JP5871499B2 - COMMUNICATION CONTROL DEVICE, COMMUNICATION CONTROL SYSTEM, COMMUNICATION CONTROL METHOD, AND PROGRAM - Google Patents

Description

  The present invention relates to a communication control device, a communication control system, a communication control method, and a program.

  In inter-process communication in which a plurality of software components on a computer node operate while communicating with each other, there is a technique capable of data communication like a subroutine call (function call) (for example, SunRPC). According to this method, data can be exchanged with another process (which may be on the same node or another node) using inter-process or inter-node communication.

  On the other hand, software components that simulate function calls in a process are arranged in a plurality of processes and a plurality of computer nodes, and the software components are reused to improve software development efficiency and quality.

  However, when performing communication (remote procedure) between processes or nodes using the same interface as the function call in the process, a communication error may occur. Software written to call functions in the process cannot directly detect communication errors in remote procedures. For example, in the specification of the function call in the process, except for the case where the called function itself returns an error, such as a specification in which the return value of the function is void or a value in which only the value indicating the execution result of the function is returned. This is because failure is not assumed.

  In order to solve this problem, methods using exception handlers and callback handlers (hereinafter collectively referred to as “handlers”) are used (catch syntax in Java (registered trademark) language, signal handler in C language). Etc.). In this method, a thread for performing error processing is prepared in the calling process while the specification of the function call does not consider the error. At the same time, when a communication error is detected, a special function for error processing or error recovery is executed by this thread, thereby detecting and responding to the communication error without affecting the function specification.

  Regarding error processing techniques that do not use handlers, Patent Documents 1 and 2 have been proposed as prior art.

  Patent Document 1 discloses a configuration in which a verification program code for error verification is added to each process in advance in a system that executes a plurality of processes. A method is disclosed in which a system error such as a communication error is detected by a verification program code and system recovery is performed. Since Patent Document 1 detects an error using a specially assigned program code, the specification of the function return value is not intended to detect the occurrence of a communication error.

  Patent Document 2 discloses a configuration for detecting an error in communication from the server side to the client side by providing a timeout mechanism on the client side in the remote call of the function. Here, the client side on which the timeout mechanism is implemented refers to the remote function calling mechanism itself, and the error detection mechanism is not included in the implementation of the function calling portion.

Japanese Patent No. 03290052 JP 2001-282656 A

  In the error detection method using a handler, it is necessary to implement an error detection unit for each process. In view of the purpose of improving the development efficiency by reusing the software that has made a function call in the process as it is as a program for performing a remote procedure, it is not desirable because the amount of code correction increases for each process. From the viewpoint of reducing the amount of program code correction, a method of implementing an error detection unit in the remote procedure call library is also conceivable. However, this method implements an error detection function, return from error, and error handling function in the remote procedure call library. The error detection function / error processing function generally depends on the hardware on which the software system operates. When software that strongly depends on hardware is implemented in a library, when porting the library to a system that operates on other hardware, there is a problem that versatility is reduced from the viewpoint of dependency on hardware. is there.

  In Patent Document 1, it is necessary to add a verification program code for error verification to each process in advance. This verification program code is generally called a watchdog, a heartbeat or the like. In this method, it is necessary to implement exception handling for each process, as in the method using exception handlers and callback handlers. Therefore, in a distributed system composed of a large number of processes, there is a problem in terms of development efficiency because of the large amount of code that must be implemented.

  In Patent Document 2, the communication error that can be detected is limited to the error at the time of transmission from the server side to the client side, and a timeout mechanism must be implemented on the client side. For this reason, there is a problem in terms of development efficiency because of the large amount of implementation code as in the method of Patent Document 1, and in the supplementary performance of error occurrence because the error from the client side to the server side cannot be detected.

  In most cases, a distributed system has a log system shared by a plurality of processes in the system, and each process has a log output function for outputting to the log system. Generally, in a distributed system, multiple processes cooperate to realize software functions, and in order to analyze system operations and errors based on logs, they do not belong to a specific process. A unified log is required for the entire system. This is the reason why the log system exists.

  An object of the present invention is to provide a communication control technique that uses a log output function shared by a plurality of processes for error detection.

A communication control device according to one aspect of the present invention that achieves the above object is as follows:
A communication process between a plurality of applications in the same process, a communication process between one application of the plurality of applications and another process other than the process in which the application exists, or the application exists A communication means for converting into communication processing with another device different from the device, and performing the converted communication processing;
Error detection means for detecting a communication error that occurs during execution of the communication processing by the communication means;
Identifying means for identifying the type of the communication error and the importance of the communication error corresponding to the type of the communication error from the correspondence between the predetermined type of communication error and the degree of importance;
Exception processing means for executing error processing when the importance identified by the identification means is a certain value or more,
It is characterized by providing.

