JPH05324134A - Duplexed computer system - Google Patents

Abstract

PURPOSE:To reduce the number of necessary power supplies without losing the reliability of a duplexed system and to reduce the size and price of the system. CONSTITUTION:Power is supplied in common from a power supply device 32 for driving one CPU 22 in the duplexed system to one shared memory 26 and a shared disk 30 in one system and power is supplied in common from a power supply device 33 for driving the other CPU 23 to a shared memory 27 and a shared disk 31 in the other system. When fault generation is detected in one CPU 22 or one shared module 26 or 30 at the time of operation based upon one CPU 22, the operation is switched to the other CPU 23, operational processing is continued by using the other shared modules 27, 31 and one power supply device 32 is disconnected to replace the module generating the fault.

Description

Detailed Description of the Invention

[0001]

The present invention relates to two CPUs (central processing units) such as a hot standby system and a duplex system, and a plurality of shared modules such as a shared memory and a shared disk between the two CPUs. A redundant computer system provided.

[0002]

2. Description of the Related Art Conventionally, when a dual computer system is constructed using two computers, data is shared between the two CPUs, so a shared memory, a shared disk, a shared I / O (input / output device) is used. ) Etc. are generally used.

In such a redundant computer system, 2
Since the shared part by one computer becomes a bottleneck in the reliability of the entire system, it is necessary to duplicate the above shared memory and shared disk.

FIG. 3 shows the configuration of a conventional redundant computer system. In FIG. 3, reference numeral 11 denotes an operating computer module for performing a given business process in a normal time, and 1
Reference numeral 2 is a standby computer module that takes over and executes the business processing when a failure occurs in the operating computer module 11.

The duplicated computer module 11,
12 are provided with the same CPUs 13 and 14, respectively, and each of the interface carts 15a-15
The shared memory 17 and 18 and the shared disks 19 and 20 that are duplicated are connected via d and 16a to 16d.

Here, each computer module 11, 1
2 is driven by respective dedicated power supplies 11a and 12a, and the shared memory 17 and 18 and the shared disk 19 and 20 that are duplicated operate as a whole of the duplicated system when a failure occurs in one of the modules. Without stopping, that is, it is necessary to continue the business processing utilizing the other system module to repair / replace the defective module. Therefore, independent power supplies 17a, 18a, 19 are provided.
a, 20a.

That is, in the normal business processing state by the operating-side computer module 11, the same data writing / reading processing is executed for both the duplicated shared memories 17 and 18 and the shared disks 19 and 20. For example, even if a failure occurs in one shared memory 17, the other shared memory 18 can be used to continue the business processing, and even if a failure occurs in the other shared disk 20, one shared disk 19 remains It can be used to continue business processing.

[0008]

As described above, in the conventional redundant computer system, the shared modules such as the shared memories 17 and 18 and the shared disks 19 and 20 are also duplicated, and each shared module is independent. Since it is driven by the power supplies 17a, 18a, 19a, and 20a of each of the shared modules, it is excellent in the reliability of business continuation when a failure occurs in the shared module. There is a problem that it becomes large and expensive.

The present invention has been made in view of the above problems,
An object of the present invention is to provide a duplicated computer system that can be reduced in size and cost by reducing the number of required power sources without deteriorating the reliability of the duplicated system.

[0010]

That is, a redundant computer system according to the present invention has two CPUs and a plurality of types of redundant shared modules.
Two power supply devices that supply power independently to each other, and one shared module that is connected between the two CPUs and that is powered by one power supply device that supplies power to one CPU , The other shared module paired with the one shared module, which is connected between the two CPUs and is powered by the other power supply device that supplies power to the other CPU, and the two CPUs And failure detecting means for detecting the occurrence of a failure in each of the redundant shared modules and one CPU to which power is supplied from the one power supply device, when the one CPU or one of the CPUs is operated by the failure detecting means. When a failure of the shared module is detected, the operation is switched to the other CPU to which power is supplied from the other power supply unit. It is constructed by a stage.

