JP2000347812A - Information processor and disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable maintenance in arbitrary disk drive units and arbitrary disk module units as to the disk array device wherein disk drives are connected by a fiber channel loop. SOLUTION: The disk array device has the fiber channel loops FC-AL composed of a main loop 6 and a subloop 7 and the individual disk modules 8 are constituted by connecting multiple disk drives 11 through the subloop 7 and a bypass switch 10 of the 2nd kind; and the respective disk modules 8 are connected to a disk adapter 3 through a bypass switch 9 of the 1st kind and the main loop 9 and disconnected from the main loop 9 without disconnecting the main loop 6 by placing the bypass switch 9 of the 1st kind in bypass mode and the disk drives 11 in the respective disk modules 8 are disconnected without disconnecting the main loop 6 by placing the bypass switch 10 of the 2nd kind in bypass mode.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing technology and a disk array technology, and more particularly, to a technology for connecting a large number of disk drives and peripheral devices in a disk subsystem and a disk array device.

[0002]

2. Description of the Related Art As means for connecting a large number of disk drives in a conventional disk array, a disk drive I is used.
A common method is to use SCSI for / F and connect a plurality of SCSI lines to a disk array controller using a crossbar switch. In recent years, a method of using a fiber channel as a disk drive I / F and connecting to a disk array controller using a fiber channel loop has been considered.

[0003]

The above-mentioned fiber channel connection method has an advantage that its high speed and a large number of disk drives can be connected with a small number of lines. However, since the communication lines must form a loop, If no change is made, all disks belonging to the loop cannot be accessed at the time of drive maintenance or the like.

For this reason, in the technology of US Pat. No. 5,768,551, each drive is provided with a fiber channel bypass switch so that the drive can be disconnected without affecting the corresponding loop during drive maintenance. Further, a drive module in which a plurality of drives are connected by a fiber channel loop via a bypass switch is used as a drive module, and further this drive module is connected to a disk array controller via a bypass switch by connecting the entire fiber channel loop. Connection and disconnection of disk modules are possible.

[0005] However, in the technique of the above-mentioned US Pat. No. 5,768,551, by introducing a second bypass switch for a drive module, connection by module can be performed.
Although it is possible to disconnect, but because it simply introduces a bypass switch for the disk module, it does not mention how to connect the disk module bypass switch to the entire fiber loop,
There is a technical problem that any disk module cannot be separated without affecting the operation of the entire fiber loop.

[0006] Further, since it does not mention how to arrange an ECC group in a disk array with respect to a disk module group, an arbitrary disk module group cannot be actually separated during operation of the disk array device. , There was a technical problem.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a disk array system for connecting a large number of disk drives using a fiber channel loop while maintaining the connection state by the fiber channel loop and arranging any disk module or disk drive during operation of the system. It is to provide a technique that enables selective disconnection and reconnection.

Another object of the present invention is to provide a disk array device in which a large number of disk drives are connected by using a fiber channel loop, while maintaining the connection state by the fiber channel loop and arbitrarily operating a disk drive or disk during operation of the device. It is an object of the present invention to provide a technology that enables maintenance work of a module.

Another object of the present invention is to control a bypass switch for connecting a disk drive unit or a disk module unit to a fiber channel loop in a disk array device that connects a large number of disk drives using a fiber channel loop. Is to provide a technology that enables a simple configuration.

[0010]

SUMMARY OF THE INVENTION The present invention relates to an information processing apparatus for connecting peripheral devices to an input / output control device by means of a fiber channel loop, wherein the fiber channel loop for connecting a plurality of peripheral devices includes a plurality of peripheral devices. Consists of a main loop for connecting a peripheral device group consisting of an I / O control device and a sub-loop for connecting a plurality of peripheral devices in the peripheral device group, and a sub-loop of each peripheral device group corresponds to each. The first type bypass switch is connected to the main loop, and when the first type bypass switch is in the through mode, the sub-loop of the corresponding peripheral device group is connected so as to form a part of the main loop. When the bypass switch goes into bypass mode, only the secondary loop of the corresponding peripheral device group is selectively disconnected from the primary loop. The main loop is obtained by configured to be maintained operations loop via the one bypass switch in a state where the corresponding peripheral device group is disconnected.

