JPH0816328A - Disk array system - Google Patents

Abstract

PURPOSE:To improve reliability by correcting/restoring data to be in a physical disk device forming a redundant group even when one or all these physical disk devices break down by composing a disk array of two hierarchies. CONSTITUTION:When a read instruction is commanded from a host computer 1 to a master disk array controller 5 in the state of breaking down a physical disk device 41-1, a read instruction is normally commanded from the master disk array controller 5 to respective logical disk devices 21-2n. A slave disk array controller 31 of the logical disk device 21 equipped with the broken clown physical disk device 41-1 among the logical disk devices 21-2n receiving the read instruction reads data from not broken down physical disk devices 41-2 to l 41-n. The other logical disk devices 22-2n perform reading, transfer data to the controller 5 and transfer those data excepting for redundant data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array system including a large number of disk devices.

[0002]

2. Description of the Related Art In a conventional disk array system, in order to obtain high reliability, a plurality of physical disk devices are arranged in an array to form a redundancy group, and data in the redundancy group is parity-checked. Data redundancy is secured by adding data and the like. Therefore, even if one physical disk device in the redundancy group fails, it has a function of correcting / restoring the data of the failed disk device from the data in the other physical disk device of the redundancy group. Such a disk array system is disclosed in, for example, JP-A-2-236714 and A Ca.
se for Redundant Array of
Inexpensive Disks (University of California Report No. UCB / CSD87 / 391, disclosed December 1987) and the like.

Further, as a technique for realizing a more highly reliable system, two physical disk devices in which parity data is recorded are provided in a redundancy group and two physical disk devices in the redundancy group are used. For example, Japanese Patent Laid-Open No. 4-287255 discloses a device having a function of correcting / restoring data of a failed physical disk device even if it fails.

As another different system, standby data is further added to the redundancy group, and when one physical disk device in the redundancy group fails,
By providing the standby data with the function of writing the data restored from the other physical disk device, the function to correct / restore the data even if two physical disk devices in the redundancy group fail Is described, for example, in Japanese Patent Laid-Open No. 5-100960.

As a further different system, by adding p data and q data in the redundancy group, there is a function of correcting / restoring the data even if two physical disk devices in the redundancy group fail. What is possessed is described in, for example, JP-A-2-236714.

FIG. 9 is a block diagram showing the configuration of a conventional disk array system. In FIG. 9, the disk array controller 22 executes commands from the host computer 21 and stores write data sent thereto. The host computer 21 and the disk array controller 22 are connected by an internal bus 23 that exchanges commands and write data from the host computer 21 with reports and read data from the disk array controller 22.

The disk array controller 22 includes a plurality of physical disk device groups 2 forming a redundancy group 24.
The function of controlling 51 to 25n in parallel and the redundancy group 24
When a certain limited number of disk devices have a failure, it has a function of correcting / restoring the data in the failed disk device.

[0008]

Since the conventional disk array system is configured as described above, when a limited number or more of physical disk devices in the redundancy group 24 fail at the same time, the failure occurs. There is a problem that it is impossible to correct / restore the data in the physical disk device.

Further, when the disk array controller 22 for controlling the physical disk devices in the redundancy group in parallel and correcting / restoring the data in the failed physical disk device in the redundancy group fails, the data is restored. There was a problem that was impossible.

The present invention has been made to solve the above-mentioned conventional problems. A first object of the present invention is to make a failure when one to all physical disk devices in a redundancy group fail. Disk array controller for correcting / restoring the data of a failed disk unit and controlling / recovering the data of a failed physical disk unit in the redundancy group in parallel by controlling the disk units in the redundancy group in parallel It is to improve the reliability of the disk array system by enabling the data restoration even when the disk fails.

A second object of the present invention is to provide a logical disk device dedicated to data restoration, and when the logical disk device fails, restore the restored data to the logical disk device dedicated to data restoration. By storing the data, it is possible to restore the data of the failed disk device at high speed and improve the reliability.

A third object of the present invention is to provide two or more master disk array controllers and continue the operation of the disk array system even when the master disk array controller fails to improve reliability.

A fourth object of the present invention is to multiplex an external bus connecting a logical disk device and a master disk array controller so that the operation of the disk array system is continued even if the external bus fails. It is to improve the reliability of the system.