  According to the present invention, it is possible to provide a communication control technique using a log output function for error detection.

  By using the log output function also for error detection, it is not necessary to install an error detection unit in each process, and the amount of code correction on the call side is small, improving software reusability and improving development efficiency. It becomes possible.

  Or, by detecting and notifying a communication error indirectly via an error status flag, the function call in the process is applied to general-purpose remote call processing that is less dependent on a specific hardware function. It becomes possible.

The figure which shows the structural example of the communication control apparatus concerning embodiment of this invention. The figure explaining the function structure with which a remote call library is provided. The figure which shows the correspondence of the classification of the log output from an information storage part, and the level of importance of a log. The figure explaining the flow of processing at the time of error detection. The figure explaining the flow of the process accompanying the monitoring of the error state flag by an error notification part, and a monitoring result. The figure explaining the flow of the error process by an exception process part.

  The present invention provides a technique for performing communication processing by applying communication within the same process to communication between processes or between nodes without significant change in program code. Here, the communication within the same process means, for example, communication for calling a function of another application from one application and transmitting / receiving information between the applications, and hereinafter, this communication is also called a function call. In addition, communication between processes or between nodes means, for example, communication between different processes or between communication control devices (nodes) for calling functions of one application from another application and transmitting / receiving information to / from each other. Say. Hereinafter, this communication is also referred to as remote call.

  A configuration of a communication control apparatus according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating a configuration example of a communication control apparatus constituting a communication control system. A CPU 102, a network controller 103, and a memory 104 are arranged on a communication control device (node 01) 101 that constitutes a communication control system, and are connected to each other by a system bus 105. The memory 104 is divided for each process by a memory protection mechanism provided in the CPU 102. When the process A and the process B are operating in the communication control apparatus (node 01) 101, the memory space 106 for the process A and the memory space 107 for the process B exist in the memory 104.

  In addition to these, on the memory 104, there are an error status flag 108, an error notification unit 109, and an exception processing unit 110 that executes error processing that exist only for each node. In the memory space 106 of the process A, an application 111 operating on the process A and a remote call library 112 linked to the application 111 are arranged. Similarly, in the memory space 107 of the process B, an application 113 and a remote call library 114 linked to the application 113 are arranged. These are arranged for each process.

  The communication control apparatus (node 01) 101 is connected to the communication control apparatus (node 02) 116 via the communication path 115 established by the network controller 103. The communication control device (node 02) 116 also has the same configuration as the communication control device (node 01) 101, and the processes A and B of the communication control device (node 01) 101 perform communication control via the communication path 115. Communication with the processes C and D of the device (node 02) 116 is possible.

  FIG. 2 is a diagram for explaining a functional configuration of the remote call library 112 in FIG. In a system having the communication control device (node 01) 101 and the communication control device (node 02) 116, the process A is operating in the communication control device (node 01) 101, and the application 111 is executed on the memory space 106 of the process A. Is working.

  The application 111 is programmed to call a function belonging to another application existing in the memory space 106 of the same process A and perform communication processing. The remote call library 112 is linked with the application 111 in the memory space 106 of the process A.

  The remote call library 112 includes a communication unit 206 that performs communication processing. The communication unit 206 converts function call processing (communication processing) in the process from the application 111 into communication processing with another process or communication processing with another communication control device, a communication protocol, It is characterized by performing communication processing. For example, the communication unit 206 converts function call processing (subroutine call) in the process from the application 111 into communication processing (remote call) between other applications, other processes, or other nodes. .

  Other applications that are communication partners of the application 111 may be arranged in the memory space 106 of the same process A as the application 111. Other applications are arranged on the same communication control apparatus (node 01) 101 or on a communication control apparatus (node 02) 116 which is a different node from the communication control apparatus (node 01) 101. Also good.

  Further, the remote call library 112 includes an error detection unit 207 that detects an error related to communication, and an information storage unit 208 that outputs an operation state of the library itself including communication. In this embodiment, it is assumed that a function for outputting an error log to a storage medium is implemented as the information storage function of the information storage unit 208. Here, the recording medium may be a specific file on the magnetic disk, a specific area on the memory, or a storage area secured on a dedicated machine for log recording.

  The remote call library 112 includes a recording information identification unit 209 that determines the type of log output from the information storage unit 208 when a communication error is detected by the error detection unit 207 and identifies the importance of the communication error. When a communication error is detected by the error detection unit 207 and an error log related to the communication error is output from the information storage unit 208, the recorded information identification unit 209 causes the exception processing unit 110 to perform exception processing. It is determined whether or not.

  Further, the remote call library 112 includes a flag setting unit 210 that changes the error state flag 108 from the non-signal state to the signal state when the recording information identification unit 209 determines that the communication error is important. Here, the non-signal state indicates that the system is operating normally, and the signal state indicates the occurrence of a communication error state.