In this information processing apparatus, each peripheral device constituting each peripheral device group is connected to a sub-loop in the peripheral device group via a corresponding type II bypass switch, and is connected to a type II bypass switch. Is in the through mode, the corresponding peripheral devices are connected so as to form a part of the sub-loop and the main loop, and when the type 2 bypass switch is in the bypass mode, only the corresponding peripheral devices are connected. The auxiliary loop is selectively disconnected from the sub-loop in the peripheral device group, and the sub-loop and the main loop are connected so that the operation loop can be maintained via the second type bypass switch even when the peripheral device is disconnected.

Further, in the disk array device of the present invention,
Provide a Fiber Channel bypass switch for each disk drive. A drive module in which a plurality of drives are connected by a fiber channel loop via a bypass switch is referred to as a drive module. Then, when a bypass switch connecting this drive module to the disk array controller (that is, the entire Fiber Channel loop) is set to the bypass mode, only the corresponding disk module group becomes the entire Fiber Channel loop. And the entire Fiber Channel loop is connected to the entire Fiber Channel loop so that its operation loop can be maintained via the corresponding bypass switch. These bypass switches are controlled by a bypass switch control device. Also, the same EC in the disk array
The disk drives constituting the C group are set to be selected from different disk drive groups.

As described above, by selectively stopping any disk drive or disk module during operation of the apparatus or by selectively disconnecting / reconnecting the entire fiber channel loop while maintaining the connection state of the fiber channel loop. Maintenance can be performed.

[0014]

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

FIG. 1 is a conceptual diagram showing an example of the configuration of a disk array device as an embodiment of the information processing device of the present invention. In the present embodiment, the disk array device
Host adapter 1, cache memory 2, disk adapter 3, disk drive 11 for storing data
And a disk processor 8 which is connected by a plurality of fiber channels, and a service processor 12 which monitors the state of these parts and serves as an interface with maintenance personnel.
(SVP12). The host adapter 1 and the cache memory 2 are connected by a data bus 4, and the cache memory 2 and the disk adapter 3 are connected by a data bus 5. A fiber channel arbitrated disk is provided between the disk adapter 3 and the plurality of disk modules 8.
They are connected by a fiber channel loop FC-AL such as a loop.

The host adapter 1 has a host I / F circuit, and performs channel I / F protocol control and data transfer. The cache memory 2 is a temporary data buffer when transferring data between the host adapter 1 and the disk adapter 3, and stores read data from the disk drive 11 and accepts a data read request for the same area again. The disk drive 11
It is also used for the purpose of increasing the processing speed by transferring the data in the cache memory 2 to the host without reading the data from the host. The disk adapter 3 is connected to the disk module 8 by a fiber channel loop FC-AL, and reads and writes data from and to each disk drive 11.

In the case of this embodiment, the disk adapter 3
A fiber channel loop FC-AL connecting the disk drive 11 and the disk drive 11 includes a main loop 6 connecting the disk adapter 3 and the disk module 8 and a sub-loop connecting the plurality of disk drives 11 in each disk module 8. The sub-loop 7 is connected to the main loop 6 by a type 1 bypass switch 9 and is connected to the main loop 6.
A part of. By setting the first type bypass switch 9 to the bypass mode, the sub-loop 7 can be separated from the main loop 6 without cutting the main loop 6.

The disk drive 11 in each disk module 8 is connected to the sub-loop 7 in the disk module 8 by a second type bypass switch 10 and forms a part of the sub-loop 7. When the type 1 bypass switch 9 corresponding to each disk module 8 is set to the connection side (through mode), the disk drive 11 also becomes a part of the main loop 6. Also, as in the case of the disk module 8, the second type bypass switch 10
Is set to the bypass mode, the corresponding disk drive 11 can be disconnected without disconnecting the sub-loop 7 and the main loop 6. As described above, in the disk array device of the present embodiment, disconnection of an arbitrary disk drive 11 or disk module 8 can be performed without affecting the operation of the main loop 6 of the fiber channel.

