JPH09288651A - Loosely coupled computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform the fast backup operations by preparing a fault notification part at every computer to receive a message showing execution of a normal operation, to monitor reception of the message and to notify the monitoring computer of a fact that no message is received any more. SOLUTION: Each of tightly coupled computers 110-1 and 110-2 consists of the CPU 100-1 and 100-2, a memory 101, an I/O 102, a shared disk controller 103 and a fault notification part 200-1 which are connected to the system buses 104-1 and 104-2, respectively. Both computers 110-1 and 110-2 share data via a shared memory 300 and a shared disk 105 and notify the interruptions via an interruption line 106 to perform the message communication and to cooperate with each other. Then a mutual communication network is connected to the I/O 102 and the simultaneous network communication can be performed between both computers. As a result, the fast backup operations can be secured between both computers.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system in which computers are loosely coupled to each other.

[0002]

2. Description of the Related Art Conventionally, for the purpose of surely completing a requested process within a fixed time, an FTC (Fault) which performs a non-stop operation by completely duplexing each resource and automatically switching when a failure is detected. Trerant Co
There was a mputter). Since the FTC has duplicated hardware, it is impossible to avoid a software failure. In a hot standby configuration or when a failure occurs, the processing capacity is lower than before the failure, but as a technology that can avoid the entire system down, a system in which computers are loosely coupled and backups are mutually Proposed. In order to make a backup, it is necessary to detect that an error has occurred. Conventionally, as disclosed in Japanese Patent Application Laid-Open No. 7-152694, a message indicating that the message is operating normally is periodically transmitted, and when the message disappears, an abnormality is detected. There is.

[0003]

SUMMARY OF THE INVENTION In such a conventional technique,
In order to detect a failure, the monitoring side must always accept a message indicating a normal operation, so that there is a problem in that extra processing increases during mutual backup and system throughput decreases. Also,
If the frequency of sending a message indicating a normal operation is reduced in order to prevent excessive processing, the failure detection will be delayed, which will delay the start of backup processing.

An object of the present invention is to provide a loosely coupled computer system that can detect at high speed that a message indicating a normal operation cannot be output due to a failure and can perform a high speed backup operation. is there.

It is an object of the present invention to provide a loosely coupled computer system capable of detecting a high-speed fault and performing a high-speed backup operation by providing a fault notifying unit for detecting the fault and notifying it to the monitoring computer.

[0006]

In order to achieve the above object, according to the present invention, each of the plurality of computers receives a message indicating that the computer is operating normally, monitors the message, and confirms the normal operation. A failure notification unit is provided to notify the computer on the monitoring side that the message indicating that it is not arriving has arrived.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION The loosely coupled computer system according to the present invention will be described below in more detail with reference to some embodiments.

First Embodiment of the Invention (Outline of System Configuration) FIG. 1 shows one embodiment of the present invention. In the loosely coupled computer system 100, the tightly coupled computers (110-1 to 110-2) share the shared disk 10 with each other.
5. The shared memory 300 and the fault notification line 106 are connected to each other. Actually, the number of tightly coupled computers (110-1 to 110-2) to be connected may be two or more, but here, two computers, which are the minimum required for the description, are shown.

Each tightly coupled computer (110-1 to 110-
2) is a system bus (104-1 to 104-2), a CPU (100-1 to 100-2), a memory 101,
The I / O 102, the shared disk controller 103, and the failure notification unit 200 are connected and configured. The tightly-coupled computers (110-1 to 110-2) share data via the shared memory 300 and the shared disk 105, and perform interrupt notification via the interrupt notification line 106, thereby performing message communication. And work in concert. Also,
A network for mutual communication may be connected to the I / O 102 to perform network communication at the same time.

The I / O 102 is a device for inputting / outputting, and is composed of, for example, a display device, a keyboard, a network, and a secondary storage such as a hard disk. In this embodiment, only one I / O is connected, but two or more I / Os may be connected. The shared memory 300 includes the tightly coupled computers (110-1 to 110).
-2) is a memory that can be accessed.

The configuration of the shared memory 300 will be described with reference to FIG. The shared memory 300 includes a bus connection mechanism (310-1 to 310-2) that receives a read and a write for the associated physical address range and a RAM 320. The bus connection mechanism is the system bus (104-1
To 104-2) are monitored, and if the address is in the corresponding range, it is a mechanism for accepting read and write, and can be constituted by logic such as TTL or CMOS.

The shared disk controller 103 is a controller for controlling a disk for sharing data, and is a SCSI (Small C) described in pages 87 to 175 of the January 1993 issue of the interface.
As the disk control hardware based on, for example, the Computers System Interface), T
It can be configured by logic such as TL or CMOS.

The shared disk 105 is composed of a hard disk.

(Fault Notification Unit) The fault notification unit 200 indicates the interrupt designation register 210 that holds the interrupt notification destination and method when the fault notification is received, the state in which the interrupt is issued when the interrupt is issued. An interrupt issuance status register 220 for notifying, a counter 230 for detecting a failure by receiving a message indicating that a normal operation is performed and decreasing at regular intervals, and a control unit 240 for controlling the register and the counter A failure notification line 106 for notifying a failure and a notification unit 250 for notifying a failure are connected by an internal bus 260. The interrupt designation register 210, the interrupt issue status register 220, and the counter 230 can be configured by logic such as TTL or CMOS or RAM. Control unit 240, notification unit 25
0 can be configured by logic such as TTL or CMOS.