[0014]

According to a first aspect of the present invention, there is provided a disk array system which is arranged in an array form.
A plurality of physical disk devices that make up a redundant group of layers, a function of controlling the plurality of physical disk devices in parallel, and redundancy, and a limited number of devices in the redundancy group due to the redundancy. When a physical disk device of the above fails, the master disk is provided with a logical disk device including a slap disk array control having a function of correcting / restoring the data in the failed physical disk device. The array controller has a function of controlling in parallel a plurality of the logical disk devices forming a second-layer redundancy group by arranging them in an array and having redundancy, and by the redundancy, a certain limit in the redundancy group is provided. When a specified number of logical disk devices fail, the function to correct / restore the data in the failed logical disk device is provided. That.

According to a second aspect of the present invention, there is provided a disk array system comprising a logical disk device dedicated to restoration for storing data of the failed logical disk device when the logical disk device fails. It is a thing.

According to a third aspect of the present invention, in the disk array system, the logical disk device and the master disk array controller are connected by an external bus, and the master disk array controller and the host computer are connected by an internal bus. Two or more master disk array controllers are connected to the external bus and the internal bus.

A disk array system according to a fourth aspect of the present invention has two or more external buses for connecting the logical disk device and the master disk array controller, and each of the external buses has the master disk array. Two or more controllers are connected and the two or more master disk array controllers are connected to the same internal bus.

[0018]

A disk array system according to the present invention is a slave disk which controls a physical disk device when a limited number of physical disk devices in a logical disk device fails. The array controller restores the data in the failed physical disk device, and if the number of physical disk devices in the logical disk device cannot be restored by the slave disk array controller, the master When the disk array controller restores the data in the failed physical disk device from another logical disk device, even if one to all of the physical disk devices forming the redundancy group have failed, the failure has also occurred. It is possible to correct / restore the data in the disk device.

In addition, the slave disk array controller in the logical disk device restores the data in the failed logical disk device by using the data in the other logical disk device by the data recovery function using the redundant data. Will be able to do.

In the disk array system according to the second aspect of the present invention, a logical disk device dedicated to data restoration is provided so that the data restored by the logical disk device can be restored to this logical disk device. By storing the data in the disk, data can be restored at high speed by using the logical disk device dedicated to the data restoration.

According to a third aspect of the present invention, in a disk array system, a master disk array controller and a plurality of logical disk devices are connected by an external bus, and a master disk array controller and a host computer for issuing a command to the master disk array controller. Are connected by an internal bus, and two or more master disk array controllers are connected to the same external bus and the same internal bus, so that when the master disk array controller in use fails, the host computer can use another master The disk array controller can be used, and the operation of the disk array system can be continued even if the master disk controller fails.

According to a fourth aspect of the present invention, a disk array system has two or more external buses for connecting a logical disk device and a master disk array controller, and two or more master disk array controllers are connected to the external buses. In addition, by connecting two or more master array controllers to the same internal bus, when the master disk array controller notices a failure of the external bus in use, the external disk can be used by using another available external bus. The operation of the disk array system can be continued even if the bus fails.

[0023]

【Example】

Example 1. Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of the disk array system of the present invention, which is a host computer 1.
Issues an input / output command and write data to the master disk array controller 5. The master disk array controller 5 has a function to store commands and executions from the host computer 1 and write data sent thereto, a data striping technique, a redundancy operation function, and a correction / restoration function. And an internal bus 8 are connected.

The master disk array controller 5 and the respective logical disk devices 21 to 2n are external buses 6
Connected by. The logical disk devices 21 to 2n are slave disk array controllers 31 to 3n and physical disk devices 41-1 to 41-n, 42-1 to 42-n, 4
Redundancy groups 41 to 4 in the first layer, which are composed of n-1 to 4n-n
n, and the logical disk units 21 ...
2n form a second hierarchical group 7. The slave disk array controllers 31 to 3n have the functions of executing the commands from the master disk array controller 5, storing the write data sent, and correcting / restoring the data striping technology and redundancy operation function. There is.

2 and 3 show three logical disk devices 2
1, 22, 23, and one of them is a logical disk device, for example, 21 is three physical disk devices 41-1, 41-2,
41 shows the data arrangement of a logical disk device and the data arrangement of a physical disk device when configured by 41-3.

The data of the logical disk device group has redundancy, and in this case, P0 is the exclusive OR of D0 and D1 and is generated by the master disk array controller 5 of FIG. The data of the logical disk device 21 is the physical disk device 41-that constitutes the logical disk device.
It is stored in 1, 41-2, 41-3. PP0, which is redundant data with respect to the continuous data P0, D2, and D4 of the logical disk device 21, is the data generated by the slave disk array controller 31 forming the logical disk device 21.