The SVP 12 is connected to the bypass switch control means 14 via the maintenance bus 13, and the bypass switch control means 14 is connected to the individual first type bypass switch 9 and second type bypass switch 9 via the bypass switch control bus 15. 10 is connected.

When the number of disk drives 11 is increased or decreased, S
By instructing the bypass switch control means 14 from the VP 12 to control the switching of the second type bypass switch 10 between the bypass mode and the through mode, the disk drive 11 is disconnected and connected without affecting the operation of the main loop 6. .

On the other hand, in the case of increasing / decreasing the disk module, the SVP 12 sends the bypass switch control means 1
By instructing the first type bypass switch 9 to switch the bypass mode / through mode of the first type bypass switch 9, the number of disk modules can be increased or decreased without affecting the operation of the main loop 6.

Each disk module 8 has a spare disk drive 11-S used in the event of a failure of the disk drive 11, in addition to the disk drive 11 normally used.

The disk adapter 3 is a disk drive 1
1 can be monitored, and when a failure occurs in a certain disk drive 11, the access destination is automatically set to the spare disk drive 11-S in the disk module 8 to which the failed disk drive 11 belongs. Switch to and continue operation. In this case, the disk adapter 3 issues an instruction to the bypass switch control means 14 through the maintenance bus 13 and automatically disconnects the failed disk drive from the fiber channel loop (sub-loop 7).

Each disk drive 11 does not necessarily need to be a magnetic disk drive, but may be constituted by an arbitrary rotary storage device such as an optical disk drive.

FIG. 2 is a conceptual diagram showing an example of the configuration of a bypass switch used in the disk array device of the present embodiment.

The bypass switch 16 (each of the first type bypass switch 9 and the second type bypass switch 10) of the present embodiment includes a bus I / F 17 with the bypass switch control bus 15, a mode setting register 18, and a selector circuit 19. , Driver circuit 20, port 0 input 21, port 0 output 22, port 1 input 23, port 1 output 24, and various signal lines connecting these. In normal use, the mode setting register 18 is set to the through side, and the selector circuit 19 selects the 0 side. Therefore, the port 0 input 21 becomes the port 1 output 24
, The port 1 input 23 is connected to the port 0 output 22, and the bypass switch 16 is in a through state (through mode).

On the other hand, when switching to the bypass mode, an instruction to switch to the bypass mode is issued via the bypass switch control bus 15. The bus I / F 17 receives the above instruction and sets the mode setting register 18 to the bypass mode. When the mode setting register 18 is set to the bypass mode, the selector circuit 19 selects the 1 side, and the signal input to the port 0 input 21 is output to the port 0 output 22 as it is. Reference numeral 16 indicates a bypass state (bypass mode).

FIG. 3 shows an embodiment of a method of setting an ECC group in a disk array when a RAID system is constructed with the disk array device of the present embodiment. In the present embodiment, as an example, the disk drive unit includes five disk modules 8 (25, 26, 27, 2) each containing nine disk drives 11 in one housing.
8, 33). However, the disk module 33 is a spare disk module and is not normally used.

The individual disk modules 25, 2
6, 27, 28 and spare disk module 3
3 has a spare disk 32 (DISK0S-DI
SK3S) exist one by one. This spare disk 3
2 is not normally used.

In this embodiment, the write data from the host CPU is divided into three parts and the three disk drives 11
Are written separately. An ECC group 31 is composed of a data block 29 composed of three blocks of the write data and a parity data block 30 calculated from the three data blocks 29.

The RAID system of FIG. 3 exemplified in this embodiment is exemplified for the RAID level 4 in which a parity block is arranged in a specific disk module for simplicity. Not limited to this, RAI
D level 1 (mirrored disk), RAID level 2
Needless to say, the present invention can be applied to any of the following three levels: level 3 (bit-wise data distribution / interleaving) and RAID level 5 (parity data is mixed with data blocks and stored in an arbitrary disk).

