JPH11212926A - System time correction method for unix system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a correction method which can gradually and slowly change the system time without affecting the coincidence of times in a network consisting of plural UNIX systems nor the application program which work in these systems. SOLUTION: The correction time data received from a network maintenance terminal 1 and a time setting request 8 including the correction width data are received in a time management process 20, and the time of a server system 2 is gradually corrected by a degree equivalent to the correction width data. Then the system 2 periodically sends a time correction request 6 including both correction time and correction width data to the client systems 3 and 4. The systems 3 and 4 receives the request 6 in each time management process 30 respectively and gradually correct their own times by a degree equivalent to the correction width data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system time correction method, and more particularly to a system time correction method for a UNIX system which gently corrects the system time so that application programs for online transaction processing are not affected.

[0002]

2. Description of the Related Art Heretofore, a system time correction method in a general UNIX system of this kind has been known as a UNIX function ad.
The time of the system is corrected using the jtime (2) system call, and the correction range varies depending on the system.
Advance 11 seconds per second. In the case of delay, the system clock is advanced by １ ／ second per one second of real time. Some systems have a large correction width.

[0003] Further, this type of conventional system time correction method is disclosed in Japanese Patent Application Laid-Open No. 5-108198, in which a plurality of facilities for distributing the reference time are installed, so that even if the equipment being distributed fails. The time is corrected by sending the reference time, so that the time correction function is not stopped.

[0004]

However, in the time correction in a general UNIX system, during the time correction of a system having a correction range such as "advance one second to 11 seconds per actual time", 5 Even if the interruption process for a second is executed, it will be one second in real time, and the normal operation of the application program cannot be guaranteed.

Also, in the system time correction method disclosed in Japanese Patent Application Laid-Open No. 5-108198, since each device receives the time data without considering the application in operation and uses it as the standard time of the device itself, a large time correction occurs. In such a case, the running application program related to the time of each device is affected.

[0006] An object of the present invention is to solve the above-mentioned problems by synchronizing the time inside a network constituted by a plurality of UNIX systems, and by executing an application program such as an online transaction process operating on a server system or a client system. To provide a system time correction method gradually and gently so as not to affect the system time.

[0007]

SUMMARY OF THE INVENTION The present invention relates to a system time correction method for a UNIX system in a network comprising a server system and a plurality of client systems, wherein the client system receives a request from the server system. A time correction request including the correction time data and the correction width data is periodically received, and the time is gradually corrected.

Further, in response to the time correction request from the server system, a time difference between the current time of the client system and a request time based on the corrected time data is calculated, and the time difference is calculated based on the request correction width data. The time is gradually corrected using the value.

The present invention also relates to a system time correction method in a UNIX system in a network including a server system to which a terminal is connected and a plurality of client systems, wherein the server system corrects the time from the terminal. When a time setting request including time data and correction width data is received, the time is gradually corrected.

Further, in response to the time setting request, a time difference between the current time of the server system and the requested time based on the corrected time data is calculated, and the time difference is calculated using the value of the requested correction width data. It is characterized in that the time is corrected step by step.

Further, the server system periodically transmits a time correction request including correction time data and correction width data to the client system.

The present invention also relates to a system time correction method in a UNIX system in a network including a server system and a plurality of client systems, wherein the client system includes a correction time data and a correction time data from the server system. A receiving step of periodically receiving correction width data, a conversion step of converting the current time of the own system and the correction time data to a relative time, and an arithmetic step of subtracting the correction width data from the relative time, If the value subtracted in the calculation step exceeds 0, the current time is corrected by the correction width data and the calculation step is repeated at regular intervals. If the value subtracted in the calculation step is 0 or less, An end step of ending the correction of the current time. It is characterized in.

[0013] The present invention also relates to a system time correction method in a UNIX system in a network comprising a server system connected to a terminal and a plurality of client systems, wherein the server system corrects the time from the terminal. A receiving step of receiving time data and correction width data, a conversion step of converting the relative time based on the current time of the own system and the corrected time data if the corrected time data is an absolute time, An arithmetic step of subtracting the correction width data, a repeating step of correcting the current time by the correction width data if the value subtracted in the arithmetic step exceeds 0, and repeating the arithmetic step at regular intervals; and If the subtracted value is 0 or less, the correction of the current time is terminated. It is characterized in that it comprises a termination step that.

Further, the server system is characterized by including a transmission step of periodically transmitting the correction time data and the correction width data to the client system.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the operation will be described below.