  Further, an error notification unit 109 is provided on the communication control apparatus (node 01) 101. The error notification unit 109 notifies the exception processing unit 110 that an error has occurred when the flag setting unit 210 changes the error state flag 108 to a signal state, and executes exception processing (error processing). Prompt.

  FIG. 3 is a diagram illustrating a correspondence relationship between the log type 301 output from the information storage unit 208 and the log importance level 302. The correspondence relationship between the communication error type (log type) and the log level indicating the importance of the communication error corresponding to the communication error type is stored in advance in the memory 104 of the communication control device or an external hard disk. It is possible to leave.

  The log importance level 302 and the log type 301 have a one-to-one correspondence. The recorded information identification unit 209 can determine the log importance level 302 corresponding to the log type by identifying the log type 301 output from the information storage unit 208. In the present embodiment, the smaller the numerical value of the log importance level 302, the higher the importance. That is, the log type LOG_FATAL 303 corresponding to “1” of the log importance level 302 (log level) is the most important log type. LOG_FATAL 303 represents an attribute of a serious error that cannot be recovered, such as a hardware error or an operating system internal error. The log type corresponding to the log level “2” is LOG_ERROR 304, which represents an important error such as a communication error that cannot be continued. The log type corresponding to the log level “3” is LOG_WARNING 305 and represents a minor abnormality that allows the processing to continue, such as a processing timeout. The log type corresponding to the log level “4” is LOG_INFO 306, which is a log that is output even in a normal state for proactively knowing the operation state of the system. The log type corresponding to the log level “5” is LOG_DEBUG307, and is a log for outputting a detailed internal operation state of the system that is used only during system debugging or failure analysis.

  The recorded information identification unit 209 can determine whether to operate the flag setting unit 210 using these log levels, and can control the flag setting unit 210 based on the determination result. For example, the recording information identification unit 209 can activate the flag setting unit 210 only when LOG_FATAL (log level 1) or LOG_ERROR (log level 2) is output as the log type. On the other hand, in the case of other log types, the recording information identification unit 209 can control the flag setting unit 210 not to operate.

  If there is no log attribute that represents the type of error according to the system specifications, the recorded information identification unit 209 can activate the flag setting unit 210 under the condition that an important error log of a certain level or more has been output. Is possible. If the log level cannot be specified, the record information identifying unit 209 can also operate the flag setting unit 210 for all error logs.

(Processing flow when an error is detected)
A processing flow at the time of error detection will be described with reference to FIG. First, when a “communication error” is detected by the error detection unit 207 during execution of the communication processing by the communication unit 206, this processing starts (step S401). Next, in step S402, the recorded information identifying unit 209 determines the type of the detected communication error log and identifies the importance of the communication error. For example, “communication error” corresponds to the log attribute LOG_ERROR 304 shown in FIG. 3 and the corresponding log level is “2”.

  When the log type is determined, the information storage unit 208 outputs a log in step S403.

  The information storage unit 208 controls a log output destination in a hard disk external to the communication control device (node 01) 101 or a specific area on the memory 104 of the communication control device (node 01) 101 according to the level of the log level. Is possible. For example, the information storage unit 208 can be controlled to output an external hard disk or memory 104 log only for log levels 1 and 2 with high importance.

  When log output is performed in step S403, in step S404, the recording information identification unit 209 identifies whether or not the log level output in the previous step S403 is a certain level or higher. The processing in this step corresponds to setting a threshold value for determining whether to perform exception processing. By setting the log level to a certain level or more, it is possible to control so that exception processing is not activated by a slight log that can be output even in a normal state.

  If it is determined in step S404 that the log level is less than a certain value (S404-N0), the process ends (S406). If the log level is more than a certain value (S404-Yes), the process proceeds to step S405. A flag setting process is executed. In step S405, the record information identifying unit 209 updates the error state flag 108 by operating the flag setting unit 210, and sets the signal level indicating the occurrence of a communication error state in which the log level is higher than a predetermined value. To do.

(Flag monitoring flow by the error notification unit 109)
FIG. 5 is a diagram for explaining the monitoring of the error state flag 108 executed by the error notification unit 109 and the flow of processing according to the monitoring result. In step S501, the error notification unit 109 starts monitoring the error status flag 108 at an arbitrary time after the system is started. Here, the monitoring is from a non-signal state indicating that the log level of the error state flag 108 set as an initial state is less than a certain value to a signal state indicating that the log level is greater than or equal to a certain value. It means detecting change. The monitoring method may be polling for checking the error state flag 108 at regular intervals, or the error notification unit 109 may receive information indicating a change in the flag state passively by a system-specific method such as an event. Also good.

