JPH01237859A - System switching system - Google Patents

Abstract

PURPOSE:To perform the switching of a system without interrupting an operation by switching a CPU connected to an input/output channel by a switching device. CONSTITUTION:Plural CPUs (101, 102,...10n) are connected to the input/output channel 500 via the switching device 400, and an input/output device 600 is connected to the input/output channel 500. The CPUs are provided with communication buses (P12,...P1n,...P2n), etc., for the communication of mutual states of them. A monitoring means 300 issues interruption to the input/output channel 500 at the time of finding the generation of abnormality in the CPU connected to the input/output channel 500, and also, supplies switching indication to connect another CPU to the input/output channel 500 to the switching device 400. The input/output channel 500 is set at a stopping state by the interruption, and restarts the operation at the time of completing the switching of the CPU.

Description

[Detailed Description of the Invention] [Summary] Regarding the improvement of a system switching method in a multiplexed computer system, the present invention relates to an improvement in a system switching method in a multiplexed computer system so that central processing units in the multiplexed computer system can be switched in a short time and without stopping operation. a plurality of central processing units, an input/output channel, an input/output device, a switching device for connecting one of the plurality of central processing units to a human output channel, and a plurality of central processing units; and a communication path for communicating mutual status between the plurality of central processing units, and the monitoring unit is configured to detect an abnormality in the central processing unit connected to the human output channel. The input/output channel is configured to give a level 2 reset signal to the input/output channel when it discovers that a problem has occurred, and to give the switching device a switching instruction to connect another central processing unit to the human output channel. The configuration requirement is that the channel is configured to be in a stopped state when a level 2 reset signal is applied.

[Industrial application field]

The present invention relates to an improved system switching scheme in a multiplexed computer system.

In industrial and social systems where computers are the core, a computer outage has a huge impact. For example, as shown in FIG. 4, in a seat reservation system, a large number of input/output devices (
In this case, the terminal devices 601, 602, etc. access the input/output channel 500 (in this case, the communication line channel) and the application program. In such a system, a plurality of central processing units 100 are used for high reliability, and by configuring a multiplexed computer system, even if a failure occurs in one central processing unit, the other central processing units A method is adopted in which the processing device takes over the processing.

[Conventional technology]

FIG. 5 is a diagram showing an example of a conventional duplex system. In the figure, 101 and 102 are central processing units, 201 and 2
02 is a memory, 300 is a monitoring device, 400 is a switching device,
500 is an input/output channel, 600 is an input/output device, and S4 to S32 are signal lines.

In the illustrated state, the central processing unit 101 is connected to the switching device 400.
and input/output device 600 via human output channel 500
The central processing unit 102 is connected to and is performing online processing, and the central processing unit 102 is performing other low-priority processing. The monitoring device 300 diagnoses the central processing units 101 and 102 and monitors the central processing unit 101 which is performing online processing.
When the occurrence of a fault is detected, the switching device 40 issues a switching instruction to connect the input/output channel 500 to the interface signal line S32 on the central processing unit 102 side via the signal line S4.
0, and also sends a reset signal to the input/output channel 500 via the signal line S5. When the input/output channel 500 receives this reset signal, it performs clearing processing,
Return to initial state.

[Problem to be solved by the invention]

When a hot standby system is configured using two central processing units 101 and 102 as shown in FIG.
A method is used in which a portion corresponding to the switching device 400 is provided with a plurality of access buses (multiported) to the input/output channel 500 as shown in FIG. 6, and is accessed from both central processing units.

However, in this type of conventional system, (1) A failure of the input/output channel itself may cause both central processing units to go down.

■ Difficult to perform hot maintenance replacement (for example, signal line S32 cannot be removed while the power is on).

■ If there are three or more central processing units, the interface circuit for each channel will become larger.

This creates a drawback. Therefore, as shown in FIG. 5, a switching device 400 is used to connect the signal line S31. In this case, the input/output channel 500 connected to the central processing unit 101 is switched to the central processing unit Z10.
2, in order to initialize the input/output channel 500, the monitoring device 300 gives a reset signal to the input/output channel 500 via the signal line S5, and the input/output channel 500
The clearing process was performed by

This clearing process is performed by the central processing unit 102 that takes over.
Since the software processing corresponds to the Rishut processing,
This is done using the same reset signal as the normal clearing process when the power is turned on. However, in this conventional method, ■ it takes time to initialize and initialize input/output channels;

■ The central processing unit software must reconnect the reset channels and logical paths, causing a temporary halt in operations.