[0011]

In other words, even if only two power supply devices are provided, when a failure occurs in one of the CPUs during the execution of business processing including the redundant shared module in which one CPU is on the operating side and the other CPU is on the standby side. By switching its operating system, the other CPU continues the business processing utilizing the other shared module, and even if one shared module fails, the other operating system can be switched to another The CPU can continue the business processing utilizing the other shared module, and the reliability of the duplex system can be ensured.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the structure of a redundant computer system. The redundant computer system 21 includes, for example, a CPU (0) 22 on the operating side and a CPU (1) on the standby side.
It is provided with two CPUs 22 and 23 that are duplicated with 23.

Between the two CPUs 22 and 23, one shared memory 26 and the other shared memory 27 are connected in parallel via the interfaces 24 and 25, respectively.

The redundant shared memories 26 and 27 are used for data sharing of the main memory and communication between the redundant CPUs 22 and 23. The pair of shared memories 26 and 2 are used.
7 perform backup operations with each other. For example, if one shared memory 26 fails, the other shared memory 2
The processing in both CPUs 22 and 23 is continued by using only 7.

Further, between the two CPUs 22 and 23, one shared disk 30 and the other shared disk 31 are connected in parallel via the interfaces 28 and 29, respectively.

The dual shared disks 30 and 31 share file data between the two CPUs 22 and 23, and the pair of shared disks 30 and 31 mutually operate in the same manner as the shared memories 26 and 27. When the backup operation is performed and, for example, a failure occurs in one shared disk 30, the processing in both CPUs 22 and 23 is continued using only the other shared disk 31.

Here, the operating system computer module 21A
Is driven by one power supply device 32, and power is supplied from the one power supply device 32 to the one CPU 22, the shared memory 26, and the shared disk 30.

Further, the standby computer module 21B is
It is driven by the other power supply device 33, and power is supplied from the other power supply device 33 to the other CPU 23, shared memory 27, and shared disk 31.

On the other hand, the CPUs 22 and 23 each have a self-diagnosis function for detecting the occurrence of a failure due to the control abnormality, and the shared memory 2 existing in the own module
6, 27 and the shared disks 30, 31 have a failure occurrence detection function associated with abnormal data, and when a failure occurrence of the CPU 22 is detected in the active computer module 21A, or the shared memory 26 or the shared disk 30.
When the failure occurrence is detected, the failure occurrence information in the operating system is notified from the CPU 22 to the standby system CPU 23 through the shared memories 26 and 27.

In this case, the standby system CPU 23 carries out the business process takeover process from the operating system computer module 21A. At this time, the operating system is the standby system computer module 2
It is switched to 1B, and the other CPU 23 continues the business processing by using the shared memory 27 and the shared disk 31 in the own system module.

That is, when the redundant computer system having the above configuration is applied to a hot standby system,
If any failure occurs in the operating system computer module 21A, the business process is taken over by the standby system computer module 21B.

That is, in the state where the business process is being executed in the active computer module 21A, the active CPU 2
The writing / reading of data by 2 is performed in duplicate on the duplicated shared memories 26 and 27 or the shared disks 30 and 31, and the necessary data for the business process takeover at the time of occurrence of a failure is also described above. It is written in the shared memory 26, 27 and the shared disk 30, 31.

In this way, in the state where the business processing is being executed in the operating system computer module 21A using the shared memories 26 and 27 and the shared disks 30 and 31 with the one CPU 22 as the center, for example, the one CPU 22
When a failure occurs in the above, the failure occurrence information is notified from the failure occurrence CPU 22 to the standby CPU 23.

Then, the standby CPU 23 carries out the business process takeover process based on the process data currently stored in the shared memory 27 and the process data stored in the shared disk 31. The operation system of the process is switched to the standby system computer module 21B using the other shared memory 27 centered on the other CPU 23 and the other shared disk 31, and the given business process is continued. In this case, the power supply from one of the power supply devices 32 to the operating system computer module 21A is cut off to replace the faulty CPU 22.

Further, in the state where the business processing is executed in the operating system computer module 21A using the shared memories 26 and 27 and the shared disks 30 and 31 with the one CPU 22 as the center, for example, the one shared memory 2 is used.
When a failure occurs in No. 6, the occurrence of the failure is the operating system C
Detected by the PU 22, the standby CPU 23 is notified of the failure occurrence information.

