JPH02138649A - Detecting method for unpackaged state of memory exchange unit - Google Patents

Abstract

PURPOSE:To prevent the loss of a system function caused when a memory is set into a duplex system by detecting the abnormality of the memory exchange unit corresponding to a non-code in the case no coincidence is obtained between the write check data and the data read out. CONSTITUTION:A prescribed access address is produced from a system MM0 which actuates a system based on the access information and applied to each memory exchange unit of a memory 201 set into a system which is incorporated as a spare system MM1 after duplication. Then the check data is written into and read out of a memory area 200 of the corresponding memory exchange unit by means of a prescribed access address. The the detection information on the presence of the mispackage or the fault of the memory exchange unit is obtained based on the coincidence or discordance decided between the write check data and the data read out. Thus it is possible to prevent the loss of a system function caused when a memory is set into a duplex system.

Description

[Detailed Description of the Invention] [Summary] This invention relates to a method for detecting uninstallation of a memory exchange unit in a duplex system without an automatic uninstallation check function, with the aim of preventing system function loss due to setting memory in a duplex system. In a redundant system that is not equipped with a non-implementation check function, access information for each memory exchange unit is prepared, and the central processing unit of the system that is running the system (hereinafter referred to as the system operation system) makes the system standby. Prior to copying the contents of the system operating system memory by the system operating system central processing unit to the memory set wholly or partially as the system memory, an access address for each memory exchange unit determined according to the access information. , write check data to the access address, read the data written to the access address, and if there is no sign between the write check data and the read data, write the corresponding non-sign,
It was configured to detect abnormalities in units of memory replacement.

[Industrial application field]

The present invention relates to a method for detecting uninstalled memory exchange units for a duplex system not equipped with an automatic uninstalled check function.

In the digital exchange system, the system has a redundant configuration in order to improve its reliability. In such a system, if a failure occurs in all or part of the main storage of one system, the redundancy function will be lost.
In order to replace it with a good one, the system must be temporarily unified, and then all or part of the faulty main memory must be incorporated into the system in order to transition to the original duplex system.

In doing so, the system must be equipped with means to ensure correct installation of the replaced parts.

The reason for this is to prevent loss of system functionality. In other words, even if data is copied from the central control unit of the active system to the built-in main memory, the copy may not be performed correctly, and as a result, the standby system may not operate correctly as the active system at system switchover. This is because you won't be able to get anything.

[Conventional technology]

Therefore, in conventional redundant digital switching systems,
The system is equipped with an unimplemented address check function to prevent loss of system functionality due to main memory replacement.

[Problem to be solved by the invention]

In a redundant digital switching system, etc., it is necessary to prevent the loss of system functionality due to replacement of the main storage device, so it was not necessary to equip the system with the above-mentioned unimplemented address check function for cost reasons. Even so, the system must be equipped with an alternative means to prevent the system from losing system functionality. Such problems also occur during expansion.

The present invention was created based on the above-mentioned viewpoints,
It is an object of the present invention to provide a method for detecting non-implementation of a memory exchange unit, which can prevent loss of system functions due to setting of memory in a duplex system.

(Means for Solving the Problems) Fig. 1 shows an explanatory diagram of the principle of the present invention. This invention relates to a technique for recovering a single system to a duplex system after a system becomes a single system.The present invention is directed to a system that causes the system to operate normally (system operation system) When replacing faulty memory in another system with new memory, the other system does not immediately become a standby system, and the replaced memory is incorporated, causing the system to become overloaded. system(
For example, each address generation for writing and reading check data to a predetermined check storage area of the memory exchange unit of the memory 201 set completely or partially from the central processing unit (for example, 200) of the 0 series) Access information for the memory exchange unit is prepared, and whether or not there is an abnormality in the memory exchange unit is detected from the difference between write check data and read data.

(Operation) If a fault occurs in the memory of any system during operation of the FFI system, the faulty memory is removed and replaced with a new memory.

In addition to cases where the memory is composed of a single memory, it may also be composed of multiple memory exchange units, but especially in the latter case, the memory exchange unit may not be implemented properly.
Sometimes things are implemented with faults. For these preliminary detections, the access information is used to obtain a predetermined access address from the operating system to each memory exchange unit of the set memory of the system that will be incorporated as a standby system if the system is duplicated. Occur.

Using the predetermined access address, check data is written to and read from the storage area of the corresponding memory exchange unit. Based on the difference between the write check data and the read data, detection information indicating that the memory exchange unit is missing or has a fault is obtained.

If no abnormality is found in any of the memory exchange units, the memories made up of these memory exchange units are installed as standby memory, and the system is made to operate under the duplex configuration again.

〔Example〕

An embodiment of the present invention will be explained based on the duplex switching system configuration shown in FIG. This duplex exchange system is not equipped with an unimplemented address check function. In addition, each of the main storage devices 100 and 101 of both systems is composed of the same number of memory packages, and the corresponding memory packages (
The same address range is assigned to the memory capacity (for example, 4 MB), and the same address range is assigned to the package identification information (for example, a bit of "°1°" corresponding to the package arrangement order) (access for each memory exchange unit). A storage area (register) for storing information (information) is provided in a specific memory package (for example, the first memory package) that constitutes the main memory, and storage there is performed from a non-volatile output section provided in advance in the system. At startup, it is created in the storage area in the main storage of the active system.

