JP2723073B2 - Computer system failure log information acquisition method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fault log information output method for outputting fault log information when a fault occurs in a computer system, and more particularly, to a fault log in a computer system in which a CPU and the like are multiplexed to enhance fault tolerance. It relates to an information output method.

[0002]

2. Description of the Related Art In a computer system, when a fault occurs during operation and normal operation cannot be performed, fault log information relating to the fault is created. The user can analyze the cause of the failure by examining the failure log information and take a countermeasure.

In a system requiring a high degree of reliability of a processing result, such as a control system for a spacecraft or an artificial satellite, a financial system, or the like, a central processing unit including a CPU or the like is multiplexed into three or more layers. Thus, a computer system is realized in which the same processing is executed, the processing results of the central processing units are compared, and an appropriate processing result is adopted by majority decision. In such a computer system, when two or more (or a majority) CPUs always output the same processing result, the processing result is adopted, so that the reliability of the processing result is extremely high. Further, even if one CPU fails, the other two or more CPUs can continue processing, thereby improving the tolerance against the failure.
Even in such a computer system, a predetermined C
If the PU cannot operate normally due to a failure, failure log information for the CPU is created.

In this type of computer system, when a failure occurs in a predetermined CPU, the CPU is disconnected from the multiplexing of the central processing unit. For this reason, the collected failure log information has not been sent to the log file and has been stored in the memory in the failed central processing unit. Therefore, in order to obtain the fault log information, conventionally, a debugging terminal device or the like is connected to read out the fault log information stored in the memory in the central processing unit, or another multiplexed central processing unit. After the normal operation of, the faulty central processing unit is accessed to obtain fault log information and store it in a log file.

Further, a central processing unit for main operation and a spare central processing unit are provided to multiplex the central processing units, so that even if a failure occurs in the central processing unit for main operation, the spare central processing unit performs processing. A computer system with improved fault tolerance by continuing the above has also been realized. In such a computer system, if the central processing unit for main operation is disconnected from the system due to the occurrence of a failure, the failure log information is stored in the memory of the failed central processing unit, as in the case of the above system. It remains stored. Therefore, in order to acquire the fault log information, a terminal device for debugging is required, and it is necessary to wait for the end of the processing by the spare central processing unit.

[0006]

In the above-described fault information acquisition method in a conventional computer system, when a failure occurs in one of the multiplexed central processing units, the central processing unit including the CPU is not required. Was disconnected from the system, and the collected failure log information could not be stored in the log file. For this reason, there has been a drawback that a quick response cannot be taken in analyzing and maintaining the failure.

Further, in a computer system in which the processing results of a plurality of central processing units are compared and an appropriate processing result is adopted by a majority decision, a failure occurs in the plurality of central processing units and the central processing unit operating normally. If there is only one copy, it is impossible to make a majority decision to adopt the processing result, so the operation of the system itself is immediately stopped. Therefore, in such a case, there is a drawback that the fault information cannot be obtained.

The present invention solves the above-mentioned drawbacks of the related art, and quickly acquires failure log information from a central processing unit disconnected from the system due to a failure, thereby contributing to failure analysis and system maintenance. It is an object of the present invention to provide a failure log information acquisition method. Also, in a computer system that compares the processing results of a plurality of central processing units and adopts an appropriate processing result by majority decision, only one central processing unit that operates normally stops operation of the system itself. It is also an object of the present invention to provide a computer system failure log information acquisition method which acquires failure log information even in a case where the error occurs and contributes to prompt recovery of the system.

[0009]

In order to achieve the above object, the present invention provides a central processing unit which is provided in a multiplexed configuration of three or more and executes the same processing, and a processing result of each central processing unit. In a computer system including a majority decision comparison unit that adopts an appropriate processing result by majority decision in comparison,
Provided in each of the central processing units, when a failure occurs in the central processing unit, the failure log information related to the failure is transferred to another central processing unit and sent from the central processing unit in which the failure has occurred. The system includes a diagnostic processor that receives the failure log information and outputs it as output data, and a communication bus that interconnects the diagnostic processors provided in each of the central processing units.

