JPH03250227A - Fault information control system for distributed system - Google Patents

Abstract

PURPOSE:To reduce the load of a line by counting fault information and transmitting only the fault information exceeding a threshold value to a concentrated monitoring host. CONSTITUTION:When the no-response fault of a terminal is generated and a terminal address and a state code are present on a record accumulated in a storage table 123, a fault information control processing part 121 adds '1' to the fault count of the accumulated record and updates polling transmission time. On the other hand, when no accumulated record is present thereon, the accumulated record is prepared from the terminal address, state code and polling transmission time and written into the accumulating table 123. Next, the fault information control processing part 121 compares the terminal address and state code of the accumulated record with each terminal address and state code of each reference record in a reference table 122. When those terminal addresses and state codes are coincident with each other and the fault count of the stored record is larger than the threshold value of the reference record, the accumulated record is transmitted to the concentrated monitoring host and the fault count is reset. Thus, the load of the line is reduced.

Description

[Detailed Description of the Invention] [Industrial Application Field] This method is applicable to a distributed machine in a network system.
This invention relates to a method for controlling the transmission of failure information to a centralized monitoring host.

C. Prior Art] As described in Japanese Patent Publication No. 62-217750, the conventional method is to provide a management center with means for accumulating polling result information from nodes and means for estimating the location of a failure. It was supposed to reduce the operator's work in estimating the location of a failure when a terminal does not respond.

[Problem to be solved by the invention]

In the above conventional method, when polling is sent from the distribution machine to the terminal,
Even if the terminal was powered off or disconnected, it would become unresponsive and send fault information to the central monitoring host. However, although information on failures caused by power-off or disconnection of terminals is unnecessary for isolating the failure, the distribution machine
Suppression of failure information was not considered, and there were problems of increased line load and increased load on the central monitoring host.

This method uses a distributed machine to suppress unnecessary unresponsive failures, reduce the load on the line, and reduce the load on the central monitoring host.

[Means to solve the problem]

In order to achieve the above object, there is provided a fault information control processing unit that controls sending fault information to a centralized monitoring host, and a terminal address where a non-response fault has occurred.

By providing the distribution machine with an accumulation table having multiple accumulated records consisting of status codes, fault counts, and polling sending times, and a reference table having multiple reference records consisting of terminal addresses, status codes, and threshold values,
If there is no response after the distributed machine sends polling to the terminal,
The terminal address, status code, and polling sending time are held in memory, and the failure information control processing unit sequentially refers to the terminal address and status code of each accumulated record in the accumulation table, and if they match, the failure count of the accumulated record is set. is added by 1 and the polling sending time is updated. On the other hand, if there is no match, an accumulation record consisting of the above terminal address, the above status code, and a failure count (value 1) is written to the above accumulation table, and the above accumulation record is If the difference between the polling sending time and the immediately preceding polling sending time is greater than the fixed time interval, the failure count of the accumulated record is reset.

Next, in order to determine whether or not to send the stored record to the centralized monitoring host, each reference record in the reference table is sequentially referenced, and if the terminal address and status code of the stored record match, the fault is detected. The counter is compared with the threshold, and if the fault counter is larger than the threshold, the accumulated record is sent to the central monitoring host and the fault counter is reset, while the terminal address and If there is no reference record with a matching status code, the accumulated record is sent to the centralized monitoring host and the failure count in the accumulated record is reset.

[Effect]

If a non-response failure occurs in a terminal and the terminal address and status code exist in the accumulated record in the accumulated table, the failure information control processing unit adds 1 to the failure count of the accumulated record and sets the polling sending time. On the other hand, if the storage record does not exist, a storage record is created from the terminal address, status record, and polling sending time and written in the storage table. Next, the fault information control processing section compares the terminal address and status code of the stored record with the terminal address and status code of each reference record in the reference table, and if they match, and the fault count of the stored record is If it is larger than the threshold value of the reference record, send the accumulated record to the central monitoring host and reset the fault count, while if the terminal address and status code in the accumulated record do not exist in the reference table. , immediately sends the accumulated record to the central monitoring host, and resets the failure count of the accumulated record in the accumulated table.

〔Example〕

Examples of this system will be described below with reference to the drawings.