In this way, the continuous data of the logical disk device 21 is stored in each physical disk device 41-1 to 41-3.
, Redundant data is added, and the data is distributed and stored.

2 and 3, let us consider a case where a write command for data D0 to D11 is issued from the host computer 1 of FIG. 1 to the master disarray controller 5. The master disk array controller 5 that has received the write command generates redundant data P0 with D0 and D1, and similarly generates redundant data P1, P2, P3, P4, P5, and logically with respect to the striped data. A write command is issued to the slave disk array controllers 31 to 33 of the appropriate disk devices 21 to 23.

For example, the slave disk array controller 31 constituting the logical disk device 21 is issued with a continuous write command of data of P0, D2, D4, P3, D8 and D10.

Logical disk device 2 that received this command
Slave disk array controller 31 of 1 is P0,
The redundant data PP0 is generated from D2, and PP1 and P are similarly generated.
Generate P2. These data are striped and written to the physical disk device 41-1, the physical disk device 41-2, and the physical disk device 41-3.

For example, in the physical disk device 41-1, P0, D4 and PP2 are written as continuous data.

Also in the other logical disk devices 22 and 23, the physical disk devices 42-1 to 42-3 and 43-1 to the same method as the slave disk array controller 31 of the logical disk device 21. The data of 43-3 is written.

2 and 3, let us consider a case where the host computer 1 issues a data read command of D0 to D11 to the master disk array controller 5. The slave disk array controllers 31 to 33, which have received the read command, issue the read command to the striped logical disk devices 21 to 23, including the parity data added by the master disk array controller 5.

For example, the slave disk array controller 31 constituting the logical disk device 21 is issued with a continuous read command of data of P0, D2, D4, P3, D8 and D10.

Logical disk device 2 that received this command
The slave disk array controller 31 of 1 stores the striped data including the added redundant data in the physical disk device 41-1, the physical disk device 41-.
2. Issue a read command to the physical disk device 41-3.

Slave disk array controller 31
Is P0, D2, D4, P3, D8, D10 which is data obtained by removing redundant data PP0, PP1, PP2 from the data read from each physical disk device 41-1 to 41-3.
Data is transferred to the master disk array controller 5 as continuous data in this order.

Similarly, the logical disk device 22 also has D0,
Data is transferred to the master disk array controller 5 as continuous data of P1, D5, D6, P4, and D11, and the logical disk device 23 is similarly D1, D3, and P.
Data is transferred as continuous data in the order of 2, D7, D9, and P5.

The mask disarray controller 5 converts the data read from each of the logical disk devices 21 to 23 into redundant data P0, P1, P2, P3, P4,
Except P5, it is transferred to the host computer 1 as continuous data from D0 to D11.

Next, physical disk devices 41-1 to 41-3
The case where the unit fails will be described.

When the physical disk device 41-1 has failed, the master disk array controller 5 issues a read command for data D0 to D11 from the host computer 1.
, The read command from the master disk array controller 5 to each of the logical disk devices 21 to 23 is executed as usual.

Of the logical disk devices 21 to 23 that have received this read command, the physical disk device 4 that has failed
The slave disk array controller 31 of the logical disk device 21 including 1-1 includes the data D2, PP1, D8, PP0, D3 from the physical disk device 41-2 and the physical disk device 41-3 which have not failed. Read the data of D10. Failed physical disk device 41-1
Data P0 from D2 and PP0, D4 from PP1 and D3
Can be restored from. The data of PP2 is not reproduced because it is not included in the data requested by the master disk array controller 5. The slave disk array controller 31 including P0, D2, D4, P
The data of 3, D8 and D10 are transferred to the master disk array controller 5.

The other logical disk device 22 and logical disk device 23 perform a normal read operation and transfer necessary data to the mask disk array controller 5. After the master disk array controller 5 has all the necessary data including redundant data, the redundant data P
The data of D0 to D11 excluding 0 to P5 is transferred to the host computer 1.

Next, a description will be given of the execution of a command when the number of physical disk devices in one logical disk device fails such that the slave disk array controller cannot restore data.

It is assumed that the physical disk device 41-1, the physical disk device 41-2, and the physical disk device 41-3 have failed. Host computer 1 to D0 to D1
The master disk array controller 5 which has received the read command for the data 1 issues a read command to each of the logical disk devices 21 to 23. Logical disk device 2
1 notifies the read command from the master disk array controller 5 that the data cannot be restored.