FIG. 2 shows a switching state of the active system when the active system fails in the redundant computer system. The standby system C receives the failure occurrence information regarding one of the shared memories 26.
The PU 23 performs the business process takeover process based on the process data currently stored in the shared memory 27 and the process data currently stored in the shared disk 31 in the same manner as when the CPU 22 fails. Thus, the operational system for business processing is switched to the standby computer module 21B using the other shared memory 27 centered on the other CPU 23 and the other shared disk 31.
The given business process will be continued.

In this case, the power supply from one power supply unit 32 to the operating system computer module 21A is cut off, and the fault occurrence shared memory 26 can be replaced.

If a failure occurs in the standby CPU 23 or the other shared memory 27 or shared disk 31 belonging to the standby power supply during the execution of business processing in the active computer module 21A, the above operating computer module is used. The business processing in 21A is continued as it is, the power supply from the standby power supply device 33 is cut off,
The faulty module can be replaced.

Therefore, according to the redundant computer system having the above-described configuration, the power of the dual shared memory 26 and the shared disk 30 is commonly supplied from the power supply device 32 for driving the CPU 22 and the other of them is supplied. Power for the shared memory 27 and the shared disk 31 is commonly supplied from the power supply device 33 that drives the other CPU 23, and when one of the CPUs 22 operates and operates,
When a failure occurrence in the one CPU 22 or the one shared module (memory 26 or disk 30) is detected, the operation is switched to the other CPU 23.
The other shared module (memory 27 and disk 31)
Is used to continue the business process, and the power supply device 32 on one side can be shut off to replace the faulty module. Therefore, all the individual shared modules have independent power supply devices. Even if the reliability of the redundant system is not impaired, the two power supply devices 32,
Only 33 is required, and this computer system can be realized in a small size and at low cost.

[0031]

As described above, according to the present invention, two C
It has a PU and multiple types of redundant shared modules,
Two power supply devices that independently supply power to each of the two CPUs, and one power supply device that is connected between the two CPUs and that supplies power to one CPU, are supplied with power. One shared module and the above two CPs
The other shared module paired with the one shared module, which is connected between the U and is supplied with power from the other power supply device that supplies power to the other CPU;
Failure detection means for detecting the occurrence of a failure in each of the two CPUs and the redundant shared module, and one CPU supplied with power from the one power supply device during operation of the one CPU by the failure detection means. Alternatively, when a failure occurrence in one of the shared modules is detected, the other C is supplied with power from the other power supply device.
Since the PU is provided with the operation system switching means for switching the operation, the number of power sources required can be reduced, the size can be reduced, and the cost can be reduced without impairing the reliability of the duplex system.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a redundant computer system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a switching state of an active system when an active system fails in the redundant computer system.

FIG. 3 is a diagram showing a configuration of a conventional redundant computer system.

[Explanation of symbols]

21 ... Redundant computer system, 21A ... Operating computer module, 21B ... Standby computer module, 22, 2
3 ... CPU, 24, 25, 28, 29 ... Interface, 26, 27 ... Shared memory, 30, 31 ... Shared disk, 32, 33 ... Power supply device.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI Technical display location G06F 15/16 470 J 9190-5L B 9190-5L

Claims (1)

[Claims] 1. A dual computer system having two CPUs and a plurality of types of redundant shared modules, and two power supply units for independently supplying power to each of the two CPUs, and the two CPUs. One shared module that is connected to each other and that is supplied with power by one power supply device that supplies power to one CPU, and the other shared module that is connected between the two CPUs and that supplies power to the other CPU. Powered by a power supply,
The other shared module that forms a pair with the one shared module, a failure detection unit that detects a failure occurrence in each of the two CPUs and the duplicated shared module, and one of the power supply units supplies power. CP
When the failure detection means detects the failure occurrence of the one CPU or the one shared module during operation by U, the operation is performed on the other CPU to which power is supplied from the other power supply device. A redundant computer system comprising: an operating system switching means for switching.