In another aspect, the majority decision comparing unit operates normally when two or more central processing units fail and only one central processing unit operates normally. And a means for executing a logout mode for receiving only the failure log information output from the central processing unit.

[0011] In another aspect, the diagnostic processor includes a failure detection unit that detects a failure that has occurred in the central processing unit, a log collection unit that acquires failure log information regarding the failure that has occurred, Between the diagnostic processors of another central processing unit via the communication bus, and IOP transfer means for outputting the fault log information received from the diagnostic processor of the other central processing unit to the majority comparison unit Are provided.

In another fault log information acquisition method which achieves the above object, in a computer system comprising a multiplexed central processing unit and a means for separating a failed central processing unit from the system, When a failure occurs in the central processing unit, the failure log information on the failure is transferred to another central processing unit, and the failure log information sent from the failed central processing unit is provided. The diagnostic processor includes a diagnostic processor that receives and outputs the output as output data, and a communication bus that interconnects the diagnostic processors provided in each of the central processing units.

[0013]

According to the present invention, a diagnostic processor provided in each CPU set and interconnected by a communication bus collects failure log information of a failed CPU set and outputs the information from a normal CPU set. Thus, the failure log information can be obtained promptly. Also, in a computer system that compares the processing results of a plurality of central processing units and adopts an appropriate processing result by majority decision, when only one normal central processing unit is used, the output from the normal central processing unit is output. If the system shifts to the logout mode that accepts only the error log information and outputs the error log information, even if the operation of the system itself stops because only one normal central processing unit is used, Can acquire the failure log information.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram illustrating a configuration of a computer system that implements a failure log information acquisition method according to one embodiment of the present invention. The computer system of the present embodiment is a computer system that compares the processing results of a plurality of central processing units and adopts an appropriate processing result by majority decision, and describes a case where the central processing units are tripled.

As shown in the figure, the computer system of the present embodiment includes a CPU set 10, 20, a central processing unit.
30; a majority decision unit 50 for inputting and comparing the execution results of the CPU sets 10, 20, and 30; and a file device 60 including a log file 61 for storing log information.

The CPU sets 10, 20, and 30 have the same configuration and perform the same processing as is apparent from the drawing.
Thus, the central processing unit of the computer system according to the present embodiment is tripled, and the fault tolerance is improved. CP
U sets 10, 20, and 30 are provided with CPUs 11, 2,
1, 31, memories 12, 22, 32, and an IOP (Input / Output Process) for data input / output.
sors) 13, 23, 33 and CPU sets 10, 2,
Diagnosis processor (Diagn) for diagnosing states of 0 and 30
ossis Processor (hereinafter, appropriately referred to as DGP) 14, 24, and 34.

The diagnostic processors 14, 24, 34 are shown in FIG.
As shown in (1), a failure detection unit 41 that detects a failure that has occurred on the CPU set 10, 20, 30; a log collection unit 42 that acquires failure log information related to the failure; Transfer unit 43 (hereinafter referred to as DG) for transferring to another diagnostic processor via
And transfer the received failure log information to the IOPs 13, 23, and 3 of each of the CPU sets 10, 20, and 30.
IOP transfer unit 4 for outputting to majority decision comparing unit 50 through 3
4 is provided. Also, as shown in FIG. 1, the diagnostic processors 14, 24, 34 are interconnected via a communication bus 80.

