JPH04242467A - Combined computer system - Google Patents

Abstract

PURPOSE:To minimize the service stop time and to quickly restart the service by setting previously a stand-by system in a start state while a working system is kept active and then taking over the operation automatically by the stand-by system when the working system is stopped. CONSTITUTION:A working host computer 100 offers the on-line service. A stand-by host computer 200 consists of an inter-system communication means 201, a communication control means 202, and an on-line service system 210. The inter-system communication equipments 103 and 203 perform the exchange of information between them. Meanwhile a fault detecting device 204 monitors the system fault of the computer 100 end informs the system 210 of a fact that the working of the system 100 has stopped by a fault. In a composite computer system, a system 110 is started when the system 210 is ready. Then the system 210 transmits the confirming information of the start declaration of a working system through the means 201.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention is for quickly restarting services when an operating computer system (hereinafter simply referred to as the operating system) is stopped due to a sudden hardware/software failure. This invention relates to a complex computer system that minimizes service stoppage time by activating a standby system in advance when the active system is running and automatically taking over operations when the active system is stopped. be.

0002

[Background Art] In a complex computer system, when the operating system starts service, a standby system is activated and ready to execute online operations, so that it can be used in the event of a system failure of the operating system. Continuous online service operation is achieved by switching systems. In conventional complex computer systems, when switching systems, the shared communication control unit is notified that the old operating system has stopped, or the system operation monitoring function of the shared communication control unit detects the stoppage of the operating system.
The shared communication control device disconnects all sessions that are open at that time and creates the same state as the initial state. The new operating system performs a session opening operation for the defined terminal, and the terminals under the shared communication control device are reconnected. Therefore, for the terminal user, it appears as if the service has been temporarily stopped and then reconnected.

0003

Problems to be Solved by the Invention In system switching when an operating system fails, the above-mentioned terminal reconnection processing becomes longer as the number of terminals increases. Furthermore, disconnecting and reconnecting a terminal session takes a very long time because a control message based on the connection procedure is sent over the online network. For this reason, the time required for the entire system switching process increases, and for terminal users, as the scale of the system increases, service outage time becomes extremely long, which poses a problem for actual business operations.

The present invention has been made to solve these problems, and its purpose is to provide a complex computer system that can quickly switch systems regardless of the number of terminals. It is in.

[0005]

[Means for Solving the Problem] The standby system also opens a standby session in advance for the terminal with which the active system has opened a session. The standby session at this time is a virtual one between the shared communication control unit and the online service system on the standby system, and when this session exists, the connection between the terminal and the shared communication control unit during system switchover is Do not disconnect sessions. Therefore, when a new operating system is activated, online services can be restarted immediately without reconnecting sessions.

[0006]

[Operation] (1) Start up the standby system in advance, put it in a state where online services can be executed, and then start the service of the active system.

[0007] The active system and the standby system open sessions to the same terminal at the same time. At this time, the session from the standby system is a virtual session only between the online service system and the shared communication control device.

[0008] If a virtual standby session exists while a session with an actual terminal exists in the active system, the shared communication control unit is configured to stop the old active system for the shared communication control unit during system switchover. In notification or shared communication control equipment operation system outage monitoring, the session with the terminal is not disconnected, and the online system of the new operation system is not reconnected to the already established session, depending on the number of terminals. A quick system switchover is realized. (2)/(3) The activation of the active system is notified to the standby system that has been activated in advance by the intersystem communication means and the intersystem communication device. On the other hand, it is extremely burdensome to transmit the session establishment event to the standby system using intersystem communication means. This is because if sessions are opened all at once in active systems, there is a risk that the capacity of the intersystem communication means will be exceeded. Furthermore, it becomes difficult to avoid a failure in the intersystem communication means.

[0009] The session connection event of an active session is transmitted by being recorded in a state inheritance file each time a session is opened. Each time a session is opened, the name of the opened terminal is recorded in the file so that it is possible to see how much has been recorded.

[0010] In the standby system, the contents of the status inheritance file are checked periodically from the time when the activation of the active system is sensed, and it is determined from the valid position of the latest information whether there has been a change in the session status in the active system. Even when sessions are opened all at once in an operating system, by recording and reading files, the session status can be reliably inherited until the end of the file is reached. (4)/(5) An active session is opened using session attributes based on the network definition of the active system. The same applies to the standby session, but the session attributes actually used for conversation control between the terminal and the online service system are, needless to say, the attributes in the active session. That is, if the session attributes of the active session and the session attributes of the standby session do not match, an inconsistency will occur in conversation control between the terminal and the online service system during system switching.