The ECC group 31 for storing the data divided by the RAID system is set so as to select one disk drive 11 from each of the disk modules 25 to 28. By setting the ECC group 31 in this way, even if one of the disk modules 25 to 28 becomes inoperable, the necessary remaining data is reproduced from the data readable from the disk drive 11 in the remaining disk modules. It is possible to do so without affecting the operation of the disk array device. The method of restoring the missing data from the parity and the remaining data is a basic method of RAID and will not be described here.

Similarly, without affecting the operation of the disk array device, one disk module 25 to
28 (and spare disk module 33) as the first
Main loop 6 by seed bypass switch 9 (34, 35)
, And can be increased or decreased in disk module units.

FIG. 4 shows the EC of the disk drive 11 managed by the disk adapter 3 when a disk drive failure occurs in the disk array device of this embodiment.
It shows an example of how the assignment information table for the C group 31 changes.

Hereinafter, a recovery operation when one disk drive has failed will be described with reference to FIGS. In the present embodiment, DISK 10 is initially ECC G
It is assigned to data stripe 1 of r0 (D1 in FIG. 4). Assuming that a failure has occurred in the disk drive (DISK 10), the disk adapter 3 that has detected the failure immediately sends an ECC group 3 of the disk drive.
1, the spare disk (DISK1S) in the same disk module 26 is allocated to the portion where the failed disk was located. Thereafter, the disk adapter 3 uses the spare disk (DISK1S) as the disk drive of the data stripe 1 (D1 in the figure) of the ECC Gr0 and continues to operate.

The disk adapter 3 issues an instruction to the bypass switch control means 14, switches the type 2 bypass switch 10 corresponding to the failed disk drive (in this example, DISK 10) to the bypass mode, and disconnects it from the sub-loop 7.

FIG. 5 shows the ECC of the disk drive managed by the disk adapter 3 when a failure occurs in the disk module in the disk array device of the present embodiment.
It shows an example of how the assignment information table for the group 31 changes.

Hereinafter, a recovery operation when one disk module fails will be described with reference to FIGS. In this embodiment, the disk module 26 (DI
SKMODULE1) is allocated to the data stripe 1 (D1 in FIG. 5) of the ECC Gr. Assuming that a failure has occurred in the disk module 26 (DISK MODULE 1), the disk adapter 3 that has detected the failure immediately sends an ECC group 3 of the disk drive.
1, the spare disk module 33 is allocated to the data stripe 1 (D1 in FIG. 5) of the ECC Gr. Thereafter, the disk adapter 3 performs the data stripe 1 of the ECC Gr (FIG. 5).
The spare disk module 33 is used as the disk drive of D1), and the operation is continued.

Note that the disk adapter 3 issues an instruction to the bypass switch control means 14 to set the first type bypass switch 9 corresponding to the failed disk module to the bypass mode and to place the failed disk module (in this case, the disk module 26) in the bypass mode. Disconnect from the main loop 6.

Prior to the disconnection, data of a healthy disk drive among a plurality of disk drives constituting the failed disk module may be copied to a corresponding disk drive in the spare disk module 33 of the switching destination. it can. As a result, the data restoration by the RAID technology can be performed only with the data of the failed disk drive, so that the load on the disk adapter 3 in the recovery operation can be reduced and the time required for the data restoration can be shortened.

The configuration of the bypass switch 16 (the first type bypass switch 9 and the second type bypass switch 10) in the disk array device of the present embodiment is as shown in FIG. The disk adapter 3 is not limited to the configuration in which the mode switching control is externally performed by using the control I / F in the fiber channel loop FC-AL as shown in FIG. May be executed.