FIG. 1 is a diagram showing an embodiment of this method, in which failure information control processing 121. The reference table 122, storage table 123, and related mechanisms are the means necessary to implement this method.

First, the distribution device 12 connected to a plurality of terminals 124 sends polling to each terminal 124 to inquire whether there is a data transfer request or not at a fixed time interval. If there is a response, polling will not be sent to the terminal from now on until the transmission/reception is completed.

When it is detected that there is no response from a certain terminal after a certain period of time has elapsed since the polling was sent, the terminal address, the status code and time at the time of polling inversion are held in memory, and the failure information control processing 121 performs the second The terminal address and status code of each stored record in the storage table 123 shown in the figure are sequentially referred to, and if a matching stored record exists, the failure count is incremented by 1, the polling sending time is updated, and the matching stored record If it does not exist, write the terminal address, status code, and polling sending time in the storage table 123 as a failure count (value 1) storage record, and refer to each storage record to find the polling sending time and the previous one of the stored record. If the difference in polling sending times is greater than the fixed time interval, the failure count of the accumulated record is reset.

Next, the fault information control processing 121 sequentially refers to the terminal address and status code of each reference record in the reference table 122 shown in FIG. , and if the failure count of the accumulated record is greater than the threshold of the reference record, the accumulated record is sent to the centralized monitoring host 11 and the failure count is reset. On the other hand, if no matching reference record exists,
The accumulated record is immediately sent to the central monitoring host 11. Note that the centralized monitoring host receives the accumulated records and displays them on the display device 111. Figure 4 is a time chart showing the values of polling (ENQ) L, normal response (ACK) or abnormal response (NAK), and fault count when the distributed machine sends polling to the terminal.After sending polling to the terminal 4 times, there is a normal response. , data transmission and reception with the terminal starts, and then there is an abnormal response (NAK), the failure count is reset, polling is retransmitted, and the threshold value (failure count value 5
) and sends fault information to the central monitoring host.

〔Effect of the invention〕

According to this method, when polling is sent from the distributed machine to each terminal,
Even if there is no response, failure information is not sent to the central monitoring host immediately, but by counting failure information and sending only failure information that exceeds a threshold to the central monitoring host, the failure information is sent to the central monitoring host. Since the amount of information can be suppressed, the load on the line can be reduced, and failures can be easily isolated from the terminal address, status code, and failure count sent to the central monitoring host.

[Brief explanation of drawings]

Figure 1 is a configuration diagram of a network that performs unified fault management from a centralized monitoring host in an embodiment of this method, and Figure 2 is a reference reference for determining whether or not to send fault information to the centralized monitoring host. Figure 3 shows the format of an accumulation table that counts failure information when a failure occurs; Figure 4 shows polling requests (ENQ) and normal responses (A
11 is a time chart showing failure count values when receiving an abnormal response (NAK) and an abnormal response (NAK). DESCRIPTION OF SYMBOLS 11... Central monitoring host, 12.13... Distributor, 14... Line, 111... Display device, 121... Fault information control processing unit, 123... Accumulation table, 122... - Reference table, 124...Terminal. Part 2 Diagram 1st Sainogu

Claims (1)

[Claims] 1. In a network system consisting of a central monitoring host, a plurality of distributed machines, and a plurality of terminals connected to the distributed machines, the distributed machine inquires about the presence or absence of a data transfer request by sending polling to the terminals, If there is no response after a certain period of time, the address of the terminal, the status code indicating the status at the time of polling, and the polling sending time are held in memory, and the failure information control processing unit stores the terminal address, status code, failure count, and Refers to an accumulation table that has multiple accumulated records consisting of polling sending times, adds 1 to the failure count of the accumulated record with the matching terminal address and status code, updates the polling sending time, and updates the polling sending time of the accumulated record. If the difference between the polling transmission time and the previous polling sending time is larger than the above fixed time interval, the failure count is reset, and then a reference table having multiple reference records consisting of terminal addresses, status codes, and failure count thresholds is referred to. and only when the terminal address and status code match and the failure count is greater than the threshold, the accumulated record is transmitted to the central monitoring host and the failure count of the accumulated record is reset. fault information control method for distributed systems.