JP2782971B2 - Debugging method for distributed information processing system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed information processing system for distributing and processing a plurality of programs by controlling a plurality of computers in relation to each other, and more particularly to a parallel, parallel or distributed processing of such programs. The present invention relates to a debugging method suitable for each process that operates in parallel in the system.

[0002]

2. Description of the Related Art In recent years, due to the spread of networks, lower prices of computers, and the complexity and diversification of processing required for computers, each of a plurality of processes communicates with each other through communication means via communication channels. Cooperate in parallel
A program consisting of processes operating in parallel is described. In particular, programming of a distributed program comprising the above-mentioned processes operating in parallel and in parallel distributed to a plurality of computers connected by a wide area network or a narrow area network is performed. In such programming, the development of a method for easily describing a process that cooperates in parallel / parallel operation and a method for debugging is a major issue.

[0003] One method of easily describing processes that operate in parallel / parallel in cooperation is to separate programs into a server that provides processing and a client that uses the processing results, and describe them. There is a way to modularize and program. In particular, one of the methods of separating the functions of the server and the client is a method known as remote procedure call (for example, “Front line of distributed computing environment”, Nikkei Electronics, 19
90, 6, 11; 502, pp. 122-148). In the remote procedure call, a unit of processing is regarded as a procedure, a procedure is designated to a process on the same or a remote computer, an argument is transmitted, and a return value of the procedure is received. The processing is performed by reflecting the side effect of the procedure in the state of the information processing system. By introducing this remote procedure call, it is possible to easily program a program consisting of processes that operate in parallel and in parallel. One of the major issues in debugging is how to handle processes that operate in parallel and in parallel in a simple manner. Conventional debugging methods deal only with a single process, but recently debugging methods that can handle processes that operate in parallel or in parallel have been proposed (for example, (1) D. Caswell, Black "Impleme
nting a March Debugger For Multithreaded Applicati
ons "Proc. of USENIX, Winter '90, Nov. 1989 pp 25-
39, (2) `` Domain Distributed Debugging Enviroment
Reference ”Apollo Computer Inc., Order No. 011024-A0
0, (3) See JP-A-1-200563). In addition, a debugging method has been proposed which incorporates a mechanism for managing information on the occurrence of communication failures and retries relating to the remote procedure call into the remote procedure call function to assist in debugging the remote procedure call (for example, R. Cooper " Pilgrim: A
Debugger for Distributed Systems '' Proc. Of 7th Int
Conf.on Distributed Computing Systems, pp.458-46
5, Sep. 1987).

[0004]

However, in the conventional program debugging method, a function capable of debugging a process that operates in parallel and in parallel is considered as an important point.
In particular, there is a problem that it is difficult to debug a program of a process operating in parallel / parallel using a remote procedure call as easily as a program of a single process. SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems of the conventional program debugging method, and has been described in an information processing system which is realized by a plurality of processes communicating with each other using a remote procedure call. It is an object of the present invention to provide a program debugging method that can be used.

[0005]

According to the present invention, as shown in FIG. 1, a plurality of processes 1 and 2 are executed on a plurality of computers connected to each other by a network.
(Process executed in 203 or 213 in FIG. 2) communicates with each other by remote procedure call to realize a parallel / parallel operation. A plurality of satellite debugging means 3 and 4 (20 and 21 in FIG. 2) for debugging each process on a computer to be executed, and a central debugging means 5 for remotely controlling the operations of the plurality of satellite debugging means 3 and 4 (22 in FIG. 2). Then, the central debugging means 5 includes the satellite debugging means 3 (or 4).
Remote procedure call issuance detecting means 51 (224 in FIG. 2) for detecting before the issuance of the remote procedure call that the process to be debugged has a remote procedure call, and ending the process called by the remote procedure call And a remote procedure call termination detecting means 52 (225 in FIG. 2) for detecting the remote procedure call is detected by the remote procedure call issue detecting means 51 in response to detection of the remote procedure call. Control means 53 for activating the satellite debugging means and in response to the remote procedure call completion detecting means detecting the end of the process, restarting the debugging by the satellite debugging means which called the process
(222 in FIG. 2).