In a network system composed of a plurality of UNIX systems, it is possible to automatically correct the time of all subordinate client systems only by correcting the time of the server system. In addition, when a client system is added / removed in the network, it is possible to cope with the problem only by modifying the contents of the client system registration table by providing the client system registration table in the server system.

In each of the server system and the client system, the system clock is corrected little by little, so that the system time can be gently corrected so that application programs for online transaction processing are not affected.

Next, embodiments of the present invention will be described with reference to the drawings.

Referring to FIG. 1, a network system comprising a plurality of UNIX systems according to an embodiment of the present invention includes a network maintenance terminal 1 for receiving a time setting request from a network administrator, and a reference within the network. The server system 2 includes a server system 2 serving as a time, client systems 3 and 4 subordinate to the server system 2, and a LAN 5 connecting the server system 2 and the client systems 3 and 4.

The server system 2 has a time management process 20 including a function of correcting a UNIX system clock by an arbitrary correction width under program control, and a client system registration table 7. Each of the client systems 3 and 4 has a time management process 30 for performing program control.

Referring also to FIG. 1, a plurality of UNIX
It shows the flow of time management processes 20 and 30 and related data within a network system which is constituted by a system. For example, when a network administrator makes a time setting request using the network maintenance terminal 1, the setting request reaches a time management process 20 inside the server system 2 as a time setting request 8, and In the time management process 20, the time of the own system is corrected little by little according to the setting request. Further, the server system 2
Obtains the current time of the server system at a predetermined period of time, and sends a time correction request 6 to perform time correction on the subordinate client systems. Similarly, the client system gently and gradually corrects the time of its own system according to the time correction request 6.

FIG. 2 is a block diagram mainly showing the details of the inside of the server system of FIG.

The server system 2 includes a time management process 2
0, client system registration table 7, and UNI
And an X system clock 26. The time management process 20 includes a request analysis 21 for analyzing a time setting request, a correction process 22 for comparing the current time of the system with the requested time and gradually correcting the system time with a specified correction width, and a current time of the system. And a synchronization process 23 for periodically obtaining a time correction request to the subordinate client system. In response to the time setting request 8 received from the network maintenance terminal 1 (this request includes “correction time data” and “correction width data”), the request analysis 21 of the time management process 20 is started, and the validity of the data is confirmed. I do. "Correction time data" and "Correction width data" whose validity has been confirmed
Is carried over to the correction processing 22, compares the “correction time data” with the current time of the system (UNIX system clock 26), and gradually changes the UNIX by the correction width of the “correction width data”.
The time correction of the system clock 26 is performed.

In the periodic process 23, the current time (UNIX system clock 26) of the system is obtained separately from the time correction process of the own system, and the subordinate client systems managed by the client system registration table 7 are obtained. Issue a time correction request 6 for time correction.

FIG. 3 is a block diagram mainly showing the inside of the client system 3 or 4 of FIG. 1 in detail.

The client system 3 or 4 comprises a time management process 30 and a UNIX system clock 36.

The time management process 30 compares a request analysis 31 for analyzing the time setting request 8 from the server system 2 with the current time of the system and the requested time, and gradually corrects the system time by a specified correction width. And a correction process 32. The validity of the time correction request 6 (this request includes “time correction data” and “correction width data”) received from the periodic processing 23 inside the server system 2 is confirmed by the request analysis 31 in the same manner as the server system. I do. "Correction time data" and "Correction width data" whose validity has been confirmed
Is carried over to the correction processing 32 as in the server system, and the “correction time data” and the current time of the system (UNI
The X system clock 36) is compared, and the time correction process of the UNIX system clock 36 is performed little by little with the correction width of the “correction width data”.

In FIG. 2, the data of the time setting request 8 received from the network maintenance terminal 1 are “correction time data” and “correction width data”. Note that three types of “correction time data” can be specified, and that the time to be set is specified by “year, month, day, hour, minute, second, and microsecond”, and the time to be set is set to “UNIX total seconds + It is possible to specify a Julian time specified by "microsecond" and a relative time specified by "± second + microsecond" to set a time to be set based on the current time. The "correction width data" can be specified by an arbitrary value within a range from a correction width of a minimum of 10 milliseconds per second of real time to a maximum time correction width permitted by the UNIX system.