Further, the 0-system central processing unit 120 is connected to the 0-system main storage device 10° and the 1-system main storage device IO, the O-system interfaces 11° and 11. An intersystem bus 15° connected to the 1st system central processing unit 12. 0 system main storage device 10° and 1 system main storage device 10. 1 system interface 13° and 13. An intersystem bus connected to 15. are shown as a single line in Figure 2 to avoid congestion in the drawing, but these buses 1
5. 15. are both address bus 4 and data bus 5.
and answer bus 6 (in Fig. 3, 0 series and °1
Shown without subscripts to distinguish between systems. ), the answer bus 6 is configured to be switched and connected only to the main storage device set as the active system. Although not shown in FIG. 2, as shown in FIG. 3, the main storage devices 100, 10I to the central processing units 120, 12. A failure notification line 7 extends to. This fault notification line 7 is connected to the main storage devices 100, 10 . writing.

This line notifies the central processing units 120 and 121 of a failure that occurs during reading, and the system configuration is such that it is switched and controlled so that it is effective only when the system operates as the active system.

2 and 3, the data bus 5 and the answer
Bus 6 corresponds to data bus 8 in FIG.

If there is a failure in all or part of the memory packages of the main storage device 101 of the standby system (for example, system 1) as a result of testing prior to operation of such a system or diagnosis during operation, For replacement, the main storage device or a part of the memory package will be removed, and measures such as its replacement will be taken.

In that case, the system - will have to operate under a unified system for some time, and will operate under reduced reliability.

In order to restore the system to its original duplex configuration, a new standby main storage device or its partial memory package is reset for use in place of the removed standby main storage device or its partial memory package. Prior to allowing the standby storage function to function normally, the following processing is performed.

When the process starts (31 in FIG. 4), each package start address is generated based on the package identification information sequentially read from the package identification information registration area, and is stored in a desired storage area, for example, the first memory. It is stored in the package storage area (32 in FIG. 4). At the same time as this storage process occurs, the answer bus 6 is transferred to the main storage device 10.
．． This causes a connection switch to . In addition, the main storage device 1
Writing to 0° is prohibited, and the fault notification line 7 remains connected.

A predetermined check pattern (check data) is written to a predetermined partial address (one or more addresses) of the memory package using the package start address (33 in Figure 4), and the written check pattern is transferred to the memory package. (34 in Figure 4). The difference between the read check pattern or check patterns and the written check pattern is checked (35 in FIG. 4).

If the values are unequal (≠ at 35 in FIG. 4), it indicates that the memory package must be removed or there is a failure, so the memory package is not installed in the standby main storage device.

If the values are equal (=35 in FIG. 4), it is determined whether there is an unchecked memory package or not based on the presence or absence of the next package identification information to be checked. If it remains (unchecked), then (
(NO in 36 in FIG. 4), the process returns to step S3 and repeats the same process.

If the check is completed (YES in 36 in Figure 4), copy the contents of the main storage device of the active system (for example, O system) to the main storage device 10+ of system 1 (37 in Figure 4), and copy the contents of the main storage device of system 1 (37 in Figure 4) It is incorporated into the system as a standby system of the active system (O system) (38 in FIG. 4).

Note that, as is clear from the above description, a separate method may be used to prepare the memory package identification information in the system in the above embodiment. It will also be clear that the present invention is not intended to be practiced solely with duplex digital switches. Semiconductor memories can be used as memories, and some of these semiconductor memories are configured with a single bus for writing and reading, so there is not necessarily a connection between the memory and the central processing unit. - It will also be clear that no bus need be provided.

〔Effect of the invention〕

As described above, according to the present invention, if a memory replacement unit is temporarily degenerated to a single system due to a memory failure, or the memory capacity is increased, or a system malfunction occurs without an automatic non-implementation check function, the memory replacement unit is Alternatively, even if there is a failure, system dysfunction caused by it can be prevented.

), 120, 12I is a central processing unit, 200, 20. is memory (main memory lO0, 101)
It is.

[Brief explanation of the drawing]

Fig. 1 is a diagram illustrating the principle of the present invention; Fig. 2 is a configuration diagram of a switching control section of a duplex digital exchange; Fig. 3 is a schematic diagram illustrating the present invention; Fig. 4 is a diagram showing the processing flow of the present invention. . In Figures 1 to 3, 4 is an address bus, 8 is a data bus (data bus 5, answer bus 6,

Claims (1)

[Claims] (1) connected to the central processing unit (12_0, 12_1) via an address bus (4) and a data bus (8);
Memory consisting of multiple memory exchange units (10_0, 10
In a redundant system that is not equipped with an automatic non-implementation check function for _1), the central processing of the system that operates the system (hereinafter referred to as the system operation system) prepares access information for each memory exchange unit. Memory (2
0_1) is the central processing unit for system operation (12_0)
Before copying the contents of the system operating memory (20_0), find the access address for each memory exchange unit determined according to the access information, write check data to the access address, and copy the data written to the access address. A method for detecting non-implementation of a memory exchange unit, characterized in that: when there is no sign between the written check data and the read data, an abnormality in the memory exchange unit corresponding to the non-sign is detected.