[0006]

In the above configuration, the first satellite debug means 3
When one process (hereinafter, referred to as a client process) 1 that is being debugged issues a remote procedure call, the remote procedure call issuance detecting means 51 detects the issuance of the remote procedure call, and executes the requested remote procedure call. Other processes including procedures (hereinafter, server processes)
2 is specified. The central debugging means 4 activates the second satellite debugging means 5 on the computer on which the server process 2 including the required remote procedure is running. Further, the switching control unit 53 performs communication between the central debugging unit 5 and the satellite debugging unit from the first satellite debugging unit 3 to the second debugging unit.
By switching to the satellite debugging means 4, the person who debugs the client process can use the server process 2
Is performed following the debugging of the client process 1. When the execution of the remote procedure ends, the remote procedure call end detecting means 52 detects the end of the execution of the remote procedure call. The switching control unit 53 instructs the second satellite debugging unit 4 to end the debugging of the server process 2. When the debugging of the server process 2 ends, the operation of the second satellite debugging means 4 ends. Then, the activation state is switched so that the first satellite debugging means 3 performs debugging. Thus, the person who debugs the server process 2 can debug the client process continuously from the debugging of the server process 2. As described above, the present invention
Since the satellite debugging means of the server process and the client process can be appropriately controlled, the person who debugs the communication procedure or the communication part of the argument when the remote procedure call is issued, or the remote procedure call is terminated. The communication procedure and the communication part of the return value are not visible, and it is as if the client process executed a procedure call on a single process that was not a remote procedure and returned the return value. Therefore, the debugger can use the conventional debugging technique for a single process, and debug a parallel / parallel distributed program that easily calls a remote procedure without being aware of the parallel / parallel operation of the process. Can be.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The features of the present invention will be specifically described below based on embodiments with reference to the drawings. FIG. 2 is a block diagram showing an embodiment of the present invention. In FIG. 2, the remote procedure call description file defines remote procedure names, types and numbers of arguments of the remote procedure, and return types. For example, in FIG. 5, as an example of the remote procedure call description file,
It indicates that the remote procedure name is printmessage, the argument has one string type, and the return value has int type. In the present embodiment, it is assumed that a first source code debugging unit 206 is connected to a client process executed by the client process execution unit 203 in advance. However, after the client process is activated, the first source code debug unit 206 may be connected after knowing the process number of the client process. The first source code debug unit 206 is connected to the debugger interface 222 of the central debug unit 22 via a communication channel of the communication unit 23 through the first TCP socket 231. This first TCP socket 231 is, specifically, a TCP connected to the first satellite debug unit 20 side.
The socket and the TCP connected to the central debugging unit 22
It consists of sockets and communication lines connecting these TCP sockets. Similarly, other sockets and access units of the communication unit 23 also include communication lines and interface units at both ends.

The person who debugs issues a command from the display terminal device 223 of the central debugging unit 22, and the command is sent to the first source code debugging unit 206 through the debugger interface 222. The first source code debugging unit 206 debugs the source code with reference to the client source file in the client source file storage unit 201. The execution result of the command given to the first source code debug unit 206 is sent to the debugger interface 222 via the communication channel by the first TCP socket 231 and displayed on the display terminal device 223.

For example, when a step execution command for instructing the execution of the source code one step at a time is issued to a client process stopped at a break point, the first source code debug unit 206
Causes the client process execution unit 203 to execute step execution, and when the client process stops again, outputs the current stop point in the source code to the first breakpoint detection unit 224.

