JP2007257395A - Fault monitoring method for application - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To ensure and facilitate a fault monitoring for an application in a computer system without a WDT function. <P>SOLUTION: The computer system is provided with a plurality of counters 2<SB>1</SB>to 2<SB>N</SB>corresponding to a plurality of applications 1<SB>1</SB>to 1<SB>N</SB>to be monitored and a monitoring task 3 to read and write count values for each counter. Each application increments the corresponding counter at a constant cycle during normal processing or at preset termination of a process. The monitoring task reads the count value of each counter at a constant cycle, sets all counters to "0" when all of each count values are within a defined range and carries out an abnormality alarm or a system reset for the application when even one of each count values is out of the range. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method of monitoring an application abnormality in a computer system that does not have a watchdog timer (WDT) function.

  In a computer system (a general hardware configuration is shown in FIG. 3) in which various information processing functions and control functions are established by installing a computer and its application, unexpected data destruction or program error on hardware or software. When present, it can cause system or application runaway.

  As a measure against this runaway, generally, a WDT function having a hardware configuration is provided, and a normal / abnormal alarm is generated or a system reset is performed depending on whether or not a reset output is output from the CPU within a predetermined time. (For example, refer to Patent Document 1).

In the monitoring by the WDT function, when a system and a high priority application run away, an alarm or a system reset is applied.
JP 2005-250524 A

  In a computer system that does not have a WDT function, a system reset cannot be performed when an application runs away.

  Even in a computer system having a WDT function, if the priority of a runaway application is low, the CPU load does not become a problem, so that the system is not reset.

  Further, in the WDT function, it is impossible to determine which application has runaway, so that subsequent system repair becomes difficult.

  An object of the present invention is to provide an application abnormality monitoring method that reliably and easily monitors application abnormality in a computer system having no WDT function.

  In order to solve the above-described problems, the present invention assigns a measurement memory (hereinafter abbreviated as a counter) to each application, and each application has a function of incrementing a counter associated with a certain period, and the count value of each counter The monitoring tasks are respectively read at regular intervals, and whether the application is normal or abnormal is determined based on whether or not they are within a defined range.

(1) An application abnormality monitoring method in a computer system having no watchdog timer function,
The computer system includes a plurality of counters associated with a plurality of applications to be monitored for abnormality among the applications, and a monitoring task capable of reading and writing the count value of each counter,
The plurality of applications are provided with a processing step for incrementing the counter associated with the end of a predetermined cycle or preset processing during the normal processing,
The monitoring task reads the count value of each counter at a constant period, sets all of the counters to “0” when all of the count values are within a defined range, and sets one of the count values. However, it is characterized in that a processing step for performing an application alarm or system reset when it is out of range is provided.

  As described above, according to the present invention, a counter is provided in association with each application, each application is provided with a function of incrementing a counter associated with a fixed period, and the monitoring task determines the count value of each counter at a fixed period. Since the normality / abnormality of the application is obtained by reading and whether or not they are within the defined range, it is possible to reliably and easily monitor the abnormality of the application in a computer system having no hardware WDT function. .

  In other words, the abnormality monitoring function can be easily realized by simply adding a plurality of counters and one monitoring task, and adding the counter increment function to the application, and directly monitoring the execution of the application process as the count value of the counter. By doing so, reliable monitoring can be performed.

  Furthermore, it becomes possible to identify an application in which an abnormality has occurred, and facilitate the repair work.

  Further, even in a computer system having a WDT function, a system reset is possible when an application with a low priority runs away.

FIG. 1 is a functional configuration diagram of an abnormality monitoring method showing an embodiment of the present invention. The computer system reads and writes a plurality of counters 2 ₁ to 2 _{N associated} with a plurality of applications (APL) 1 ₁ to 1 _N to be monitored for anomalies among the applications installed therein, and the count value of each counter A possible monitoring task 3 configuration is taken.

Each of the applications 1 ₁ to 1 _N to be monitored for abnormality is provided with a function for generating a command for incrementing the counter at a constant cycle or at the end of a preset process during the normal process. It should be noted that the function for generating the increment command may be limited to a case where it is provided in all applications installed in the computer system, and an application having a complicated software configuration with a high possibility of causing a runaway or a high priority application.

The counters 2 _{1 to} 2 _N are provided, for example, on a shared memory of the computer system in correspondence with an application having a function of generating an increment instruction, and from the increment operation by the corresponding applications 1 ₁ to 1 _N and other tasks. The count value can be read.

  The computer system includes, in the local memory in the monitoring task, a pair of an upper limit value and a lower limit value indicating the normal range of the count value of the counter corresponding to counter 1 to counter N as shown in Table 1.

Monitoring task 3, a sufficiently long predetermined period than increment instruction generation period by each application, the counter 2 ₁ to 2 reads the count value of _N, the upper and lower limits that all the count values have been established in Table 1 within the scope of, and has a function of issuing a counter reset instruction to set all the counters 2 ₁ to 2 _N to "0". If any _{one of} the counters 21 to _2N is out of the range of the upper limit value and the lower limit value defined, the monitoring task 3 performs an alarm or system reset (software reset).

FIG. 2 shows a processing flow provided for the application and the monitoring task. Each of the applications 1 ₁ to 1 _N executes a process (general process) inherent to each application (S1), and waits for a certain time as necessary (S2). As described above, this fixed time is a fixed period or the end of a preset process. After this fixed time, the associated counter value is incremented (S3), and if the application process is not finished, the process returns to process S1 and proceeds to the next process (S4).

In the monitoring task 3, first, a counter reset command is issued, all count values of the counters 2 ₁ to 2 _N are set to “0” (S11), and a predetermined time is waited (S12). As described above, this fixed time is sufficiently longer than the increment command generation cycle by each application. After this fixed time, the count values of the counters 2 ₁ to 2 _N are checked (S13), and it is determined whether or not these count values are within the upper limit value and the lower limit value defined in Table 1 (S14). When all the count values are within the range of the upper limit value and the lower limit value defined in Table 1, the process returns to step S11 and all the count values of the counters 2 ₁ to 2 _N are set to “0”. If even one of the count values is outside the range defined by the upper limit value and the lower limit value defined in Table 1, an alarm or system reset is performed as an application error (S15).

  Note that the values defined for the upper limit value and the lower limit value in Table 1 are not limited to the same value, and may be different values for each application. In this case, it becomes possible to cope with a difference in increment generation cycle due to a difference in application.

  In addition, when an alarm is generated, the monitoring task writes application identification information associated with a counter that falls outside the range of the upper limit value and the lower limit value defined in Table 1 to the non-volatile memory. It is possible to easily analyze the cause of abnormalities.

The function block diagram of the abnormality monitoring method which shows embodiment of this invention. The processing flow of the application and the monitoring task in the embodiment. The block diagram of a computer system.

Explanation of symbols

1 ₁ to 1 _N application 2 ₁ to 2 _N counter 3 Monitoring task

Claims (1)

An application abnormality monitoring method in a computer system that does not have a watchdog timer function,
The computer system includes a plurality of counters associated with a plurality of applications to be monitored for abnormality among the applications, and a monitoring task capable of reading and writing the count value of each counter,
The plurality of applications are provided with a processing step for incrementing the counter associated with the end of a predetermined cycle or preset processing during the normal processing,
The monitoring task reads the count value of each counter at a constant period, sets all of the counters to “0” when all of the count values are within a defined range, and sets one of the count values. However, an application abnormality monitoring method characterized by providing a processing step for performing an application abnormality alarm or system resetting when out of range.

Images