For example, referring to FIG. 4A, assuming that the system clock is advanced by a correction width of 100 milliseconds per second of real time, when the real time is 1.0 second, the system clock becomes 1.1. Second, the system clock is 5.5 seconds when the real time is 5.0 seconds, and the system clock is 1 when the real time is 10.0 seconds.
This is corrected to 1.0, which means that the system clock is advanced by 1 second to the actual time of 10 seconds. Referring to FIG. 4B, assuming that the system clock is delayed by the same correction width, the system clock is 0.9 seconds when the real time is 1.0 second, and the system clock is 5.0 seconds when the real time is 5.0 seconds. Is 4.5 seconds, real time 10.
At 0 seconds, the system clock is corrected to 9.0 seconds, when the real time is 5.0 seconds, the system clock is corrected to 4.5 seconds, and when the real time is 10.0 seconds, the system clock is corrected to 9.0 seconds. That is, the system clock is delayed by one second per second.

Next, the operation of the embodiment of the present invention will be described with reference to FIGS.

Now, in order to correct the time of the server system 2, the maintainer inputs “corrected time data” as the time to be adjusted and “corrected time width data” as the time width to be corrected from the network maintenance terminal 1. And Then, the time setting request 8 (“correction time data” and “correction width data”
A) is sent to the server system 2. The “correction time data” and “correction width data” from the network maintenance terminal 1 are received by the request analysis 21 of the time management process 20 of the server system 2 in FIG. 2, and the absolute time specification or the Julian time specification data is converted to the relative time data. (Fig. 5
Steps S1, S2, S3). In the calculation method of the relative time at this time, if the absolute time is, for example, the absolute value of (current time−absolute time) is the relative time. However, the current time is a value obtained by reading the UNIX system clock 26. The reason for converting to the relative time is to unify the internal processing of the program of the correction processing 22.

In the correction process 22, the "correction time data" (hereinafter, simply referred to as "relative time") and the "correction width data" (hereinafter, simply referred to as "correction width") converted into relative time data are stored in the memory. Then (steps S4 and S5), the correction process is started. In the correction processing, the designated "correction width" is set as an argument of the UNIX function adjtime (2), which is a UNIX system call (step S7), and this system call is executed. That is, in the correction process at this time, a subtraction process is performed when the time is advanced, and an addition process is performed when the time is delayed. Thereafter, the “correction width” is subtracted from the “relative time” (step S8),
The process is suspended for one second (step S9). The end of this processing is to compare the “relative time” with the “correction width” and determine that the value is 0.
It is determined based on the following (step S6). That is, until the comparison value becomes 0 or less, for example, if 200 ms is designated as the “correction width”, the UNIX system clock 26 is gently corrected by 200 ms per second.

Next, it is assumed that "correction width data" is set in advance so that the time can be corrected at a fixed period (for example, on a daily basis). At this time, the correction time (for example, 8:
00), a time correction request 6 is issued every day.

The time correction method between the server system 2 and the client systems 3 and 4 based on the time correction request in FIG.
The periodic processing 23 of the server system 2 executes the time of its own system (UNIX system clock 2) at a predetermined cycle.
6) is obtained, and “year / month / day / hour / minute / second + microsecond” data, which is a type of absolute time specification, and standard correction width data specified in advance are transmitted to the subordinate client systems 3 and 4 as time correction data. A time correction request 6 is issued to make the request. Subordinate client systems 3, 4
“Data” is data managed by a “periodic processing program” in the server system and needs to be registered in the client system registration table 7 in advance. The time correction request 6 from the server system 2 is received by the request analysis 31 of the time management process 30, and the UNIX system clock 36 is gradually increased by the correction process 32, similarly to the processing of the server system 2.
Is corrected. FIG. 6 shows the correction processing flow of the time management process 30 at this time.

In order for the server system 2 to periodically correct the time to the client systems 3 and 4, a time correction request 6
(With "correction time data" and "correction width data") are sent to the client systems 3 and 4. “Correction time data” and “correction width data” from the server system 2 are received by the request analysis 31 of the time management process 30 of the client system in FIG. 3, and the absolute time designation data is converted into relative time data (FIG. 6). Step S22). In the calculation method of the relative time at this time, the absolute value of (current time−absolute time) becomes the relative time similarly to the processing of the server system 2. However, the current time is a value obtained by reading the UNIX system clock 36.