The majority decision comparing section 50 includes a CPU set 10,
A data comparing unit 51 that compares the contents of the data 70 output from the devices 20 and 30; and a file storage unit 52 that stores the failure log information in the log file 61 of the file device 60 when the input data 70 is failure log information. And is provided. Here, the data 70 includes a tag 71 indicating the type of the data and a data body 72 as shown in FIG.
Consists of

In a normal operation, the data comparing section 51 stores data 7 inputted from each of the CPU sets 10, 20, and 30.
Compare 0. And two or more CPU sets 10,
When data 70 of the same content is input from 20, 30, the data 70 of the same content is input to the CPU sets 10, 20,.
This is adopted as the processing result of Step 30. It should be noted that the data 70 input from each of the CPU sets 10, 20, and 30 have different contents, or that any two CPU sets 1
If a failure occurs in 0, 20, and 30 and only one normal processing result is obtained, the system itself of this embodiment is stopped because a majority decision cannot be made. However, this does not apply to the case where the operation of the system is in the logout mode and failure log information is output, as described later.

The file storage unit 52 includes a data comparison unit 51
As a result of the comparison, when it is confirmed that the contents of the data 70 are the same and that the type of the data 70 is not the failure log information with reference to the tag 71, the data 70 is stored in a predetermined area of the file device 60. Store.

The operation of the computer system configured as described above will be described. Usually, CPU set 10,
20 and 30 always execute the same processing while maintaining synchronization. Each execution result output by each of the CPU sets 10, 20, 30 is output to the majority decision comparing section 50 via the IOPs 13, 23, 33. At this time, a tag 71 indicating the type of the data is added in front of the data body 72 as the execution result, and the majority comparison unit 5
Output to 0.

Data 70 output from each of the CPU sets 10, 20, and 30 is input to the majority decision comparing section 50. Then, the data comparing unit 51 checks each data 7
A comparison is made to determine whether the contents of 0 are the same. When the data comparison unit 51 confirms that the contents of the respective data 70 are the same and that the data body 72 is data other than the failure log information, the file storage unit 72
Stores the data in a predetermined area of the file device 60.

Next, a description will be given of an operation of acquiring a fault log according to the present embodiment. First, a triple CPU set 1
The acquisition of failure log information when one of 0, 20, and 30 has failed will be described with reference to the block diagram of FIG. 3 and the flowchart of FIG.

It is assumed that a failure has occurred in the CPU set 10 for some reason. At this time, the failure detection unit 41 of the diagnostic processor 14 of the CPU set 10 detects the failure, and the log collection unit 42 collects failure log information (steps 401 and 402). Next, D of the diagnostic processor 14
The inter-GP transfer unit 43 stores the collected fault log information in FIG.
Are transferred to the diagnostic processors 24 and 34 of the CPU sets 20 and 30 via the communication bus 80 as shown by arrows (step 403).

From the CPU set 10 to the CPU set 20,
The failure log information transferred to the diagnostic processor 2
4 and 34 are received by the inter-DGP transfer unit 43. When the failure log information is received, IO of the diagnostic processors 24 and 34 is performed.
The P transfer unit 44 adds a tag 71 indicating that the data body 72 is failure log information, and outputs the data to the majority comparison unit 50 via the IOPs 23 and 33 (step 40).
4).

When the data 70 output from the CPU sets 20 and 30 is received by the data comparing section 51 of the majority comparing section 50, the majority comparing section 50 transmits the data from the CPU set 10 corresponding to the received data 70. Wait for reception of 70. However, due to a failure in the CPU set 10, C
There is no output of data 70 from PU set 10. Therefore, the data comparing unit 51 disconnects the CPU set 10 from the computer system of the present embodiment at a predetermined timing such as when a predetermined time elapses (steps 405 and 406). Next, the data comparison unit 5
1 is data 70 received from CPU sets 20 and 30
Is transferred to the file storage unit 52 (step 407).

The file storage 52 stores the received data 7
When it is confirmed from the tag 71 added to "0" that the data is fault information, the data body 72 as fault information is stored in the log file 61 of the file device 60 (steps 408 and 409). As described above,
Failure log information in the case where a failure has occurred in one of the tripled CPU sets 10, 20, 30 is acquired.