It has the disadvantage of.

The present invention was created in view of this point, and is a system switching system that enables switching of central processing units in a multiplexed computer system in a short time and without temporarily stopping operation. The purpose is to provide a method.

[Means to solve the problem]

FIG. 1 is a diagram showing the principle of the present invention. The computer system of the present invention includes a plurality of central processing units 101, 102, .
..., 10n, an input/output channel 500, and an input/output device 600 connected to the lower side of the input/output channel 500.
and a plurality of central processing units 101, 102 . ..., 10
a switching device 400 that connects one of n to an input/output channel (500), and a plurality of central processing units 101, 102.
．． - A communication path P12., for communicating the mutual status between the monitoring means 300 for diagnosing 10n and the plurality of central processing units. ...Pin, ...P2n, ... are provided.

When the monitoring means 300 discovers that an abnormality has occurred in the central processing unit connected to the human output channel 500,
A level 2 reset signal is given to input/output channel 500, and other central processing units are connected to input/output channel 5.
A switching instruction to connect to 00 is given to the switching device 400.

When the switching device 400 is given a switching instruction to connect another central processing unit to the input/output channel 500,
Connect the input/output channels to the other central processing unit.

The input/output channel 500' enters a stopped state when a level 2 reset signal is applied.

〔Example〕

FIG. 2 is a block diagram of one embodiment of the present invention.

In the figure, 101 and 102 are central processing units, 201
and 202 are memories, 301 and 302 are monitoring devices, and 400
500 is a switching device, 500 is an input/output channel, 501 is a reset circuit, 502 is a microprocessor, and 503 is a communication interface. Also, Sl.

S21. ..., S7 indicates a signal line.

The central processing unit 101 executes a program stored in the memory 201, and the central processing unit 102 executes a program stored in the memory 202.
Execute the program stored in .

The central processing units 101 and 102 perform mutual diagnosis via the signal line S1 and also periodically notify the status.

The central processing unit 101 connects to the switching device 4 via the signal line S31.
00, and the central processing unit 102 outputs the signal vAS32.
It is connected to the switching device 400 via. Switching device 400
connects either one of central processing units 101 and 102 to input/output channel 500. The monitoring device 301 diagnoses the central processing unit 102, and the monitoring device 302 diagnoses the central processing unit Lot. In an actual system, the monitoring devices 301 and 302 are equipped with central processing units 101 and 10 so that they do not have to have independent housings and power supplies to reduce costs.
Although it is designed to be implemented in a second case, it logically operates independently of the central processing unit. Monitoring device 301.
Diagnosis of the central processing units 102 and 101 by 302 is performed as follows.

(2) The monitoring device 302 has a timer and sends an interrupt signal to the central processing unit 101 via 521 at regular time intervals.

(2) The central processing unit 101 recognizes an interrupt from the monitoring device 302 based on the interrupt code, and starts a self-diagnosis program.

(2) If there is a failure in the central processing unit 101, the self-diagnosis program will terminate abnormally or will not terminate even after a certain period of time.

(2) The monitoring device 302 monitors response data from the central processing unit 101. If there is no response even after a certain period of time has elapsed after the interrupt according to item (3), or if there is a response but it is found that it is impossible to continue the operation of the central processing unit 101 based on the contents of the response data, the following switching process is performed. Start.

Note that the relationship between the monitoring device 301 and the central processing unit 102 is also similar. The monitoring device 301 can instruct the switching device 400 to switch, and similarly the monitoring device 302 can instruct the switching device 400 to switch. The monitoring device 301 can give a level 1 reset signal to the input/output channel 500 via the signal line S51,
A level 2 reset signal can be applied to the input/output channel 500 via the signal line S52. Monitoring device 302
The same is true. The input/output channel 500 operates as a communication line channel. 500 input/output channels
has a reset circuit 501, a microprocessor 502, and a communication interface 503. Signal line S5
When a level 1 reset signal is sent via 1,
Reset circuit 501 returns channel 500 to its initial state.