  When the error state flag 108 detects a setting change from the non-signal state to the signal state during monitoring (S502-Yes), in step S503, the error notification unit 109 notifies the exception processing unit 110 of the occurrence of an error (error notification). ) The specific processing contents of the exception processing unit 110 that has received the error notification will be described in detail with reference to FIG.

(Error processing flow by the exception processing unit 110)
FIG. 6 is a diagram illustrating the flow of error processing by the exception processing unit 110 that has received an error notification from the error notification unit 109.

  First, in step S601, when the exception processing unit 110 receives an error notification from the error notification unit 109, this processing starts. In step S602, the exception processing unit 110 displays that an error has occurred (error state) on the user interface (UI) screen. Thereafter, in step S603, the exception processing unit 110 terminates all processes belonging to the system. In step S604, the exception processing unit 110 ends the exception processing by shutting down the hardware (step S605).

  This exception processing is processing performed when the system falls into a serious abnormal state and depends on the hardware on which the application operates. The exception processing in FIG. 6 is an example of stopping the entire system and shutting down the hardware. In the case of a recoverable error, the process may be restarted without ending. Further, a plurality of exception processing flows may be executed according to the degree of error (log level). Also, it may be a process of stopping or restarting hardware, a process of stopping or restarting all or part of an application, or a process of recovering from a system-specific error. May be.

  According to the present embodiment, a communication control technique using the log output function for error detection is provided. By using the log output function also for error detection, it is not necessary to install an error detection unit in each process, and the amount of code correction on the call side is small, improving software reusability and improving development efficiency. It becomes possible. By detecting and notifying communication errors indirectly via error status flags, function calls in the process can be applied to general-purpose remote call processing with less dependence on specific hardware functions. It becomes possible.

(Other examples)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

Claims (8)

A communication process between a plurality of applications in the same process, a communication process between one application of the plurality of applications and another process other than the process in which the application exists, or the application exists A communication means for converting into communication processing with another device different from the device, and performing the converted communication processing;
Error detection means for detecting a communication error that occurs during execution of the communication processing by the communication means;
Identifying means for identifying the type of the communication error and the importance of the communication error corresponding to the type of the communication error from the correspondence between the predetermined type of communication error and the degree of importance;
Exception processing means for executing error processing when the importance identified by the identification means is a certain value or more,
A communication control apparatus comprising:   The apparatus further comprises flag setting means for changing the setting of a flag indicating a normal communication state to a setting indicating the occurrence of a communication error when the importance is identified as a certain value or more by the identification means. The communication control apparatus according to 1. Storage means for storing a correspondence relationship between the type of the communication error and the importance of the communication error corresponding to the type of the communication error;
3. The information storage unit according to claim 1, further comprising: an information storage unit that outputs the type of the communication error to the storage unit when the importance level identified by the identification unit is equal to or greater than a predetermined value. Communication control device.   The communication control apparatus according to claim 1, wherein the exception processing unit executes a plurality of exception processing flows according to the importance level identified by the identification unit. 5. The communication control device according to claim 1, wherein the communication unit performs communication by converting a subroutine call between the plurality of applications into a communication protocol with another device. 6. . In a plurality of communication control devices, a communication control system for calling a function of an application of the other communication control device from an application of one communication control device and transmitting / receiving information to / from each other,
The communication control device includes:
A communication process between a plurality of applications in the same process, a communication process between one application of the plurality of applications and another process other than the process in which the application exists, or the application exists A communication means for converting the communication processing between the other communication control device different from the communication control device and performing the converted communication processing;
Error detection means for detecting a communication error that occurs during execution of the communication processing by the communication means;
Identifying means for identifying the type of the communication error and the importance of the communication error corresponding to the type of the communication error from the correspondence between the predetermined type of communication error and the degree of importance;
An exception processing unit that executes an error process when the degree of importance identified by the identification unit is equal to or greater than a predetermined value. A communication control method in a communication control device,
The communication means of the communication control device performs communication processing between a plurality of applications in the same process between one application of the plurality of applications and another process other than the process in which the application exists. A communication process for converting the communication process into a communication process with another communication control apparatus other than the communication control apparatus or the communication control apparatus in which the application exists , and performing the converted communication process;
An error detection step in which the error detection means of the communication control device detects a communication error that occurs during the execution of the communication process by the communication step;
Identification in which the identification means of the communication control device identifies the type of the communication error and the importance of the communication error corresponding to the type of the communication error from the correspondence between the predetermined type of communication error and the degree of importance Process,
Exception processing means of the communication control device, when the importance identified by the identification means is a certain value or more, an exception processing step of executing error processing;
A communication control method characterized by comprising:   The program for functioning a computer as each means of the communication control apparatus of any one of Claims 1 thru | or 5.

Images