In the flowchart of FIG. 4, when the majority decision comparing section 50 receives data from the CPU sets 20 and 30 and then receives an input from the CPU set 10 (step 405), the CPU set 10 is disconnected. When it is determined that the data is not the failure log information when the data is transferred to the file storage unit 52, the process ends. This is because attention has been paid to the operation of acquiring the failure log information. In practice, it goes without saying that in each case, the processing shifts to an appropriate processing.

Next, the triple CPU set 10, 2
The acquisition of failure log information when two failures 0 and 30 occur will be described with reference to the block diagram of FIG. 5 and the flowchart of FIG.

It is assumed that a fault has occurred in the CPU set 20 while the fault has occurred in the CPU set 10 described above. At this time, similarly to the diagnostic processor 14 of the CPU set 10, the failure detecting unit 41 of the diagnostic processor 24 of the CPU set 20 detects the failure of the CPU set 20, and the log collection unit 42 collects the failure log information (step S1). 601 and 602). Next, the inter-DGP transfer unit 43 of the diagnostic processor 24 transmits the collected failure log information to the diagnostic processor 3 of the CPU set 30 via the communication bus 80.
4 (step 603).

The fault information transferred from the CPU set 20 is stored in the DG of the diagnostic processor 34 of the CPU set 30.
It is received by the inter-P transfer unit 43. When the failure log information is received, the IOP transfer unit 44 of the diagnostic processor 34 sends a tag 7 indicating that the data body 72 is the failure log information.
1 is added and output to the majority decision comparing unit 50 via the IOP 33 (step 604).

Data 7 output from CPU set 30
When 0 is received by the data comparing section 51 of the majority decision comparing section 50, the majority comparing section 50 waits for reception of the data 70 from the CPU sets 10 and 20 corresponding to the received data 70. However, there is no output of the data 70 from the CPU sets 10 and 20 due to a failure that has occurred in the CPU sets 10 and 20. Therefore, at a predetermined timing, the data comparing unit 51 disconnects the CPU sets 10 and 20 from the computer system of this embodiment (step 605,
606). Thereafter, the majority comparison unit 50 shifts to a logout mode in which only the failure log information output from the CPU set 30 is received (step 606). When confirming that the data comparison unit 51 is in the mouthout mode, the data comparison unit 51 transfers the data 70 received from the CPU set 30 to the file storage unit 52 without performing majority processing (step 607).

The file storage 52 stores the received data 7
When it is confirmed from the tag 71 added to 0 that the data is fault log information, the data body 72 as fault log information is stored in the log file 61 of the file device 60 (steps 608 and 609). As described above,
Failure log information in the case where a failure has occurred in two of the tripled CPU sets 10, 20, 30 is obtained.

In the flowchart of FIG. 6, when the majority decision comparing section 50 inputs data from the CPU set 30 and then receives an input from the CPU set 10 or 20 (step 605), the CPU set 10 is disconnected. When it is determined that the data is not the failure log information when the data is transferred to the file storage unit 52, the process ends. This is because attention has been paid to the operation of acquiring the failure log information. In practice, it goes without saying that in each case, the processing shifts to an appropriate processing.

Although the present invention has been described with reference to the preferred embodiments, the present invention is not necessarily limited to the above embodiments. For example, in the present embodiment, the computer system in which the central processing unit is tripled has been described. However, the present invention can be similarly applied to a computer system in which four or more multiplexes are performed.

In this embodiment, a computer system which compares the processing results of a plurality of central processing units and adopts an appropriate processing result by majority decision has been described. However, a central processing unit for main operation and a spare central processing unit are used. The present invention can be similarly applied to a computer system in which a central processing unit is multiplexed by providing a central processing unit. in this case,
In some cases, the central processing unit has a duplicated configuration, but even in that case, the majority comparison means and the like are no longer necessary.
It does not affect the failure log information acquisition method which is a characteristic technique of the present invention.

