit number 3 in the disk device 103 and stored in the logical unit number conversion table 102. The logical unit 1004b is numbered and managed inside the disk device 103, but does not correspond to any logical unit number of the host computers 301a and 301b. Also inaccessible.

In this embodiment, the logical unit correspondence setting unit 23
1 is the disk unit 1 corresponding to the logical unit number 0 of the host computer 301b in the logical unit number conversion table 102, for example.
By changing the internal logical unit number of 03 from 0 to 1, the logical unit 1004b that could not be accessed from either the host computer 301a or the host computer 301b becomes a logical unit that stores information unique to the host computer 301b. Settings can be changed.

The logical unit correspondence setting unit 231 monitors the access frequency to the logical unit 1004c, and when the access frequency exceeds a certain threshold value or when one of the hard disk drives 1002f and g constituting the logical unit 1004c fails. Then, when the reliability of the RAID group 1003b is reduced, the data stored in the logical unit 1004c is copied to the logical unit 1004b having the same capacity and high speed. By changing the logical unit number inside the disk device 103 corresponding to the logical unit number 0 of the host computer 301b from 0 to 1, without changing the logical unit management inside the host computer 301, the logical unit number unique to the host computer 301b can be changed. The speed of the logical unit for storing information can be increased. In addition, a decrease in reliability can be prevented.

[0101]

According to the present invention, in a computer system comprising a plurality of host computers and one or a plurality of disk devices having a plurality of logical units therein, a logical unit of a disk device is used for exclusive management. Exclusive management can be performed without using SCSI commands. The host computer can store information unique to the host computer, such as the OS, in the logical unit of the disk device shared by multiple host computers, eliminating the need for a dedicated disk device for the host computer, and reducing the cost of the computer system. Can be achieved.

Further, shared applications and data used by a plurality of host computers can be stored in a logical unit of a disk device shared by a plurality of host computers, thereby enabling centralized maintenance and improving maintainability.

[Brief description of the drawings]

FIG. 1 is a diagram showing a logical unit configuration of a disk device according to a first embodiment of the present invention and a table format of a logical unit number conversion table.

FIG. 2 is a diagram showing a configuration of a disk control unit according to the first embodiment of the present invention.

FIG. 3 is a configuration diagram of a computer system that can be configured by a host computer and a disk device according to the embodiment of this invention.

FIG. 4 is a flowchart of a process in which a logical unit number conversion unit of the embodiment of the present invention converts a logical unit number designated for each SCSI command from a host computer into a logical unit number assigned to a logical unit inside a disk device.

FIG. 5 is a diagram showing a flow of processing in which the disk device of the embodiment of the present invention executes a SCSI command received from a host computer.

FIG. 6 is a diagram showing a page format of a parameter page for notifying a number internally assigned to a logical unit by a host computer by a Mode Select command.

FIG. 7 is a diagram showing the mode number of the disk unit for notifying the number internally assigned to the logical unit by the host computer of this embodiment.
FIG. 4 is a diagram illustrating values of page parameters transmitted by an ect command.

FIG. 8 is a diagram showing an internal configuration of a disk device according to a second embodiment of the present invention.

FIG. 9 shows values stored in a logical unit number conversion table according to the second embodiment of the present invention.

[Explanation of symbols]

103: disk device, 102: logical unit number conversion table, 201: disk control unit, 250: disk drive, 301: host computer

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Naoto Matsunami 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Inside System Development Laboratory, Hitachi, Ltd. (72) Inventor Hideki Kamimaki 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture F-term in Hitachi Systems Development Laboratory 5B014 EB04 GC07 HA11 HB05 HB13 HB26 5B065 BA01 CC02 CC03 ZA15

Claims (13)

[Claims] At least one disk device having a plurality of logical units so that a plurality of host computers share the disk device, wherein the disk device comprises a logical unit number designated by the host computer and the disk device. A logical unit number correspondence storing means for storing a correspondence with a logical unit number of the host computer; a logical unit number converting means for converting the logical unit number designated by the host computer into a logical unit number of the disk device; And a logical unit correspondence setting means for storing the correspondence between the specified logical unit number and the logical unit number of the disk device in the logical unit number correspondence storage means. 2. The disk device according to claim 1, wherein the logical unit number correspondence storage means includes a unique address of the host computer, an identification code of the logical unit number managed by the host computer, and the disk device. And the logical unit number of the disk device. 3. The disk device according to claim 2, wherein
The disk device, wherein the logical unit number conversion means recognizes the address and the identification code input from the host computer and outputs the logical unit number corresponding to the identification code. 4. The disk device according to claim 3, wherein
The logical unit correspondence setting means associates a logical unit number designated by a page parameter transmitted from the host computer with a logical unit number inside the disk device and sets the logical unit number in the logical unit number correspondence storage means. Disk device. 5. The disk device according to claim 4, wherein
The logical unit correspondence setting means allocates a logical unit of the disk device shared by the plurality of host computers to a logical unit which can be written / read at high speed among the logical units. Disk device. 6. The disk device according to claim 4, wherein
The logical unit correspondence setting means assigns a frequently accessed logical unit of the disk device, which stores information unique to the host computer, to a high-speed writable / readable logical unit among the logical units. A disk device characterized by that: 7. The disk device according to claim 4, wherein
The logical unit correspondence setting means changes the logical unit of the disk device corresponding to the logical unit X designated by the host computer A indicated by the logical unit number conversion means from X to Y, whereby any host computer The inaccessible logical unit X is transferred to the host computer A
A disk unit which can be changed to the logical unit Y for storing information unique to the disk unit. 8. The disk drive according to claim 7, wherein
The logical unit correspondence setting means instructs, when the frequency of access from the host computer to the logical unit X exceeds a predetermined number, to copy the contents of the logical unit X to another logical unit, and converts the logical unit number. The logical unit number of the disk device corresponding to the logical unit number X designated by the host computer A indicated by the means is changed from X to Y, whereby information unique to the host computer A is stored in the logical unit Y. A disk device characterized by that: 9. The disk device according to claim 8, wherein
The logical unit correspondence setting means, when a failure occurs in any of the plurality of disk devices, instructs to copy the contents of the logical unit X to another logical unit, and the host indicated by the logical unit number converting means. The logical unit number of the disk device corresponding to the logical unit number X designated by the computer A is changed from X to Y, whereby information unique to the host computer A is stored in the logical unit Y. Disk device. 10. The disk device according to claim 1, wherein the total number of said logical units included in said disk device is equal to or greater than the number of said host computers. 11. The disk device according to claim 1, wherein said disk device constitutes a disk array comprising a plurality of disk devices. 12. A method in which a plurality of host computers share
In a computer system having at least one disk device having a plurality of logical units, a plurality of host computers independently managing the logical units of the disk device, and a plurality of host computers connected to the plurality of host computers via a network A disk device according to claim 1, wherein said disk device is:
Logical unit number correspondence storage means for storing a correspondence between a logical unit number designated by the host computer and a logical unit number of the disk device; and a logical unit of the disk device having the logical unit number designated by the host computer A logical unit number converting means for converting the logical unit number into a number; and a logical unit number correspondence setting for storing the correspondence between the logical unit number designated by the host computer and the logical unit number of the disk device in the logical unit number correspondence storage means. Computer system characterized by comprising means. 13. The computer system according to claim 12, wherein said disk device constitutes a disk array composed of a plurality of disk devices.