[0010] The first breakpoint detecting section 224 includes:
The first remote file access unit 232 refers to the client source file in the client source file storage unit 201 from the current stop point in the source code of the client process, and determines whether the current stop point is a procedure call. Is checked, and if it is not a procedure call, nothing is output to the debugger interface 222. If the current stop point is a procedure call, the file in the remote procedure call description file storage unit 226 is referred to. If a remote procedure name that matches the procedure name of the procedure call exists, the current stop point is It is regarded as a line for calling a remote procedure in the client process, and the debugger interface 222 is notified of the remote procedure name, the channel number of the communication channel used for the remote procedure, and the fact that the remote procedure is to be started. In the present embodiment, a third TCP socket 233 is used as an interface in the present embodiment.
1060 (March 1990, J. Reynolds J. Pstel ISI) Internet protocol port numbers. In the client source code of FIG. 3,
The client process returns "result = printmessage
(& message, c1); ", the first breakpoint detection unit 224 determines whether the line c in the file of the remote procedure call description file storage unit 226 illustrated in FIG. "Int PRINTMESSAGE (string) = 1"
To detect that the procedure printmessage corresponds to the remote procedure PRINTMESSAGE, and
To the remote procedure printmessage is started.
The debugger interface 222 records the remote procedure activation notification from the first breakpoint detection unit 224.

Next, when the debugger instructs to execute step execution again, the debugger interface 222 instructs the first source code debug unit 206 to execute the next suspended “line 8” after the execution of the currently stopped “line 8” is completed. To set a first breakpoint at "line 9" as the line to be executed. Also, the debugger interface 222
The presence / absence of the recorded start notification of the remote procedure is checked, and if there is a record of the notification, the name of the computer having the process including the remote procedure is checked. This computer name is used, for example, when a remote procedure method is used in which the argument given to the invocation of the notified remote procedure name includes the name of a computer having a process including the remote procedure. Instead of instructing the first source code debug unit 206 to execute a step, a command for displaying the detailed contents of the argument of the procedure call is given, and the result of the display can be checked. Alternatively, at the start of the remote procedure, one computer may be determined from a plurality of computers having a process including the remote procedure by the remote procedure input destination computer determination unit 227 without specifying a specific computer name. Good. In this case, the debugger interface 222 uses the output from the remote procedure input computer determination unit 227 to
The name of the computer having the process including the remote procedure can be known.

The debugger interface 222 activates the second source code debug unit 217 on a computer having a process including the remote procedure, and activates the activated second source code via a communication channel through the second TCP socket 234. Connect to the debug unit 217. Next, the debugger interface 222 uses the process table access unit 236 to determine, from the channel number of the communication channel used for the notified remote procedure, the process of the process having the remote procedure and waiting for a connection from the client process. The number is specified, notified to the second source code debug unit 217 as the process number of the server process to be connected, and the server process execution unit 213 and the second source code debug unit 217 are connected. The debugger interface 222 notifies the second source code debug unit 217 of the recorded name of the remote procedure, and
Is set to “line C” of the source code illustrated in FIG. 4, and the third breakpoint is set to “line E”.
Set to.

Next, a second source code debug unit 21
7, the server process execution unit 213 waits for a connection from the client process execution unit 203, and the second source code debug unit 217 reaches the second breakpoint. Continue until you do. After that, the debugger interface 222 starts the first source code debugging unit 20.
6 is notified of a continue command for instructing continuous execution. The client process finishes executing the remote procedure and continues execution until the first breakpoint is reached. After that, the debugger interface 222 receives an input from the display terminal device 223 mainly to the second source code debug unit 217, and outputs a result output from the second source code debug unit 217 to the display terminal device 223. It is switched as follows.

Next, when the remote procedure is issued from the process being executed by the client process execution unit 203, the server process of the server process execution unit 213 stops at the second breakpoint. As described above, by controlling the first source code debug unit 206 and the second source code debug unit 217, the debugger can communicate with the communication procedure and the communication part of the argument when the remote procedure call is issued. Does not appear, as if the client process executed a procedure call that is not just a remote procedure on a single process. The debugger can debug a remote procedure without being aware that the server process is being debugged.

When the debugging of the remote procedure is completed and, for example, the debugger instructs the execution of the continuation, the debugger interface 222 records that the execution of the continuation is instructed, and the server process execution unit 213 sets the "line e" to Stopping at the third breakpoint, the debugger interface 222 knows it from the second source code debugger 217 through the second TCP socket 234 and then
To remove the second and third breakpoints. Second breakpoint and third
Is removed, the debugger interface 222 switches so that the command from the display terminal device 223 is sent to the first source code debug unit 206, and the first source code debug unit 2
The output is switched so that the output 06 is displayed on the display terminal device 223.

