JPH05151187A - Fault logging system for plural processors - Google Patents

Abstract

PURPOSE:To perform processing by surely delivering fault log data to master processor by saving unprocessed fault log data in a file when the system down of a slave processor occurs. CONSTITUTION:When a fault occurs in the main body system/peripheral system device of the master processor (a), a master processor log data extraction means 11 extracts information with respect to the number of a device in which the fault occurs and the kind of the fault. etc. Thence, an exclusive control mechanism 5 checks whether or not data being written on a log file 13 exists, and instructs standby when it exists, and informs that write can be performed after all the data being written are processed. At a time when the write can be performed. a data write means 6 writes the data extracted by the master processor log data extraction means 1 on the log file 13.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fault logging system for multiple processors, and more particularly to a fault logging system for collectively managing fault logs in a system having multiple processors.

[0002]

2. Description of the Related Art In the conventional fault logging method for a plurality of processors, as shown in FIG. 2, when a fault occurs in a main system / peripheral system device of a mass processor, a master processor log data collecting means 21 for collecting the fault information. And, when a failure occurs in the main system / peripheral system device of the slave processor, the log data collecting means 22 for collecting the failure information and the log data collected by the log data collecting means 2 are passed from the slave processor to the master processor. An inter-communication mechanism 23, a log data receiving means 24 for receiving log data from a slave processor by the inter-processor communication mechanism 3, a log file 30 for storing log data of a master processor and a slave processor,
Data writing means 26 for writing the log data collected by the master processor log data collecting means 21 or the log data of the slave processor received by the log data receiving means 4 to the log file 30, and a plurality of log data at the same time in the log file 30. Exclusive control mechanism 25 that controls writing so as not to cause data inconsistency
A data designation input means 27 for receiving the designation of the log data to be edited and output from the terminal device, and a data editing means 28 for editing the data of the log file 30 based on the designation received by the data designation input means 27. It has a data output means 29 for outputting the data edited by the data editing means 28 to an output device.

Next, the operation of this prior art will be described.

When a failure occurs in the main system / peripheral system device on the master processor, the master processor log data collection means 21 collects information such as the number of the failure device and the type of failure.

The exclusive control mechanism 25 checks whether or not there is data being written in the log file 30, and if there is any, instructs the waiting. Then, when there is no data being written, the fact that writing is possible is notified.

When the writing becomes possible, the data writing means 26 writes the log data in the log file 30.

When a failure occurs in the operation of the main system / peripheral system on the slave processor, the log data collecting means 22
Collects information such as the number of the device that caused the failure and the type of failure. Next, the inter-processor communication mechanism 3 notifies the master processor of the log data collected by the slave processor. Then, the log data receiving means 24 receives the notified log data and reflects it in the buffer on the master processor. After that, when the exclusive control mechanism 5 notifies that the data can be written, the data writing unit 26 writes the log data to the log file 30.

When the data of the log file 30 is to be edited and output, first, the data designation input means 27 receives from the terminal device designation such as device designation processor designation of the log data to be edited and output. Then, the data editing means 2
Reference numeral 8 retrieves from the log file 30 the log data matching the designation received by the data designation input means 27 and edits it. Finally, the data output means 29 is the data editing means 28.
The data edited by is output to the output device.

[0009]

In this conventional fault logging method, unprocessed log data on the slave processor is saved in a file when a shutdown command is executed or a fault occurs and the system is stopped. Since there is no log data, there was a risk that the collected log data would be lost.

Even if the log data is transmitted using the inter-process communication mechanism during the system stop processing, there is a possibility that the system may stop before the log data can be sent if the remaining data is large.

An object of the present invention is to prevent failure log data that has not been processed from being taken over even after the system of slave processors is stopped when collectively managing failure log data of a plurality of processors. To provide a logging method.

[0012]

According to the fault logging method for a plurality of processors of the present invention, when a fault occurs in a main system / peripheral system device of a master processor, master processor log data collecting means for collecting the fault information, When a failure occurs in the main / peripheral devices of the slave processor,
Log data collection means for collecting the failure information, an inter-processor communication mechanism for passing the log data collected by the log data collection means from the slave processor to the master processor, and a log for receiving log data from the slave processor by the inter-processor communication A data receiving unit, a log file storing log data of the master processor and the slave processor, and log data collected by the master processor log data collecting unit or log data of the slave processor received by the log data receiving unit in a log file. A data writing unit for writing, an exclusive control mechanism for simultaneously writing a plurality of log data to a log file so as not to cause data inconsistency, and a system for stopping a slave processor system due to a failure or the like. A save file that temporarily stores the log data of the slave processor, a log data save means that saves the log data of the slave processor to the save file when necessary, and interprocessor communication after the failure of the master processor is recovered In the case of, or when the slave processor is started up, log data loading means for extracting log data from the save file, data specification input means for receiving specification of log data to be edited and output from the terminal device, and the data specification input means. And a data output means for outputting the data edited by the data editing means to the output device.

[0013]

The present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing the configuration of the present invention.

The present invention collectively manages fault logging data in a system having a plurality of processors. In the present embodiment, a fault logging method in the case of having two processors will be described.

FIG. 3 is a flow chart showing the processing of the log data collection / storing method when a failure of the main system / peripheral system device occurs on the master processor.

Referring to FIG. 3, the main system / peripheral fault log collection and storage processing on the master processor includes a log collection processing A1, an exclusive control processing A2, and a log writing processing A3.