In order to avoid this, the session attributes of the active session and the session attributes of the standby session are compared, and if they are different, the standby session is not opened. As a result, the session for that terminal is disconnected and reconnected in the same manner as in the past at the time of system switching. (6) In a system having a shared communication control device in which a standby session may be opened regardless of whether there is an active session, it is possible to keep a standby session open all the time. By doing this, it is possible to record whether or not the session opening information has been recorded in the session state inheritance file, and once the session opening event is recorded, from then on, by not recording the session opening event in the session state inheritance file. To realize reliable opening of a standby session by preventing the file from reaching the end even in a system where an active session repeatedly connects and disconnects to a finite session state inheritance file. (7) When a failure occurs in the shared communication control device, the standby session is also disconnected. Therefore, when reconnecting after an active session is disconnected due to a failure in the shared communication control device, even after the session opening event has been recorded once, the session opening event must be recorded in the active system.

[0012] In this case, the handover information is not limited. The inheritance file realizes permanent session state inheritance by reusing the file from the beginning when the end of the file is reached. To avoid overwriting inherited data due to reuse before it is read by the standby system in a finite inheritance file, it is possible to avoid overwriting the inherited data due to reuse before the standby system reads it. , Reliable session state inheritance is achieved by reconnecting after failure recovery.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of the present invention. In the figure, 100 is an operating host computer that provides online services, and includes intersystem communication means 101, communication control means 102, startup control means 111, which is a feature of the present invention, and session establishment state recording means 112. Session opening means 113, session reconnection means 114, terminal name 116, session attribute 117, session takeover history 118, failure record 119,
Online service system 11 including a terminal status table 115 including an information recording area of session takeover time 120
It consists of 0.

Further, 200 is a standby host computer, which includes intersystem communication means 201, communication control means 202, and switching control means 211, which is a feature of the present invention.
session opening means 213 including session opening state reading means 212 and standby session checking means 214;
A terminal status table 21 including a takeover file monitoring means 215 and an information storage area for terminal name 217 and session attribute 218
6, and an online service system 210 including 6 and 6.

Further, 103 and 203 are intersystem communication devices for exchanging information between both systems, and 204 is a system monitoring device for monitoring system failures in the operating host computer 100, and when the system is stopped due to a system failure, the online service system 20
This is a failure detection device that sends a notification to 0. Also, 400 is the terminal name 411, the session attribute 412, and the record generation 41.
a session state takeover file consisting of session state records 410 for more than the number of terminals, including a storage information area of 3;
Reference numeral 300 denotes an active/standby shared communication control device (hereinafter referred to as
800 is a terminal device.

For convenience of explanation, sessions from the online service system 110 are referred to as operating sessions 500,
A session from the online service system 210 is represented by a standby session 600, and a session with the actual terminal device 800 is represented by a terminal session 700.

Next, the operation of this embodiment will be explained with reference to the drawings. (System Startup) In the composite computer system, before starting up the online service system 110, the online service system 210 is started up. When the online service system 210 is ready for online service, the online service system 11
Start 0.

The startup control means 111 first prepares the services of its own system. In addition, the communication control means 102
Execute the running system declaration for. by this,
The communication control means 102 controls the opening of a session for the terminal device 800 requested by the online service system 110 so that the terminal session 700 is actually opened as the active session 500. When the startup control means 111 completes the preparation, it uses the intersystem communication means 101 to transmit an active system startup declaration to the online service system 210.

This request is sent to the intersystem communication devices 103 and 203.
The information is transmitted to the intersystem communication means 201 via. The inter-system communication means 201 communicates this with the online service system 21.
0, the takeover file monitoring means 215 activates a timer to check the session opening status reading means 212.
The session state inheritance file 400 is periodically read using the . Thereby, the online service system 210 can acquire the latest session status record 410 information that is recorded when an event change such as session opening occurs in the online service system 110.

The online service system 210 uses the intersystem communication means 201 to send a confirmation notification of the activation system activation declaration. This notification is delivered to the online service system 110 by the intersystem communication means 101. When the activation control means 111 receives this confirmation notification, the operating system starts the service. Intersystem communication device 103
, 203 may be faulty. In such a case, the startup declaration and confirmation notification will be lost. These retransmission controls are performed by the intersystem communication means 101.
This is also performed by the activation control means 111. (Opening a session) Online service system 110
opens a session for the terminal device 800 to be serviced when the online service system 110 is activated or when a session opening command is input. Session opening method 11
3 searches the terminal status table 115 corresponding to the target terminal device 800 and opens a session by issuing a session opening instruction to the communication control means 102. Since the communication control means 102 has received the active system declaration from the activation control means 111, it requests the shared communication control device 300 to open an active session 500. At this time, if the failure record 119 is recorded as after failure, the session takeover time 120 is referred to and the current time is determined by the takeover file monitoring means 215.
Determine whether enough time has elapsed than the period monitored by . If sufficient time has not elapsed, the session reconnection means 114 requests reconnection of the session after the elapse of time. The session reconnecting means 114 waits until a sufficient time has elapsed for the terminal, issues a session opening request to the session opening means 113, and re-executes the above process. If sufficient time has elapsed, the failure record 119 is cleared and a session is opened.

