JP2001184138A - Hardware system and its fault solving method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To more quickly and surely solve a problem caused by a fault of a hardware module in a computer system. SOLUTION: The system is provided with plural switches for individually switching the on/off of power supply to individual hardware modules such as an I/O device and an I/O controller. After temporarily stopping power supply to a hardware module whose fault is detected by controlling these switches, power is supplied again to the module to independently reset the module.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a hardware system such as a computer system and a method for resolving a fault in a hardware module connected to the system.

[0002]

2. Description of the Related Art Most transient faults occurring in a hardware system such as a computer system are eliminated by reinitializing (resetting) the entire system. However, in general, resetting of the entire system is performed in a number of steps, and thus takes a long time. Therefore, there is a method such as a checkpoint / recovery method for solving a failure in a short time.

In this checkpoint / recovery method, when a failure occurs in a certain hardware module (equipment), an attempt is made to recover from the failure state by supplying a reset signal while the power is supplied to the module. However, this method is based on the premise that the failure of the module does not reach the reset circuit. If the failure of the module spreads to the reset circuit, the reset does not work properly and the failure cannot be recovered.

[0004]

As a countermeasure against the occurrence of a fault in a hardware system such as a computer system, a method of resetting the entire system and a method of resetting a faulty module while the power is on are provided. There is. However, although the method of resetting the entire system is reliable, it takes a long time until the reset is completed, and the method of resetting the failed module while the power is on is not reliable.

The present invention has been made to solve such a problem, and a hardware system and a method for solving the hardware system which can more quickly and surely solve a system problem caused by a hardware module failure. The purpose is to provide.

[0006]

In order to achieve the above-mentioned object, a hardware system according to the present invention comprises a central controller, a plurality of hardware modules operated by the central controller, and the individual hardware. It is configured to include switch means for individually switching on / off of power supply to a module, and switch control means for controlling the switch means to reset the hardware module in which a failure has been detected.

According to the present invention, by controlling the switch means to reset only the hardware module in which a failure is detected, temporarily stopping the supply of power to the hardware module, and then supplying the power again, The system can be recovered from the faulty state faster than the known system for resetting the entire system, and the failure can be more reliably eliminated than the known system for resetting the faulty module while the power is on.

Further, by adding a configuration for controlling the switch means to turn off the power supply for an optimum time for each hardware module, it is possible to reset the hardware module with higher reliability.

Further, there are provided identification information reporting means for reporting identification information of the hardware module, and switch connection management means for managing a connection relationship between each switch and each hardware module based on the identification information reported from each hardware module. By providing such a system, even in a system in which hardware modules can be exchanged, the connection relationship between each switch and each hardware module becomes clear, and reliability can be maintained.

[0010]

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

FIG. 1 shows a configuration of a hardware system according to an embodiment of the present invention.

The hardware system includes, for example, a processor 1 which is a central control device for controlling the entire system, and a plurality of hardware modules 2 (hereinafter simply referred to as a module 2) operated under the control of the processor 1.
Call. ), A power supply 3 for supplying necessary power to the system, a power switch 4 for individually turning on / off the power supply from the power supply 3 to each module 2,
A module setting information storage device 5 for storing setting information of each module 2, a module checking device 6 for specifying a module 2 in which a failure has occurred, and a module 2 based on the setting information stored in the module setting information storage device 5.
And a power switch control device 8 for controlling each power switch 4.

This hardware system can be applied to any system such as a computer system using the module 2 as various I / O devices and a process system using the module 2 as a processing device for various processes.

Each of the module setting information storage device 5, the module check device 6, and the module resetting device 7 may be realized by a program executed by the processor 1 or by dedicated hardware under the control of the processor 1. It may be realized.

Next, the operation of the hardware system will be described.

At the time of normal operation of the hardware system, a control signal is supplied from the processor 1 to the power switch control device 8 so that each power switch 4 is turned on. Power is supplied to each module 2 through the power supply.

When a failure occurs in any one of the modules 2, the processor 1 checks the status of each module 2 by the module check device 6. When the module check device 6 determines the module 2 in which the failure has occurred, the processor 1 subsequently switches the power switch 4 for switching power supply to the module 2 in which the failure has occurred to the off state by the power switch control device 8. Then, the power supply to the module 2 is stopped for a certain time. After a certain period of time, the power switch 4 is turned on by the power switch control device 8 and the module 2
Resumes power supply to the

When the module 2 is started, the processor 1 sets the module setting information stored in the module setting information storage device 5 in the module 2 by the module resetting device 7. As described above, the reset of the failed module 2 is completed.