The master disk array controller 5 is another logical disk unit 22 and logical disk unit 2
D0, P1, D5, D6, P4, D read from 3
11 and D1, D3, P2, D7, D9, P5, P0, D2, D4, P3, D8 of the logical disk device 21.
The data of D10 is restored, and the data of D0 to D11 is transferred to the host computer 1. In this way, the physical disk device 41-1 that constitutes the logical disk device 21 is
It is possible to correct / restore the data even if one to all 41-3 fail.

When the logical disk device 21 fails and the slave disk array controller 31 fails, the master disk array controller 5 is notified that the slave disk array controller 31 cannot read data due to its own failure. If the slave disk array controller 31 cannot report its own failure, the master disk array controller 5 has determined that the failure has destroyed the number of disk devices in the logical disk device in which data cannot be restored. Data is restored by the same function as that of the case, and the data is transferred to the host computer 1. In this way, the logical disk device 2
Slave disk array controller 31 that composes 1
Data can be corrected / restored even if the device fails.

Example 2. FIG. 4 is a block diagram showing a configuration of a disk array system including a logical disk device 14 for restoration including a slave disk array controller for restoration 15 and a redundancy group 16 for restoration. Are denoted by the same reference numerals and redundant description will be omitted. In the second embodiment, when there is no failure in any of the logical disk array systems, the master disk array controller 5 does not use the logical disk device 14 for restoration and executes the normal processing as described above. Run.

Next, when the logical disk device fails, the master disk array controller 5 restores the data of the failed logical disk device by the above procedure, and restores the restored data to the logical disk for restoration. It is stored in the device 14. For example, when the logical disk device 21 in FIGS. 2 and 3 fails, as shown in FIG.
0, D2, D4, P3, D8, D10 ... Are stored in the logical disk device 14 for restoration.

After the restoration is completed, the master disk array controller 5 reads and writes data by using the restoration logical disk device 14 instead of the logical disk device 21 before the failure. When one more logical disk device fails, the master disk array controller 5 restores data from other logical disk devices including the logical disk device for restoration by the method described above.

As described above, by providing the logical disk device 14 for data restoration and storing the restored data, the data restoration can be performed at high speed in the disk array system.

Example 3. In FIG. 6, the master disk array controller 5 is connected to an internal bus 8 connected to the host computer 1 and an external bus 6 connected to a logical disk device.
FIG. 2 is a block diagram showing a configuration of a disk array system equipped with two or more 1 to 5n, and the same parts as those in FIG. In the third embodiment, as a simple example, a case where the master disk array controller 5 is two will be described as an example.

In normal use, the host computer 1 commands input / output to the master disk array controller 51. Master disk array controller 51
If the host computer 1 reports its own failure to the host computer 1 or if there is no termination report from the master disk array control 51 within a limited time, the host computer The computer 1 issues an input / output command to the master disk array system 52. Since the master disk array controller 52 has the same function as the master disk array controller 51, the host computer 1 has the logical disk devices 21 to 2n in each logical disk device.
It becomes possible to access the data of.

Example 4. FIG. 7 shows a logical disk device 2
1 to 2n are respectively connected to two or more external buses 61 to 6n, and two or more master disk array controllers 51 to 5n are connected to the two or more external buses 61 to 6n.
FIG. 2 is a block diagram showing a configuration of a disk array system connected to the same. The same parts as those in FIG. In the fourth embodiment, as the simplest example, a case where each of the master disk array controller 5 and the external bus 6 is two will be described.

In normal use, the host computer 1 commands the master disk array controller 51 to perform input / output. The master disk array controller 51 uses the external bus 61 to issue output commands to the logical disk devices 21 to 2n.

Next, the operation when the external bus 61 is defective will be described. When the master disk array controller 51 detects an abnormality in the bus sequence, it stops using the external bus 61 and uses the external bus 62 to issue I / O commands to the respective logical disk units 21 to 2n. . Each logical disk unit 21 to 2n is an external bus 6
Bus control for each of 1 and external bus 62
It has a function to perform processing regardless of which bus the input / output instruction arrives, and has a function to transfer the report and data using the bus where the input / output instruction arrives.

At this time, the input / output instruction and data that have used the external bus 62 are processed, and after execution is completed, the data is transferred for reporting and if necessary. The master disk array controller 51 performs necessary processing by using the data and the report sent from the external bus 62.