When a fault occurs in the main system / peripheral system device on the master processor, the master processor log data collecting means 1 collects information such as the number of the faulty device and the type of fault (step A1). Next, the exclusive control mechanism 5
It is checked whether or not there is data being written in the log file 13 and, if there is, an instruction to wait. Then, when there is no more data being written, it is notified that writing is possible (step A2). When it becomes writable, the data writing means 6 writes the data collected by the master processor log data collecting means 1 in the log file 13 (step A3).

Further, referring to FIG. 4, the main system / peripheral system fault logs on the slave processor are collected and stored as a log collection process B1, a communication process B2, and an exclusive control process B3.
And a log writing process B4. When a failure occurs in the main system / peripheral system device of the slave processor, the log data collecting means 2 collects information such as the number of the device in which the failure has occurred and the type of failure (step B1). Next, the inter-processor communication device 3 notifies the master processor of the failure log data collected by the slave processor. Further, the log data receiving means 4 receives the notified failure log data and reflects it in the buffer on the master processor (step B2). After that, the exclusive control mechanism 5 checks whether or not there is data being written in the log file 13 and, if there is, instructs to wait. When there is no more data being written, the fact that writing is possible is notified (step B3). When it becomes writable, the data writing means 6 writes the data received from the slave processor in the log file 13 (step B4).

When the slave processor is stopped, the log data is temporarily saved in a file by the log data saving process shown in FIG. 5, and when restarted, the log data is restored by the log data loading process shown in FIG. ..

The log data saving process of FIG. 5 includes a saving determination process C1 and a log data writing process C2.

The log data loading process of FIG. 6 comprises a save file reading process D1, a notification judging process D2, and a communication process D3. When stopping the system of the slave processor, the log data saving means 7 first checks whether or not there is log data that has not been notified to the master processor (step C1). If there is unreported log data, the log data saving means 7 writes all unreported log data in the save file 9 (step C
2). When restarting the slave processor, the log data loading means 8 first reads the log data from the save file 9 (step D1). Check whether there is log data saved when the system is stopped (step D
2). If there is saved log data, the interprocessor communication mechanism 3 notifies the master processor (step D3). The log data notified to the master processor is the exclusive control process B3 and the log write process B4 shown in FIG.
Is stored in the log file 13.

Next, a process for editing and outputting the log data of the log file 13 will be described. Referring to FIG. 7, the data edit output process includes a data designation input process E1, a data edit output process E2, and a data output process E3. First, the data designation input means 10 receives designations such as device designation, date designation, and processor designation of log data to be edited and output from the terminal device (step E1). Then, the data editing means 11 takes out the log data specified by the data specification input means 10 from the log file 13 and edits it (step E2). Finally, the data output means 12 outputs the data edited by the data editing means 11 to the output device (step E3).

In the above manner, even if the system of the slave processor is stopped midway, the failure log data of a plurality of processors can be properly managed and analyzed collectively.

Incidentally, even when there are two or more slave processors, it is possible to collectively manage and analyze them by the same principle.

[0026]

As described above, according to the present invention, the unprocessed fault log data can be saved in a file when the system of the slave processor is stopped, and can be loaded from the file when the system is restarted and processed. This has an effect that the failure log data of can be reliably passed to the master processor for processing.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of the present invention.

FIG. 2 is an overall configuration diagram of a conventional technique.

FIG. 3 is a flow chart showing the processing of log data collection and storage method when a failure occurs in the main system / peripheral system device on the master processor.

FIG. 4 is a flow chart showing the processing of log data collection and storage method when a failure of the main system / peripheral system device occurs on the slave processor.

FIG. 5 is a flowchart showing log data saving processing when the slave processor system is stopped.

FIG. 6 is a flowchart showing a log data loading process when the system of the slave processor is restarted.

FIG. 7 is a flowchart showing the processing of the log data edit output method.

[Explanation of symbols]

 1 master processor log data collecting means 2 log data collecting means 3 inter-processor communication mechanism 4 log data receiving means 5 exclusive control mechanism 6 data writing means 7 log data saving means 8 log data loading means 9 save file 10 data designation input means 11 data Editing means 12 Data output means 13 Log file

Claims (1)

[Claims] 1. In a system having a plurality of processors, when a failure occurs in a main processor / peripheral device of a master processor, master processor log data collecting means for collecting the failure information, and a main processor / peripheral system of a slave processor. When a failure occurs in the device, log data collection means for collecting the failure information, an inter-processor communication mechanism for passing the log data collected by the log data collection means from the slave processor to the master processor, and the inter-processor communication Log data receiving means for receiving log data from the slave processor, a log file for storing log data of the master processor and the slave processor, log data collected by the master processor log data collecting means or received by the log data receiving means A data writing unit that writes the log data of the slave processor that was taken to a log file, an exclusive control mechanism that controls multiple log data to the log file at the same time so as not to cause data inconsistency, and the slave processor is stopped due to a failure or other reason. When saving the log data of the slave processor, the save file that temporarily stores the log data, the log data save means that saves the log data of the slave processor to the save file when necessary, and the communication between the processors after the failure of the master processor Log data loading means that retrieves log data from the save file when the slave processor is started up, data specification input means that receives the specification of the log data to be edited and output from the terminal device, and the data specification Received by input means A fault logging system for a plurality of processors, comprising: a data editing means for editing the data in the log file based on the designation; and a data output means for outputting the data edited by the data editing means to an output device. ..