As described above, the failed module 2
By temporarily stopping the supply of power to the power supply and then restarting the power supply again and resetting the module 2 alone, the system can be recovered from the faulty state faster than the known method of resetting the entire system. In addition, the failure can be eliminated more reliably than the known method of resetting the failure generating module 2 while the power is on.

Next, an embodiment will be described in which this hardware system is applied to a computer system such as a personal computer.

FIG. 2 shows the configuration of the computer system. In this computer system, the module setting information storage device 5, the module check device 6, and the module resetting device 7 of the hardware system of FIG.
This is realized by U11 and a program loaded in the main storage unit 12. Further, as the module 2, drives such as a hard disk drive and an optical disk drive,
At least an I / O device 13 such as a display, a communication device such as a modem or a network adapter, and an I / O controller 14 for controlling the device 13 are present.

The I / O device 13 and the I / O controller 14 are connected to the power supply 3 through separate power switches 4a and 4b, respectively, and the power supply is switched on / off individually by the power switch controller 8. Is The control of the power switch control device 8 is based on a control signal sent from the CPU 11 through the processor bus 15. In the figure, the I / O devices 13 and I
Although only one set of the / O controller 14 is shown, there are a plurality of these sets.

Consider a case where a failure occurs in a certain I / O device 13 of a group of modules connected to the system. The occurrence of a failure in the I / O device 13 is determined by the CPU 1
1 and the communication between the I / O device 13,
The determination can be made based on the absence of a response from the O device 13 or a response of an error occurrence signal from the I / O device 13. When it is determined that a failure has occurred in the I / O device 13, a control signal for stopping supply of power to the I / O device 13 is sent from the CPU 11 to the power switch control device 8. The power switch control device 8 receives the control signal and controls the switch 4b for switching the power supply to the I / O device 13, thereby stopping the power supply to the I / O device 13 for a certain period of time. Then, the switch 4b is turned on again to restart the power supply to the I / O device 13. After the power is turned on, the setting of the I / O device 13 is performed, and the reset of the I / O device 13 is completed.

A failure of the I / O device 13 may spread not only to the I / O device 13 but also to the I / O controller 14 that controls the I / O device 13. In this case, I
Resetting only the / O device 13 does not solve the system problem.

Therefore, if the failure is not resolved after resetting the I / O device 13, the power supply to the I / O device 13 is turned off again and the I / O device 13 is sent to the I / O controller 14 again. The power supply is also turned off, and after a certain period of time, the power supplies of both the I / O device 13 and the I / O controller 14 are turned on. That is, the set of the I / O device 13 and the I / O controller 14 is reset. As a result, even if the failure of the I / O device 13 has spread to the I / O controller 14, the failure of the system can be resolved. In addition, I /
If the failure is not solved even after resetting the pair of the O device 13 and the I / O controller 14, the higher control module may be reset together. Since the modules of the entire system are connected in a tree structure, until the problem is solved, the range of the module hierarchy to be reset is expanded from the end side to reset the minimum number of required modules, so that efficient The problem of the system can be solved quickly and reliably.

The optimum power-off time for resetting differs depending on the type of the module. For example, in the case of a network adapter, about 1 s, SCSI
In the case of equipment, it is about 5 s. Therefore, it is preferable to prepare in advance a table in which information of the optimum power-off time for each module is registered, and to select the power-off time with reference to this table. This table may be created manually by the user, or the information of the power-off time is recorded in the module 2 in some form, and the system reads the information from the module 2 when the table is created and automatically creates the table. You may make it.

Next, the correspondence between modules and switches will be described.

It is assumed that the computer system has a function of recognizing which module 2 is currently connected and which module 2 has a fault. When the failed module 2 is determined, C
Since the PU 11 notifies the power switch control device 8 of the failed module 2, the power switch control device 8 checks the power switch 4 to which the module 2 is connected and controls switching of the switch. However, in the case of a computer system in which the modules 2 can be replaced, the power switch control device 8 needs to know what modules 2 are currently connected to the individual power switches 4.