The session opening means 113 acquires the session attribute on the network definition of the active host computer 100 when the communication control means 102 accepts the session opening, and records it in the session attribute 117. Furthermore, the session takeover history 118 is checked, and if the contents are cleared, a session establishment recording request is issued to the session establishment status recording means 112 by specifying the terminal status table 115. If the session takeover history 118 indicates that the session has been taken over, nothing is done. Upon receiving the session establishment recording request, the session establishment state recording means 112 records the terminal name 116 of the specified terminal status table 115 and the session attribute 1.
17, the terminal name 411 and session attribute 412 are set, and the session state record 410 is written next to the last written record in the session state inheritance file 400. In addition, the current time is written to the record generation 413 and session takeover time 120, and the current time is written to the record generation 413 and session takeover time 120, and
The session has been taken over. If the last record written is the last record in the file, it is written at the beginning in the same way. (Opening a standby session) Online service system 2
10, the session opening state reading means 212 at the cycle started by the takeover file monitoring means 215 at the time of startup.
The session state inheritance file 400 is read by. Records whose record generation 413 of the read records is new are recognized as information to be inherited, and are read one after another. If the record generation 413 goes back in time, the record is still in the online service system 1.
10 is considered to have not been recorded. In addition,
The determination time of the record generation 413 at the first reading when the online service system 210 is started is included in the startup declaration notification when the online service system 110 is started, so this is referred to.

The takeover file monitoring means 215 searches the corresponding terminal status table 216 from the terminal name 411 of the read session status record 410 and determines the session attribute 4.
12 is recorded in the session attribute 218, and a session opening instruction is issued to the session opening means 213. When the session opening means 213 issues a session opening instruction to the communication control means 202, the communication control means 202 requests the shared communication control device 300 to open a standby session 600. The session opening means 213 can acquire the session attribute on the online definition of the standby host computer 200 when the communication control means 202 accepts the session opening. However, since this session attribute does not necessarily match the one that controls the actual terminal session 700, the standby session checking means 214 performs an attribute check. The standby session check means 214 compares the acquired session attribute with the content recorded in the session attribute 218 when reading the handover information. If they are not equal, the establishment of the standby session 600 is rejected. (Operation at the time of session failure) When the shared communication control device 300 becomes a failure, a failure notification of the communication control means 102 is notified to the online service system 110. When the session reconnection means 114 receives the failure notification, it searches the terminal status table 115 for the relevant session, clears the session takeover history 118, and records ``after failure'' in the failure record 119. If the terminal is a target for automatic recovery in the event of a failure, the session handover time 120 is referred to to determine whether the current time has passed sufficiently longer than the period monitored by the takeover file monitoring means 215. If sufficient time has elapsed, a session opening request is issued to the session opening means 113, and the session opening process described above is attempted. If this is not the case, the terminal waits for the processing for that time, and then re-executes the processing. When the standby session 600 is opened, a failure notification of the communication control means 202 is notified to the online service system 210, but the online service system 210 assumes that the session is simply disconnected. (Role of standby session) When the activation control means 111 declares the operating system to the communication control means 102, when the online service system 110 issues a session opening request for the terminal device 800 to the communication control means 102, the communication control means 102 opens an active session 500 to the shared communication control device 300. The shared communication control device 300 establishes a terminal session 700 for the terminal device 800.
, and the online service system 110 is able to send and receive messages to and from the terminal device 800.

On the other hand, in the communication control means 202 which has not received the declaration of an operating system, the online service system 2
10 issues a session opening request to the terminal device 800 to the communication control means 202, the communication control means 202
is a standby session 600 for the shared communication control device 300.
will be established. When the shared communication control device 300 receives this, the standby session opening means 301 maintains control of the session. The standby session opening means 301 controls the standby session 600 regardless of whether the active session 500 is opened or not.

The shared communication control device 300 monitors the operating host computer 100, and when a system stoppage is detected, or when the fault detection device 204 detects a fault stoppage of the operating host computer 100, the shared communication control device 300 monitors the operating host computer 100. When receiving the stoppage notification of the active host computer 100, the standby session opening unit 301 holds the session in the session holding unit 302 only for the terminal device 800 for which the active session 500 and the standby session 600 have been established. 600 standby sessions with no active sessions 500
causes the communication control means 202 to disconnect the standby session 600 due to failure, and the active session 500 without the standby session 600 causes the terminal device 800 to disconnect the terminal session 700.
Disconnect the faults respectively.