In the correction process 32, the "correction time data" (hereinafter abbreviated as "relative time") and the "correction width data" (hereinafter abbreviated as "correction width") converted into relative time data are stored in the memory. (Steps S24 and S25), and the correction process is started. The correction process is performed by adding an argument of a UNIX function adjtime (2), which is a UNIX system call,
The designated "correction width" is set (step S27), and this system call is executed. That is, in the correction process at this time, a subtraction process is performed when the time is advanced, and an addition process is performed when the time is delayed. afterwards,
Subtract "correction width" from "relative time" (step S
28) Execute processing interruption for one second (step S2)
9). The end of this process is determined by comparing the “relative time” with the “correction width” and determining that the value is 0 or less (step S26). That is, until the comparison value becomes 0 or less, for example, if 200 ms is specified for the “correction width”, 2
Correct the UNIX system clock gently every 00 milliseconds.

[0037]

The effect of the present invention is as follows. In a network constituted by a plurality of UNIX systems without using special heartware, a network administrator simply corrects the time of a server system, and a subordinate client system can be used. This means that all times can be automatically corrected. In addition, the network administrator can cope with the work of adding / removing the client system inside the network only by modifying the client system registration table.

Further, a server system or a client system on which an application program for online transaction processing of the present network operates. In this case, by designating "correction data" in a range that is not affected by the present program, it becomes gentler little by little. It is possible to correct the system time.

[Brief description of the drawings]

FIG. 1 is a network diagram composed of a plurality of UNIX systems.

FIG. 2 is a configuration diagram of a server in a system according to the present invention.

FIG. 3 is a configuration diagram of the present invention in a client system.

FIG. 4 is a relationship diagram between a system clock and an actual clock in a system time correction process.

FIG. 5 is a flowchart of a correction process of a time management process of the server system.

FIG. 6 is a flowchart of a correction process of a time management process of the client system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Network maintenance terminal 2 Server system 3, 4 Client system 5 LAN 6 Time correction request 7 Client system registration table 8 Time setting request 21, 31 Request analysis 22, 32 Correction processing 23 Periodic processing 26, 36 UNIX system clock

Claims (8)

[Claims] 1. A system time correction method in a UNIX system in a network including a server system and a plurality of client systems, wherein the client system includes a correction time data and a correction width data from the server system. U which periodically accepts a time correction request including
A system time correction method in a NIX system. 2. In response to the time correction request from the server system, a time difference between a current time of the client system and a requested time based on the corrected time data is calculated, and the time difference is calculated as a value of the correction width data. 2. The UNI according to claim 1, wherein the time is corrected little by little by using.
System time correction method in X system. 3. A system time correction method in a UNIX system in a network including a server system connected to a terminal and a plurality of client systems, wherein the server system corrects the correction time data from the terminal and the correction time data from the terminal. The time is gradually corrected when a time setting request including width data is received.
System time correction method in the system. 4. In response to the time setting request, a time difference between a current time of the server system and a request time based on the corrected time data is calculated, and the time difference is gradually changed using the value of the correction width data. 4. The system time correction method in a UNIX system according to claim 3, wherein the time is corrected. 5. The server system according to claim 1, wherein the server system periodically transmits a time correction request including correction time data and correction width data to the client system.
Or the system time correction method in the UNIX system according to 3. 6. A system time correction method in a UNIX system in a network including a server system and a plurality of client systems, wherein the client system includes a correction time data and a correction width data from the server system. A receiving step of periodically receiving a time correction request including: a conversion step of converting the current time of the own system and the corrected time data into a relative time; and an arithmetic step of subtracting the correction width data from the relative time. If the value subtracted in the calculation step exceeds 0, the current time is corrected by the correction width data and the calculation step is repeated at regular intervals; and the value subtracted in the calculation step is 0 or less. End step for ending the correction of the current time, if any. And a system time correction method in a UNIX system. 7. A system time correction method in a UNIX system in a network including a server system to which a terminal is connected and a plurality of client systems, wherein the server system corrects the correction time data and the correction time data from the terminal. A receiving step of receiving width data; a conversion step of converting the corrected time data into a relative time based on the current time of the own system and the corrected time data if the corrected time data is an absolute time; and An operation step of subtracting, if the value subtracted in the operation step exceeds 0, a repetition step of correcting the current time by the correction width data and repeating the operation step at regular intervals, and a value subtracted in the operation step Ending step of ending the correction of the current time if is less than or equal to 0 And UNI.
System time correction method in X system. 8. The UNIX system according to claim 6, wherein said server system includes a transmission step of periodically transmitting said correction time data and said correction width data to said client system. System time correction method.