Therefore, in the present embodiment, as shown in FIG. 3, a correspondence table 21 between the power switch 4 and the module 2 is prepared in the power switch control device 8, and the CPU 1
When a faulty module 2 is notified from 1 to the power switch control device 8, the power switch 4 connected to the module 2 can be checked with reference to the correspondence table. This correspondence table 21
Can of course be created manually, but in the present embodiment, a function of automatically creating the correspondence table 21 is employed.

FIG. 3 shows a configuration for this automatic table creation. The module 2 is provided with an ID generation circuit 22. The ID generation circuit 22 puts a modulation signal based on a high-frequency pulse including ID information of the module 2 on the power supply line 23. As the power switch 4, a power transistor, FET, or the like that can transmit a modulation signal by a high-frequency pulse put on the power line 23 to the switch control port 24 of the power switch control device 8 is used.

The power switch control device 8 is provided with a high-pass filter 25 and a decoder 26 for each switch control port 24. The high-pass filter 25 extracts a high-frequency component from the signal of the switch control port 24 and introduces the high-frequency component into the decoder 26. The decoder 26 demodulates the signal to obtain ID information of the module 2 connected to the power switch 4. The power switch control device 8 creates the correspondence table 21 using the ID information of the module 2 obtained in this manner. Note that the module 2 requires a low-pass filter 27 for cutting high-frequency pulses from the ID generation circuit 22 onto the power supply line 23 and supplying the high-frequency pulses to the main circuit at the subsequent stage.

Here, the module 2 is connected to the power supply line 23.
Although the method of transmitting the ID information as a high-frequency pulse and transmitting it to the power switch control device 8 has been described, another method may be used as long as the ID information can be notified from the module 2 to the power switch control device 8. For example, the number of the power switch 4 may be recognized on the module 2 side, and the switch number and the ID information of the module 2 may be transmitted to the switch control device 8 via the CPU 11.

Further, in this embodiment, the correspondence table between the switch number and the module ID information is managed in the power switch control device 8, but the management of the correspondence table may be performed by the CPU 11. .

[0034]

As described above, according to the present invention,
The system can be recovered from the faulty state faster than the known system for resetting the entire system, and the failure can be more reliably eliminated than the known system for resetting the faulty module while the power is on. Further, by turning off the power supply for an optimal time for each hardware module, it is possible to reset the hardware module with higher reliability. Furthermore, by managing the connection relationship between each switch and each hardware module based on the identification information declared by each hardware module, even in a system where hardware modules can be exchanged, the connection between each switch and each hardware module can be performed. The connection relationship becomes clear, and reliability can be maintained.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing an embodiment in which the hardware system of FIG. 1 is applied to a computer system

FIG. 3 is a diagram showing a configuration for automatically creating a correspondence table between a power switch and a module in the computer system of FIG. 2;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Processor 2 ... Hardware module 3 ... Power supply device 4 ... Power supply switch 5 ... Module setting information storage device 6 ... Module check device 7 ... Module resetting device 8 ... Power supply switch control device 11 ... CPU 12 ... Main storage unit 13 ... I / O device 14 ... I / O controller 15 ... Processor bus

Claims (7)

[Claims] 1. A central controller, a plurality of hardware modules operated by the central controller, switch means for individually turning on / off power supply to each of the hardware modules, and detecting a fault. And a switch control means for controlling the switch means to reset the hardware module. 2. The control unit according to claim 1, wherein the control unit controls the switch unit to turn off power supply for a time determined for each hardware module.
Hardware system as described. 3. The hardware system according to claim 1, wherein said switch means comprises a plurality of switches for switching on / off a connection to a power supply for each of said hardware modules. 4. An identification information reporting unit for reporting identification information of each hardware module, and each switch and each hardware module based on identification information declared by the identification information reporting unit of each hardware module. Switch connection management means for managing a connection relationship between the switch module and the hardware module, based on a connection relationship between each of the switches and each of the hardware modules managed by the switch connection management means. 4. The hardware system according to claim 3, wherein the hardware system is controlled. 5. A method for resolving a fault in a hardware system having a central controller and a plurality of hardware modules operated by the central controller, comprising: detecting the failed hardware module; Temporarily turning off the supply of power to the detected hardware module and resetting the hardware module. 6. The power supply is turned off for a time determined for each hardware module.
A method for resolving a hardware system failure as described. 7. The hardware according to claim 5, wherein when the failure of the hardware module in which the failure is detected fails to be resolved by resetting the hardware module, the number of hardware modules to be reset is increased and the failure is resolved again. How to troubleshoot the system.