The control means 2 switches the detection of a failure stop of the operating host computer 100 by the failure detection device 204.
11, after notifying the above system stop detection,
The switching control means 211 executes a takeover process of the online transaction service that was being executed by the online service system 210. When this handover process is completed, the switching control means 211 notifies the communication control means 202 of an active system declaration in order to resume the service of the online service system 210 as an active system. When the communication control means 202 receives the operating system declaration, it takes over the communication state of the terminal session 700 with the shared communication control device 300. The session holding unit 302 regards the standby session 600 as an active session, and thereafter allows the standby host computer 200 to perform transmission and reception via the terminal session 700.

[0026]

Effects of the Invention As explained above, in the present invention, when a system switchover occurs when an operating system fails in a complex computer system, the sessions of the terminals that were being serviced can be continued without disconnecting and reconnecting, so the number of terminals can be reduced. Quick system switching without dependency can be realized. Furthermore, no matter how many times a terminal is connected or disconnected, it will not be recorded in the session state inheritance file unless the shared communication control device becomes a failure. Therefore, even if the session state inheritance file is limited, as long as there are records for the number of terminals, the terminal state can be carried over reliably. Furthermore, in the event of a failure of the shared communication control device, the standby session will also be disconnected, so it is necessary to record the session takeover.
The re-opening of a session after failure recovery is written after waiting at least the monitoring cycle time of the standby system, so that the terminal state can be reliably inherited.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the present invention.

[Explanation of symbols]

100 Operating host computer 101
Intersystem communication means 102 Communication control means 103 Intersystem communication device 110 Online service system 111
Startup control means 112 Session opening state recording means 113
Session opening means 114 Session reconnection means 115 Terminal status table 200 Standby host computer 201
Intersystem communication means 202 Communication control means 203 Intersystem communication device 204 Failure detection device 210 Online service system 211
Switching control means 212 Session opening state reading means 213
Session opening means 214 Standby session checking means 215
Handover file monitoring means 216 Terminal status table 300 Active/standby shared communication control device 301
Standby session opening means 302 Session holding means 400 Session state inheritance file 800
terminal device

Claims (7)

[Claims] Claim 1: A standby system having the same device configuration and software configuration as an operating computer system is placed on standby in advance in a state where online services can be immediately executed, and when the operating computer system is providing online services, In a composite computer system configured to switch to operation in the standby system when a failure occurs in the computer system, a standby session opening means for simultaneously opening a connection to the same terminal from the operating computer system and the standby system, and the operating computer an active/standby shared communication control device having a session holding means for restarting the session when the standby system starts operation without disconnecting the session in the event of a system failure; A standby session is created in the shared communication control device by opening a session for the same terminal, and the standby session is maintained by notifying the shared communication control device that system switching has occurred in the standby system when the operating computer system fails. A complex computer system characterized by realizing high-speed system switching without disconnecting and reconnecting sessions in the event of a failure by restarting sessions that were previously in use. [Claim 2] The operating computer system according to claim 1 includes a startup control means for controlling when the system is started, a session opening means for opening a session with a terminal, and recording information about the session opening in a session state inheritance file. 2. The composite computer system according to claim 1, further comprising: an online service system including session opening state recording means for recording session opening status. 3. The standby system according to claim 1 includes a switching control means for controlling system switching based on detection information of a failure detection means for detecting a failure stop of an operating computer, and a session establishment for opening a session with a terminal. 2. An online service system according to claim 1, comprising: a means for reading a session establishment state, a means for reading a session establishment status for reading the contents of a session status inheritance file, and an inheritance file monitoring means for monitoring session establishment of an operating computer system. complex computer system. 4. The session state inheritance file and session establishment state recording means of claim 2 includes a session state inheritance file including a session attribute field capable of recording session attributes of an active session, and a session state inheritance file that records session attributes of an active session. 3. The composite computer system according to claim 2, further comprising session establishment state recording means including session attribute recording means. [Claim 5] The session opening state reading means and the session opening means of claim 3 include a session opening state reading means including a means for acquiring session attributes of the active session, and an active computer system that takes over the session attributes of the standby session when the session is opened. 4. The composite computer system according to claim 3, further comprising a session opening means including a standby session check means for comparing the attributes of the active session with the attributes of the active session and checking not to open a standby session when the two are different. . [Claim 6] The shared communication control device according to claim 1 establishes a standby section regardless of the presence or absence of an active section, and stores a session inheritance history indicating whether information on session establishments has been recorded in a session state inheritance file. 2. The composite computer system according to claim 1, wherein the information is recorded in a terminal status table to realize status inheritance that is independent of the amount of change in the session establishment status. 7. The online service system according to claim 2 includes a terminal state table that records fault record information of the shared communication control device and handover time information that records the recording time of the first session connection event, and 3. The composite computer system according to claim 2, further comprising session reconnection means for causing the connection to be established after a predetermined period of time has elapsed from the recording time of the first session connection event.