That is, the bypass switch 16A (the first type bypass switch 9 and the second type bypass switch 10) exemplified in FIG. 6 has its own FC-AL interface 17A. Like the drive 11 and the like, it is recognized as one of the input / output devices connected to the fiber channel loop FC-AL (main loop 6, sub-loop 7). Then, the disk adapter 3 can control the switching of each mode of bypass / through in the bypass switch 16A by the command interface by the fiber channel loop FC-AL.

The bypass switch 16A having the configuration shown in FIG.
Is used as the first type bypass switch 9 and the second type bypass switch 10, an extra configuration such as the bypass switch control bus 15 is not required as a control interface, and a simpler configuration, that is, low cost Thus, control such as selective disconnection / reconnection of disk modules and disk drives in the main loop 6 and the sub-loop 7 can be performed.

As described above, according to the disk array device of the present embodiment, the operation of the fiber channel loop FC-AL such as the main loop 6 and the sub-loop 7 is affected during the operation of the disk array device. Without any disk drive 11 or disk module 8
Can be selectively stopped, or selectively disconnected from the entire fiber channel loop FC-AL, or reconnected to perform maintenance. As a result, for example, an effect is obtained that a failed disk drive or a disk module can be accurately replaced in a RAID system without stopping.

Further, even if a failure occurs in any disk drive 11 or disk module 8, the other disk elements 11 are not affected by the failure by selectively disconnecting the disk drive 11 or disk module 8 which has failed. Has the effect that the operation can be continued.

Also, by providing the FC-AL interface function to the bypass switch and controlling the bypass / through mode by recognizing the bypass switch as one of the FC-AL connected devices, the control system of the bypass switch can be simplified. The manufacturing cost of the disk array device can be reduced.

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

For example, in the description of the above embodiment,
As an example of the information processing device, a case where the present invention is applied to a disk array device using a disk drive group as a peripheral device has been exemplified, but the present invention can be widely applied to a general configuration in which a large number of peripheral devices are connected by a fiber channel loop. .

[0049]

According to the disk array device of the present invention,
In a disk array device that connects a large number of disk drives using a fiber channel loop, selectively disconnecting and reconnecting any disk module or disk drive during operation of the device while maintaining the connection state by the fiber channel loop. be able to,
The effect is obtained.

Further, according to the disk array device of the present invention, in a disk array device that connects a large number of disk drives using a fiber channel loop, while the connection state by the fiber channel loop is maintained, an arbitrary The advantage is that the maintenance management work of the disk drive or the disk module can be performed.

According to the disk array device of the present invention, in a disk array device connecting a large number of disk drives using a fiber channel loop, a bypass switch for connecting to a fiber channel loop in disk drive units or disk module units Can be achieved with a simple configuration.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram illustrating an example of a configuration of a disk array device that is an embodiment of an information processing device according to the present invention.

FIG. 2 is a conceptual diagram showing an example of a configuration of a bypass switch used in a disk array device according to an embodiment of the present invention.

FIG. 3 is a conceptual diagram showing an example of a method of setting an ECC group in a disk array when a RAID system is constructed with a disk array device according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating an example of a process performed when a failure occurs at a disk drive level in a disk array device according to an embodiment of the present invention;
It is explanatory drawing which shows the transition example of the assignment table of CC group.

FIG. 5 is an explanatory diagram showing a transition example of an ECC group assignment table when a failure occurs at a disk module level in a disk array device according to an embodiment of the present invention.

FIG. 6 is a conceptual diagram showing a modification of the configuration of the bypass switch used in the disk array device according to the embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Host adapter, 2 ... Cache memory, 3 ... Disk adapter (input / output control device), 4 ... Data bus, 5
... Data bus, 6 ... Main loop, 7 ... Sub loop, 8 ... Disk module (peripheral device group), 9 ... Type 1 bypass switch, 10 ... Type 2 bypass switch, 11 ...
Disk drive (peripheral device), 11-S spare disk drive, 12 service processor, 13 maintenance bus, 14 bypass switch control means, 15 bypass switch control bus, 16 bypass switch, 16
A: bypass switch, 17: bus I / F, 17A: F
C-AL interface, 18 mode setting register,
19 selector circuit, 20 driver circuit, 21 port 0 input, 22 port 0 output, 23 port 1 input, 24 port 1 output, 25 to 28 disk module, 29 data block, 30 parity data Block, 31: ECC group, 32: Spare disk, 33: Disk module, FC-AL: Fiber channel loop.