Further, the master disk array controller 5
When 1 fails and the host computer 1 starts using the master disk array controller 52, the master disk array controller 52 first uses the external bus 61 to input / output data to / from each logical disk device 21 to 2n. Execute an instruction. At this time, the process when the external bus 6 is always felt is the same as that of the master disk array controller 51. The external bus 62 is used to perform processing.

With the above processing, it becomes possible to access the data of the respective logical disk units 21 to 2n when the external bus 61 fails or when the master disk array controller 51 fails.

Example 5. The present invention is not limited to the above-described embodiments. For example, as shown in FIG. 8, the logical disk devices 21 to 2n and the logical disk devices 141 and 142 for restoration are connected to separate external buses ( 61-1, 61-
2, 61-n) to (62-1, 62-2, 62-n) can be variously modified without departing from the scope of the invention, such as a system connected to the master disk array controllers 51 to 5n.

[0061]

As described above, according to the first aspect of the present invention, since the disk array is configured in two hierarchies, one to all physical disk devices forming a redundant group fail. Also, the data that should be in this physical disk device can be modified / restored, and the reliability is improved.

According to the second aspect of the present invention, a logical disk device dedicated to restoration is provided, and when a certain logical disk device fails, the data of the failed logical disk device is restored. Since it is configured to store in a logical disk unit dedicated to restoration, this logical disk unit dedicated to data restoration can be used to perform data restoration at high speed and to create a logical disk that can be restored by the master disk array controller. It is possible to increase the number of failed disk devices and improve the reliability of the disk array system.

According to the third aspect of the invention, since two or more master disk array controllers for controlling the disk array system are connected to the external bus and the internal bus, even if the master disk array controller fails. , The operation can be continued, and the reliability of the disk array system is improved.

According to the fourth aspect of the invention, since the master disk array controller that constitutes the disk array system and each logical disk device are connected by a plurality of external buses, a failure occurs in the external buses. However, the operation can be continued and the reliability of the disk array system is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a disk array system according to a first embodiment of the present invention.

FIG. 2 is a schematic diagram showing an arrangement example of data in a logical disk device according to the first embodiment.

FIG. 3 is a schematic diagram showing an arrangement example of data in a physical disk device according to the first embodiment.

FIG. 4 is a block diagram showing a disk array device including a logical disk device dedicated to restoration according to a second embodiment of the present invention.

FIG. 5 is a schematic diagram showing an example of data arrangement in a logical disk device including a logical disk device dedicated to restoration according to a second embodiment.

FIG. 6 is a block diagram showing a configuration of a disk array system including two or more master disk array controllers showing a third embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of a disk array system including two or more master disk array controllers and two or more external buses according to a fourth embodiment of the present invention.

FIG. 8 is a block diagram showing a configuration of a disk array system having a plurality of separate external buses showing a fifth embodiment of the present invention.

FIG. 9 is a configuration diagram of a conventional disk array system.

[Explanation of symbols]

1 host computer, 21 to 2n logical disk device, 41 to 4n redundancy group formed by physical disk device, 41-1 to 41-n physical disk device, 5 (51 to 5n) master disk array Controller, 6 (61 to 6n) external bus, 7 redundant group formed by logical disk devices, 8 internal bus, 1
4 Logical disk unit dedicated to restoration, 16 Redundancy group for restoration.

Claims (4)

[Claims] 1. A plurality of physical disk devices that are arranged in an array to form a first-tier redundancy group, a function of controlling the plurality of physical disk devices in parallel, and redundancy. Due to the redundancy, when a limited number of physical disk devices in the redundancy group of the first layer fail, a slap having a function of correcting / restoring data in the failed physical disk device A logical disk device including a disk array control, a plurality of the logical disk devices that are arranged in an array to form a second level redundancy group, and the plurality of logical disk devices are controlled in parallel. Function and redundancy, and when a limited number of logical disk devices in the redundancy group fail due to the redundancy, the data in the failed logical disk device is A disk array system which is characterized in that a master disk array controller having a function of correcting / restore data. 2. A logical disk device dedicated to restoration, which stores data of a failed logical disk device when the logical disk device fails, according to claim 1. Disk array system. 3. The logical disk device and the master disk array controller are connected by an external bus, the master disk array controller and a host computer are connected by an internal bus, and the master disk is connected to the external bus and the internal bus. The disk array system according to claim 1, wherein two or more array controllers are connected. 4. An external bus for connecting the logical disk device and the master disk array controller is provided in two or more, and two or more master disk array controllers are connected to each of the external buses and the two are connected. 4. The disk array system according to claim 3, wherein the master disk array controller is connected to the same internal bus.