Thereafter, the debugger interface 222 instructs the second source code debugging unit 217 to end the debugging of the server process by the server process execution unit 213. As a result, the server process execution unit 213
Is disconnected from the second source code debug unit 217 and is continuously executed. The return value of the remote procedure from the server process execution unit 213 is
03, and the client process temporarily stops at the first breakpoint.
Reference numeral 06 indicates that the debugger has instructed the continuous execution in the remote procedure of the server process, so that the first breakpoint is removed and the continuous execution is instructed to the first source code debug unit 206. . Thus, the first source code debugging unit 206
And the second source code debug unit 217, the debugger does not see the communication procedure when the remote procedure call is completed and the communication part of the return value. It appears as if the return value was returned from a non-procedural procedure. Therefore, the debugger can debug the remote procedure without being aware that the server process is being debugged.
In the above-described embodiment, the debugging method of the present invention has been described by taking a remote procedure call between one client process and one server process as an example. However, in general, a connection is made in parallel with one client process. Needless to say, the debugging method of the present invention can be applied to debugging between an arbitrary client process and a server process in a distributed information processing system including a plurality of server processes. Further, the debugging method of the present invention can be similarly applied to debugging of a system including a relationship in which a server process is connected to a client process and the server process is connected to another server process as a client process.

[0017]

As described above, according to the present invention, by appropriately controlling the satellite debugging means of the server process and the client process, the person who debugs can receive the remote procedure call when the remote procedure call is issued. The communication part of the communication procedure and arguments, or the communication part when the remote procedure call ends, and the communication part of the return value are not visible, and it is as if the client process executes a procedure call on a single process that is not a remote procedure. It looks as if the return value was returned. Therefore, the debugger can use the conventional debugging method for a single process, and debug a parallel / parallel distributed program that easily calls a remote procedure without being aware of the parallel / parallel operation of the process. Can be.

[Brief description of the drawings]

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

FIG. 2 is a block diagram showing an embodiment of the present invention;

FIG. 3 shows a client process for calling a remote procedure.
Diagram showing an example of source code

FIG. 4 is a diagram showing an example of a source code of a server process.

FIG. 5 is a diagram showing an example of a remote procedure call description file.

[Explanation of symbols]

1, 2, process, 3, 4 satellite debugging means, 5 central debugging means, 51 remote procedure call issue detection means,
52: remote procedure call end detection means, 53: switch control means, 20: first satellite debug unit, 201: client source file storage unit, 202: client process execution format file storage unit, 203: client process execution unit, 204 ... Data storage unit, 206 first source code debug unit, 21 second satellite debug unit, 211 server source file storage unit, 212 server process execution format file storage unit, 213 server process execution unit, 214 ... data storage unit, 217 ... second
Source code debug unit, 218 process table 218, 22 central debug unit, 222 debugger interface, 223 display terminal device, 224
First breakpoint detecting section, 225... Second breakpoint detecting section, 226... Remote procedure call description file storage section, 227.
7, 23 communication part, 231 first TCP socket, 2
32 first remote file access unit, 233 third TCP socket, 234 second TCP socket,
235 ... second remote file access unit, 236 ...
Process table access unit 236.

Claims (1)

(57) [Claims] 1. A plurality of networks connected to each other by a network
On a computer, in a debugging method of a distributed information processing system in which a plurality of processes perform network communication mutually by remote procedure call to realize parallel / parallel operation, each of the plurality of processes is executed. On the calculator,
A plurality of satellite debugging means for debugging a process, and a central debugging means for remotely controlling the operation of the plurality of satellite debugging means. The central debugging means includes a remote procedure call to a process to be debugged by the satellite debugging means. Before the remote procedure call is issued.
A remote procedure call issued detector for output, and remote procedure call termination detection means for detecting the end of the called process in the remote procedure call, detection of a remote procedure call to the remote procedure call issued detector
Called by the remote procedure call in response to the call
Activate satellite debugging means on computer with process
And a process of the remote procedure call end detecting means.
In response to the detection of the end, Mamoru of the caller of the process
A debugging method for a distributed information processing system, comprising: control means for restarting debugging by star debugging means . [0001]