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B065 BA01 CA01 CA30 5D066 BA02 BA05 5K031 CB12 DA04 DA19 DB10 DB14 EB03 EB09 EC04

Claims (5)

[Claims] 1. An information processing apparatus for connecting a peripheral device to an input / output control device by a fiber channel loop, wherein the fiber channel loop for connecting a plurality of the peripheral devices includes a plurality of the peripheral devices. A main loop for connecting a device group to the input / output control device, and a sub-loop connecting a plurality of the peripheral devices in the peripheral device group, wherein the sub-loop of each of the peripheral device groups is The main loop is connected to the main loop through a corresponding type 1 bypass switch, and when the type 1 bypass switch is in the through mode, the sub-loop of the corresponding peripheral device group forms a part of the main loop. And when the type 1 bypass switch is in the bypass mode, the sub-group of the corresponding peripheral device group is Only the main loop is selectively disconnected from the main loop, and the main loop is configured to be able to maintain the operation loop via the first type bypass switch even when the corresponding peripheral device group is disconnected. Information processing device. 2. The information processing apparatus according to claim 1, wherein each of the peripheral devices constituting each of the peripheral device groups is connected to the sub-loop in the peripheral device group via a corresponding type II bypass switch. Connected to the second
When the type bypass switch is in the through mode, the corresponding peripheral device is connected to form a part of the sub-loop and the main loop, and when the type 2 bypass switch is in the bypass mode, Only the corresponding peripheral device is selectively disconnected from the sub-loop in the peripheral device group, and the sub-loop and the main loop operate via the second type bypass switch even when the peripheral device is disconnected. An information processing device configured to maintain a loop. 3. A disk array device for connecting a disk drive to a disk array control device by a fiber channel loop, wherein said fiber channel loop for connecting a plurality of said disk drives comprises a disk comprising a plurality of said disk drives. A main loop for connecting a module to the disk array controller; and a sub-loop for connecting a plurality of the disk drives in the disk module. The sub-loop of each disk module corresponds to a corresponding one of the sub-loops. Connected to the main loop via a type 1 bypass switch, and when the type 1 bypass switch is in the through mode, the sub-loop of the corresponding disk module is connected to form a part of the main loop;
When the first type bypass switch is in the bypass mode, only the sub-loop of the corresponding disk module is selectively disconnected from the main loop, and the main loop is connected to the disk module even when the corresponding disk module is disconnected. A disk array device configured to maintain an operation loop via a type 1 bypass switch. 4. The disk array device according to claim 3, wherein each of said disk drives constituting each of said disk modules is connected to said sub-loop in said disk module via a corresponding type 2 bypass switch. When the second type bypass switch is in the through mode, the corresponding disk drive is connected so as to form a part of the sub loop and the main loop, and the second type bypass switch is in the bypass mode. In this case, only the corresponding disk drive is selectively disconnected from the sub-loop in the disk module, and the sub-loop and the main loop are connected to the second type bypass switch even when the disk drive is disconnected. Connected so that the operation loop can be maintained via The disk array device according to. 5. The disk array device according to claim 3, wherein an ECC group of the disk array is a first configuration that is set orthogonal to a plurality of the disk modules, and an ECC group of the disk array is configured. A second configuration in which the plurality of disk drives are arranged in different disk modules respectively; a bypass switch that controls switching between the through mode and the bypass mode in at least one of the first type and the second type bypass switch; A third configuration having control means, wherein each of the first type and second type bypass switches has a connection interface to the fiber channel loop, and is configured to switch between the through mode and the bypass mode via the fiber channel loop. Fourth structure for controlling switching The disk array device, which comprises at least one.