Method for processing a diagnosis of a processor, information processing system and a diagnostic processing program

• US 20050166089A1
• Filed: 12/22/2004
• Published: 07/28/2005
• Est. Priority Date: 12/25/2003
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Ito, Masanao, Sakakibara, Tadayuki
• Granted Patent: US 7,421,618 B2
• Time in Patent Office: Days
• US Class Current: 714/11
• CPC Class Codes: G06F 11/0724 in a multiprocessor or a multi-core unit multiprocessors per se G06F15/80
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring

Abstract

In an information processing system including one or a plurality of processors, a diagnostic program is executed with a predetermined frequency to diagnose the processors. The diagnostic program generates one or a plurality of processes or threads at a predetermined frequency, at predetermined time intervals, for example, to diagnose the processors. A generated process or thread executes diagnosis on each processor, and the process or thread that detected a fault in the processor finishes its execution by storing fault information in storage. The process or thread of another processor other than the faulty processor refers to the fault information about the faulty processor and executes a troubleshooting process.

Claims

1. A method for performing diagnosis of an information processing system including one or a plurality of processors by using a diagnostic program, comprising the steps of:

setting a frequency of execution of diagnosis;

starting the diagnostic program to generate processes or threads for diagnosing the information processing system at said frequency; and

executing specified diagnosis on each of said processors by using said process or thread generated.

2. The method for performing diagnosis according to claim 1, wherein said setting step is to set the frequency and amount of diagnosis executed and said executing step is to terminate the diagnosis within the specified amount.

3. The method for performing diagnosis according to claim 2, wherein said frequency is a time interval between diagnoses executed and said amount is a length of time spent for one diagnosis.

4. A method for performing diagnosis of one or a plurality of processors constituting an information processing system by a diagnostic program, comprising the steps of:

setting a frequency of execution of diagnosis;

starting the diagnostic program to generate processes or threads for diagnosing each of the processors at said frequency; and

executing specified diagnosis on each of said processors by said process or thread generated.

5. The method for processing diagnosis according to claim 4, wherein said setting step is to set a frequency and an amount of diagnosis executed and said execution step is to terminate the diagnosis within said amount.

6. The method for processing diagnosis according to claim 5, wherein said frequency is a time interval between diagnoses executed and said amount is a length of time spent for one diagnosis.

9. The method for performing diagnosis according to claim 4, further comprising the step of obtaining the number of said processors, wherein said generating step is to generate said processes or threads according to said number of the processors.

10. The method for performing diagnosis according to claim 5, further comprising a step of specifying the contents of diagnosis according to the amount of said diagnosis executed, wherein said executing step is to execute said specified contents of diagnosis.

11. The method for performing diagnosis according to claim 5, wherein said contents of diagnosis are an aggregate of a plurality of independently executable diagnostic items, and wherein said specifying step is to select and specify at least one of the diagnostic items that can be executed within said amount.

12. The method for performing diagnosis according to claim 4, wherein in said generating step, said diagnostic program is started by a monitoring program, the method further comprising a step of monitoring a running state of said diagnostic program started by said monitoring program, and stopping the diagnostic program in response to an error found in the diagnostic program.

7. The method for processing diagnosis according to claim 7, further comprising the steps of:

on detecting occurrence of a fault on said one of the processors, the related process or thread storing fault information in memory means, and terminating its operation; and

another process or thread not having detected occurrence of the fault in said processor referring to said fault information on the other processors and, if any fault is detected, performing a specified fault process.

8. The method for performing diagnosis according to claim 7, wherein the step of performing said fault process includes a step of cutting off said faulty processor from the system.

13. An information processing system comprising:

a plurality of processors; and

storage for holding OS, a diagnostic program, execution frequency specification information, diagnostic contents specification information, and diagnosed state information,

wherein said OS starts said diagnostic program on an arbitrary processor, wherein said diagnostic program generates processes or threads for diagnosing each of said processors with a frequency based on said execution frequency specification information, and wherein specified diagnosis is executed on each of said processors by said process or thread generated.

14. A computer program for diagnosing one or a plurality of processor constituting an information processing system comprising the steps of:

deciding a frequency at which to execute diagnosis;

starting said diagnostic program to generate processes or threads for diagnosing each of said processors; and

executing specified diagnosis on each of said processors by said process or thread generated.