A level 1 reset signal is applied to the input/output channel 500 when a failure occurs in the input/output channel itself, when the power is turned on, or when it is necessary to reset the entire system. When a level 2 reset signal is sent via the signal line S52, the microprocessor 502 is in a stopped state only while the level 2 reset signal is active (on).

Even when the central processing unit 101 is connected to the communication line channel 500 and communicating with an external device (terminal or other system) via the communication line S7, the application program of the central processing unit 102 remains in hot standby. state, and the central processing unit 1 is sent via the signal cotton S1.
You can know the latest status of 01 and the connection status with the communication line.

When the monitoring device 302 discovers that a failure has occurred in the central processing unit 101, it turns on the level 2 reset signal on the signal line S52, stops the input/output channel 500, and connects the input/output channel 500 to the central processing unit 102. A switching instruction indicating that the connection should be made is sent to the switching device 400 via the signal line S4.
and input/output channel 5 to central processing unit 102.
Notify that it is connected to 00. 500 input/output channels
When the switching device 400 receives a switching instruction to connect the switch to the central processing unit 102, the switching device 400 connects the switch to the central processing unit 102.
02 side and connect the human output channel 500 to the central processing unit 1.
Connect to 02.

Upon receiving the notification that the input/output channel 500 and the central processing unit 102 are connected, the central processing unit 102 instructs the monitoring device 302 to stop the level 2 reset signal, and disconnects the signal line. A command is given to the input/output channel 500 based on the state of the central processing unit 101 received via S1 and the latest state data of the application program running on the central processing unit 101, and communication between the central processing unit 101 and the external device is performed. Resume data communication. central processing unit 101,
The processing flow of step 102 is shown in FIG. Note that FIG. 3 shows a case where the central processing unit 101 is performing business processing (data communication with a terminal through a communication line). This reduces the switching time and reduces the need for external devices (terminals and other systems).
It becomes possible to switch central processing units while maintaining the logical connection (session) with the central processing unit.

[Effects of the Invention] As is clear from the above description, according to the present invention, (a) Even when it becomes necessary to switch the central processing unit, the logical path between the input/output device and the application program ( session) remains connected, so there is no operational pause.

(b) Since initial setting processing of input/output channels is not required at the time of switching, switching time can be shortened.

It is possible to achieve remarkable effects such as:

[Brief explanation of the drawing]

FIG. 1 is a diagram of the principle of the present invention, FIG. 2 is a block diagram of an embodiment of the present invention, FIG. 3 is a diagram showing the operation of the central processing unit, and FIG. 4 is a block diagram of an example of a seat reservation system. FIG. 5 is a block diagram of an example of a conventional duplex system, and FIG. 6 is a diagram showing an example of multi-board input/output channels. lOl and 102... central processing unit, 201 and 202...
...Memory, 301 and 302...Monitoring device, 400.
...Switching device, 500...I/O channel, 501.
...Reset circuit, 502...Microprocessor,
503...Communication interface. Patent applicant: Pfn Co., Ltd.

Claims (1)

[Claims] A plurality of central processing units (101, 102, ..., 10n
), an input/output channel (500), an input/output device (600) connected to the lower side of the input/output channel (500), and a plurality of central processing units (101, 102, . . . , 10n)
) to the input/output channel (500), and a plurality of central processing units (101, 102, . . . , 10n).
) and communication paths (P12, . . . P1n, . . . P2n, . . . ) for communicating mutual status between the plurality of central processing units.
), wherein the monitoring means (300) switches the input/output channel (500) when it discovers that an abnormality has occurred in the central processing unit connected to the input/output channel (500). configured to give a level 2 reset signal to the output channel (500) and give a switching instruction to the switching device (400) to connect another central processing unit to the input/output channel (500); When receiving a switching instruction to connect another central processing unit to the input/output channel (500), the switching device (400) connects the input/output channel (500) to the other central processing unit. A system switching system characterized in that the input/output channel (500) is configured to be in a stopped state when a level 2 reset signal is applied.