Trace information collecting system and program

Abstract

Occurrence of a failure in a computer system is appropriately detected, and information required for removing the failure is automatically collected. A trace information collecting system of the present invention which collects trace information indicating the course of processing of a program comprises a history recording section for recording a history of run and stop of a target program for which the trace information is to be produced, a similarity calculating section for calculating a degree of similarity between a first operation pattern in which the target program ran and stopped at the last time and a second operation pattern in which the target program ran and stopped at the time before last when the target program, which once stopped, starts to run, a failure occurrence determining section for determining that a failure occurred when the target program ran at the last time on condition that the similarity calculated by the similarity calculating section is lower than a reference similarity, and a trace information collecting section for collecting the trace information from the target program in response to the determination that the failure occurred, wherein the trace information is not collected if it is determined that the failure did not occur.

1 - 14 . (canceled) 
   
   
       15 . A trace information collecting system for collecting trace information showing a processing history of a program, comprising:
 a history recording subsystem for recording a run/stop history of a target program that is a target for generating trace information;   a similarity calculating subsystem for calculating, as a degree of similarity, a value indicating a deviation in a probability distribution of time ratios of a running time and a stop time each time there is repetitive running and stopping, a last time that the target program was run, and prior to the last time that the target program was run, of the time ratios of the running time and the stop time in a repetitive running and stopping the last time that the target program was run, when starting an operation of the target program that has been stopped;   a failure occurrence determining subsystem for determining that a failure occurred the last time the target program ran, on condition that a similarity calculated by the similarity calculating subsystem is lower than a reference similarity; and   a trace information collecting subsystem for collecting sequentially trace information, which is not collected when there is no determination that the failure has occurred, from a started target program in response to a determination that the failure occurred.   
   
   
       16 . The trace information collecting system according to  claim 15 , further comprising:
 a run/stop time ratio calculating subsystem for calculating a run/stop time ratio each time there is repetitive running and stopping of the target program, that is a ratio of an elapsed stop from the stopping of the target program until the running and an elapsed running time from running until stopping; wherein
 the similarity calculating subsystem calculates the value indicating the deviation of the run/stop time ratio of a last run and stop in a probability distribution of the run/stop time ratio at each repetition; and 
 the failure occurrence determining subsystem determines that the similarity is lower than the reference similarity if the value indicating the deviation calculated by the similarity calculating subsystem is greater than a reference value. 
   
   
   
       17 . The trace information collecting system according to  claim 15 , wherein the failure occurrence determining subsystem further determines that a failure has occurred when an elapsed stop time from the target program stopping until running is shorter than a reference stop time is shorter than the reference stop time. 
   
   
       18 . The trace information collecting system according to  claim 15 , wherein the history recording subsystem further records a type of a stop procedure for stopping the target program when the target program is stopped; and
 the failure occurrence determining subsystem determines that the failure has occurred on condition that the stop procedure which stopped the target program at the last time is an immediate stop procedure that interrupts and stops processing already started, and does not determine that the failure has occurred on condition that the stop procedure is a stop after completion procedure, which stopped a processing after completion of the already started processing.   
   
   
       19 . The trace information collecting system according to  claim 18 , further comprising a score calculating subsystem for calculating a score indicating the possibility that the failure occurred in the target program by taking a higher value when the similarity calculated by the similarity calculating subsystem is lower, taking the higher value, when compared with a case of the stop after completion procedure, when the stop procedure that stopped the target program at the last time was the immediate stop procedure, and taking the higher value when the elapsed stop time until running the target program after a previous stop was shorter than a reference stop time; and
 the failure occurrence determining subsystem determines that the failure has occurred on the condition that the score this calculated by the score calculating subsystem is higher than a reference score that is set in advance, and does not determine that the failure has occurred on condition that the score is at or lower than the reference score.   
   
   
       20 . The trace information collecting system according to  claim 18 , wherein the failure occurrence determining subsystem further determines that the failure has occurred on condition that the stop procedure that stopped the target program at the last time was an immediate stop procedure, and a number of processes interrupted in the immediate stop procedure is greater than a reference number. 
   
   
       21 . The trace information collecting system according to  claim 15 , further comprising a run time deviation calculating subsystem for calculating the value indicating the deviation of an elapsed run time from when they target program started to run last time until the target program stopped last time, in a probability distribution of a length of a run time in the target program repeating a run and stop, based on a history of the run and stop recorded by the history recording subsystem; wherein
 the failure occurrence determining subsystem further determines that the failure has occurred if the value indicating the deviation calculated by the run time deviation calculating subsystem is greater than a reference value.   
   
   
       22 . The trace information collecting system according to  claim 15 , further comprising: an operation parameter collecting subsystem for collecting operation parameters to be supplied to the target program in response to a determination by the failure occurrence determining subsystem that the failure has occurred. 
   
   
       23 . The trace information collecting system according to  claim 15 , further comprising:
 an average run time calculating subsystem for calculating an average run time from the time when the target program started to run until the target program stopped, based on a history of a run as recorded by the history recording subsystem; wherein   the trace information collecting subsystem collects the trace information, which is not collected it is not determined that the failure has occurred, in response to the determination that the failure has occurred, and finishes a collection of the trace information if the average run time has elapsed from a start of running of the target program.   
   
   
       24 . The trace information collecting system according to  claim 15 , further comprising:
 a file input/output use ratio detecting subsystem for detecting a frequency of file input/output of an information processing device that executes the target program; wherein   even if collecting the trace information, the trace information collecting subsystem finishes a collection of the trace information if the frequency of the file input/output exceeds a predetermined reference frequency.   
   
   
       25 . The trace information collecting system according to  claim 15 , further comprising:
 a central processing device use ratio detecting subsystem for detecting a use ratio of the central processing device of an information processing device that executes the target program; wherein   even if collecting the trace information, the trace information collecting subsystem finishes a collection of the trace information if the use ratio of the central processing device exceeds a predetermined reference use ratio.   
   
   
       26 . A trace information collecting program for causing an information processing device to function as a system for collecting trace information showing a processing history of a program, wherein:
 the information processing apparatus is caused to function as:
 a history recording subsystem for recording a run/stop history of a target program that is a target for generating trace information; 
 a similarity calculating subsystem for calculating, as a degree of similarity, a value indicating a deviation in a probability distribution of time ratios of a running time and a stop time each time there is repetitive running and stopping, a last time that the target program was run, and prior to the last time that the target program was run, of the time ratios of the running time and the stop time in a repetitive running and stopping the last time that the target program was run, when starting an operation of the target program that has been stopped; 
 a failure occurrence determining subsystem for determining that a failure occurred the last time the target program ran, on condition that the similarity calculated by the similarity calculating subsystem is lower than a reference similarity; and 
 a trace information collecting subsystem for collecting sequentially trace information, which is not collected when there is no determination that a failure has occurred, from a started target program in response to the determination that the failure occurred.