The notification of the failure is performed as follows. C
The PUs (100-1 to 100-2) specify the interrupt destination and method to be issued when a failure is detected.
The interrupt designation register 210 is set. CPU (100-1 to 100-2) while the process is proceeding normally
Writes a value to the counter 230 at regular intervals. The control unit 240 decrements the value of the counter 230 at regular time intervals, and when the value becomes 0, notifies the alternative computer via the notifying unit 250 of a failure. CPU (100-1 to 100-2) while the software is operating normally
Since the value is written in the counter 230 at regular intervals, the value never becomes zero. However, at the time of abnormal operation, the counter cannot be updated for a certain period of time or more, which causes a failure notification to the alternative computer.

(Alternative Processing) Hereinafter, alternative preliminary processing will be described as initialization processing for performing the alternative operation according to FIG.
First, as shown in step 310, the I / O device 102
The substitution target information set in advance is read out. next,
As shown in step 320, the interrupt designation register 210 is set in order to designate the interrupt at the time of failure detection. Finally, as shown in step 330, a spare program for performing a substitute operation is activated and the substitute process is set to be performed by an interrupt issued when a failure is detected.

When an interrupt occurs, it can be detected that it is a failure notification interrupt by checking the interrupt issue status register 220, and an alternative operation is started.

<Second Embodiment of the Invention> Instead of providing the fault notifying section 200 connected by the fault notifying line 106 as in the first embodiment, as shown in FIG. 4, each tightly coupled computer ( It is also possible to provide the monitoring unit 200 connected to 110-1 to 110-2) and to notify the failure by this monitoring unit 200. The monitoring unit 200 includes a memory 210 and a CPU 220 that can be accessed by each tightly coupled computer (110-1 to 110-2).
During normal operation, the operation check flag is updated, the CPU 220 detects that the update is no longer performed, and this is regarded as a failure occurrence, and the failure assignment is determined based on the information held in the interrupt notification information. Include notification. With the above method, the same operation as that of the first embodiment can be performed and the same effect can be obtained.

[0019]

As described above, according to the present invention, since it is not necessary for the alternative computer to monitor the message indicating that the normal operation is being performed, the soft overhead of the alternative computer can be reduced. Further, instead of monitoring that the message indicating that normal operation is not performed for a certain period of time, it is possible to directly notify that a failure has occurred, so that failure recovery processing can be performed at high speed.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a loosely coupled computer system according to the present invention.

FIG. 2 is a diagram showing a configuration example of a shared memory in the device of FIG.

3 is a flowchart of an alternative preparatory process used in the apparatus of FIG.

FIG. 4 is a diagram showing a second configuration of the loosely coupled computer system according to the present invention.

[Explanation of symbols]

104-1, 104-2 ... System bus, 260 ... Internal bus.

(72) Inventor Masahiko Yamauchi 12 shares at 890 Kashimada, Sachi-ku, Kawasaki City, Kanagawa Prefecture, Ltd. Information & Telecommunication Development Division, Hitachi, Ltd. (72) Inventor, Takehisa Hayashi, 890 12 Stock Company Hitachi, Ltd. Information & Telecommunication Development Headquarters (72) Inventor Yasuhiro Ishii 810 Shimoimaizumi, Ebina, Kanagawa Prefecture Office Systems Division, Hitachi, Ltd.

Claims (4)

[Claims] 1. A plurality of computers and a connection mechanism for transferring data, the plurality of computers for normal operation confirming means for notifying continuation of operation and for notifying an alternative computer of failure detection. And a notification unit for notifying the normal operation and notifying the normal operation confirmation unit that the notification has not been made for a predetermined time or longer. 2. A plurality of computers, a connection mechanism for transferring data, or a shared memory and a shared disk for sharing data, the plurality of computers being normal for notifying continuation of operation. A loosely coupled computer characterized by having an operation confirmation means and a notification means for notifying the alternative computer of failure detection, and notifying that the normal operation confirmation means has not been performed for a certain period of time or more. system. 3. A second shared memory having a plurality of computers and a connection mechanism for transferring data, accessible from the computer, and a processor different from the computer for monitoring the second shared memory. An alternative configured to monitor the normal operation confirmation area for notifying the continuation of the operation by each computer held in the second shared memory, and to hold that the normal operation is not continued in the notification information area A loosely coupled computer system characterized in that notification is made by information for notifying a computer. 4. A second shared memory and a second shared memory which have a plurality of computers, a connection mechanism for transferring data, or a shared memory and a shared disk for sharing data, and which can be accessed from the computers. It is composed of a processor different from the above computer that monitors the shared memory, and this processor monitors the normal operation confirmation area for notifying the continuation of the operation by each computer held in the second shared memory and continues the normal operation. A loosely coupled computer system characterized by notifying by means of information for notifying an alternative computer held in the notifying information area that this has not been done.