[0037]

As described above, according to the present invention, the above-mentioned conventional disadvantages can be solved, and failure log information can be promptly obtained from the central processing unit disconnected from the system due to the occurrence of a failure. Efficiency in failure analysis and system maintenance can be improved. Also, in a computer system that compares the processing results of a plurality of central processing units and adopts an appropriate processing result by majority decision, only one central processing unit that operates normally stops operation of the system itself. In such a case, the failure log information can also be acquired, so that the system can be quickly restored.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a computer system that implements a failure log information acquisition method according to one embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration of a diagnostic processor.

FIG. 3 is a block diagram illustrating a state of the computer system when a failure occurs in one CPU set.

FIG. 4 is a flowchart illustrating an operation of acquiring failure log information when a failure occurs in one CPU set.

FIG. 5 is a block diagram illustrating a state of the computer system when a failure occurs in two CPU sets.

FIG. 6 is a flowchart illustrating an operation of acquiring failure log information when a failure occurs in two CPU sets.

[Explanation of symbols]

 10, 20, 30 CPU set 11, 21, 31 CPU 12, 22, 32 Memory 13, 23, 33 IOP 14, 24, 34 Diagnostic processor 41 Failure detection unit 42 Log collection unit 43 Inter-DGP transfer unit 44 IOP transfer unit 50 Majority comparison unit 51 data comparison unit 52 file storage unit 60 file device 61 log file 70 data 71 tag 72 data body 80 communication bus

Claims (5)

(57) [Claims] 1. A central processing unit which is multiplexed into three or more types and executes the same processing, and a majority comparing unit which compares processing results of the central processing unit and adopts an appropriate processing result by majority decision. A computer system provided with each central processing unit, wherein when a failure occurs in the central processing unit, failure log information relating to the failure is stored in all other units.
And transfers to the central processing unit of the Te, and the diagnostic processor to be output to the majority comparing unit for receiving the fault log information which the failure is sent from the other central processing unit occurs, provided to the central processing unit Dedicated for connecting the diagnostic processors to each other and transmitting and receiving the fault log information
A communication bus , wherein the majority comparison unit is configured to output the fault from the diagnostic processor.
A failure log information acquisition method for a computer system, comprising: means for storing harm log information . 2. The system according to claim 1, wherein the majority comparison unit has a fault in two or more central processing units, and only one of the central processing units operates normally.
Other central processing that failed from the central processing unit
When the trouble log information is received by the
Check that there is no data input from the processing
Separating the processing unit, without performing the majority decision processing
The diagnostic processor of the central processing unit operating normally;
Failure of other central processing unit where the failure output from the server has occurred
Shift to logout mode that accepts only log information
The method according to claim 1, further comprising means . 3. The diagnostic processor according to claim 1, wherein the diagnostic processor detects a failure that has occurred in its own central processing unit, a log collection unit that acquires failure log information related to the failure that has occurred, and stores the acquired failure log information in the diagnostic processor. Communication means for transferring to a diagnostic processor of another central processing unit via a communication bus; and transfer means for outputting fault log information received from the diagnostic processor of another central processing unit to the majority comparison unit. The failure log information acquisition method for a computer system according to claim 1 or 2, wherein: 4. A computer system comprising a multiplexed central processing unit and means for separating a failed central processing unit from the system, wherein the computer system is provided in each of the central processing units, and the central processing unit has a failure. If an error occurs, the error log information on the error
And a diagnostic processor that receives the failure log information sent from the other central processing unit in which the failure has occurred and outputs it to a log file, and a diagnostic processor that transfers the failure log information to the central processing unit. Communication for interconnecting the diagnostic processors and transmitting and receiving the fault log information
A failure log information acquisition method for a computer system, comprising: a bus ; 5. The diagnostic processor according to claim 1, wherein the diagnostic processor detects a failure that has occurred in its own central processing unit, a log collection unit that acquires failure log information regarding the failure that has occurred, and stores the acquired failure log information in the diagnostic processor. Communication means between diagnostic processors for transferring to a diagnostic processor of another central processing unit via a communication bus, and transfer means for outputting failure log information received from the diagnostic processor of another central processing unit to the log file The method for acquiring fault log information of a computer system according to claim 4, wherein: