JPH0448256B2 - - Google Patents

Description

[Detailed Description of the Invention] [Summary] A data processing system that exchanges messages to be processed between application programs (referred to as message communication) temporarily suspends data processing due to a system failure, and then removes the cause of the failure. In a message recovery processing method for restoring data processing, when a message is communicated, a copy is stored in the file memory DASD in the form of a message queue, and at the start of system restart processing after a system down. , the file memory
When restoring unprocessed messages from DASD onto the main memory buffer, the number of messages on the queue waiting to be processed on the main memory buffer is monitored, and when the number of messages waiting to be processed falls below a certain number, The next message is restored from the file memory DASD to the main memory buffer and connected to the queue waiting for processing.

The message recovery processing described above is performed in parallel with the processing of the application program by a dedicated task separate from the application program.

As a result, during system recovery processing after a system failure, all unprocessed messages (held on DASD) are expanded to the main memory, and the main memory buffer area becomes larger than necessary for the messages. The processing balance of the data communication management system DCMS is adjusted so that the processing efficiency of other data processing (conversation processing performed between terminals and application programs via the line, etc.) that is occupied by the data communication management system and performed independently of the recovery processing is not reduced. The aim is to improve

[Industrial application field]

The present invention relates to a message recovery processing method when a failure occurs in a data processing system that communicates messages between application programs X and Y.

In recent years, in various industries, as business operations have expanded, the amount and types of data to be processed in real time have increased, and there has been a trend toward greater reliance on online data processing systems.

Therefore, from the viewpoint of processing capacity, load distribution and function distribution are attempted among multiple systems, and from the standpoint of risk distribution, so-called distributed systems are increasingly being adopted, in which file memory is divided into multiple systems. Ta.

Furthermore, in online systems, data (messages) input from a terminal are generally processed in the main memory and then filed in the form of a database or the like.

In such a computer system that is a complex system, for example, when data is input from a terminal connected to one system, if the data belongs to a file in the own system, it is immediately transferred to that file. , but it does not belong to the file of the own system, or even if it belongs to the file of the own system, if the application program X that received the data from the terminal cannot directly process the data. , it is necessary to send the data to another application program Y. This data transmission between application programs X and Y is what is called communication between application programs.

At this time, if the other party's system to which the data is to be sent has a failure, the data received from the terminal cannot be sent to the other party's system. When this happens, a method is adopted in which the data is sent to the other party's system.

Also, in communication between two application programs within the own system, copies of data (messages) are stored in the file memory in order to avoid loss of data (messages) when the system goes down. The data is expanded onto the main memory and processed.

In this type of processing, when a large amount of data (messages) is accumulated in the file memory, if the data is expanded to the main memory at once, the main memory will be occupied by the data, and it will be difficult to restore the data. There is a problem in that it becomes impossible to allocate main storage space to application programs that perform data processing other than processing, and as a result, the processing efficiency of the entire system decreases, and an effective data (message) recovery processing method is awaited. Ta.

[Prior art] Figure 3 schematically shows the concept of communication between general application programs.
When 1 sends a message using a communication macro for sending a message (data), for example, "PUT," the message is stored in the main memory 3 as configured by the user according to the scale of the system. The message is written to the main memory buffer 31, and is also notified to the application program Y2 which is issuing the communication macro 'GET' for receiving the message.

In the application program Y2, by reading the message from the main memory buffer 31 and processing it, the message in the main memory buffer 31 becomes "open", that is, "empty".

In this way, conventionally, for example, two application programs X and Y have executed communication between them by issuing the communication macros "PUT" and "GET" to each other.

Next, the improved communication system between application programs X and Y will be explained with reference to FIG.

In this method, the application program
At the same time as the message is stored on the main memory buffer 31, a copy of the message is kept on the file memory DASD 4, so that the message is not lost from the main memory buffer 31 due to a system failure or the like. I made sure to get a guarantee when the time came.

[Problem that the invention seeks to solve]

Therefore, in the conventional system shown in (2), there was a problem in that if the system went down while communication between application programs was being executed, data (messages) in the main memory buffer 31 would be lost.

In addition, in the improved conventional method, when the system goes down, unprocessed messages are stored in the file memory DASD 4, so by expanding the messages to the main memory buffer 31, the application programs However, when deploying the message to the main memory buffer 31, conventionally all unprocessed messages stored on the file memory DASD 4 are transferred to the main memory buffer 31. Data communication management system DCMS
There was a problem in that it had a significant impact on the operation of the terminal (for example, conversation processing between the terminal and the application program).

In view of the above conventional drawbacks, the present invention dynamically manages the number of messages recovered for each application program in the message recovery process when a system down occurs during communication between the application programs X and Y. The purpose of this invention is to provide a method that can efficiently perform message recovery processing without affecting the operation of the data communication management system DCMS.

[Means for solving problems]

FIG. 1 is a block diagram of the principle of message recovery processing according to the present invention.

In this figure, a1 and a2 are application program Y2
At the message queue terminal 31' waiting for processing,
When the system goes down and the application program Y2 is started, the message queues (Ma1-1, Ma1-2,...,
Ma2-1, Ma2-2, . . . ), the messages are restored to the main memory buffer 31 for each terminal 31'.

At this time, a dedicated recovery task provided in the data communication management system DCMS performs a recovery task based on the request from the application program Y2.
In parallel with the message processing in step 2, the number of unprocessed messages remaining on the main memory buffer 31 is monitored for each terminal 31' of the message queue waiting to be processed, and the number of unprocessed messages is monitored. When the number becomes less than a certain number, for example, the number of application program tasks x 2, the next message on the file memory DASD 4 is restored to the main memory buffer 31 and connected to the message queue waiting to be processed.

[Effect]

That is, according to the present invention, application programs X, Y
If a system failure occurs during communication between the The number of unprocessed messages in the queue is monitored, and when the number of unprocessed messages falls below a certain number, the next message on the file memory DASD 4 is restored onto the main memory buffer 31 and processed. Connect to the waiting message queue.

At this time, the main memory buffer 31 is designed so that the restoration process can be performed by a dedicated task provided in the data communication management system DCMS in parallel with the message processing in the application program Y2. Since the restored message is not occupied and the message restoration processing is performed in a dedicated task separate from the application program Y2, the application program Y2 can concentrate on processing the message, improving the processing efficiency of the entire system. There is an effect that can be done.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.
Figures 2A and 2B schematically show an embodiment of the present invention, and of the continuous processes related to the present invention, (1) to (3) are shown in Figure A, and the subsequent processes (4) to (6) are shown in drawing B.

In this figure, the same reference numerals as in FIGS. 1, 3, and 4 indicate the same objects. The explanation will be given below in order starting from (1).

(1) When application program X1 sends data (message) ~ to application program Y2, application program
Data communication management system DCMS (hereinafter referred to as DCMS)
(abbreviated as ), the DCMS writes it to the main memory buffer 31 in the form of a message queue waiting to be processed, and also copies the data.
is stored in the same queue format on the file memory DASD 4, and the application program Y2 is notified of data transmission.

Application program Y2 that issues the communication macro “GET” to receive the message.
reads data from the main memory buffer 31 and processes it.

Main memory buffer 31 and file memory
The data on the DASD 4 is processed by the application program Y2 and released by the DCMS when the communication macro GET for receiving the next data is issued.

(2) When the system that is communicating with the application programs X and Y goes down, a system start-up process is immediately performed, and when the start-up is completed, the DCMS enters the next process (3).

(3) When the DCMS5 recognizes the startup from the application program Y2, it first loads the file memory DASD4.
Then, only the data is restored to the main memory buffer 31 on the main memory (memory) 3.

(4) DCMS5 notifies the application program Y2 of the restored data.

The application program Y2, which issues the communication macro "GET" to receive the message, reads the data to one address and processes it.

When the data processing is completed, the data is released from the main memory buffer 31 and file memory DASD 4.

(5) DCMS5 has a dedicated task that performs the restoration process, and this task is the application program Y.
2, the number of unprocessed messages remaining in the waiting message queue on the main memory buffer 31 is monitored for each waiting message queue terminal 31', and the number of unprocessed messages is kept constant. When the number (for example, the number of application program tasks x 2) or less is reached, restoration of the next data stored on the file memory DASD 4 is executed.

Therefore, in the present invention, in response to a request from the application program Y2, a dedicated task of the DCMS 5 monitors the number of unprocessed messages remaining in the processing waiting message queue on the main memory buffer 31, and calculates the number of messages. When the number becomes below a certain number, the application program Y2 is controlled to restore the next data, and the application program Y2 itself can concentrate on message processing without being conscious of restoring the data.

(6) Upon receiving the restoration request, the DCMS 5 restores the data on the file memory DASD 4 onto the main memory buffer 31 and notifies the application program Y2. At this time, if the data has already been processed, the data is also restored onto the main memory buffer 31 and notified to the application program Y2.

After that, the operation returns to (5) and the file memory DASD
4 until there is no unprocessed saved data (5)
(6) Repeat the loop of (5).

In this way, data is dynamically restored from the file memory DASD 4 to the main memory buffer 31, so that the DCMS 5 and the application programs under it can efficiently process data.

〔Effect of the invention〕

As explained above in detail, the message recovery processing method of the present invention is designed to recover unprocessed messages stored on the file memory DASD 4 when a system failure occurs during communication between application programs X and Y. When restoring on the memory buffer 31 in the form of a message queue waiting to be processed, the number of unprocessed messages in the queue of messages waiting to be processed is always monitored, and if the number of unprocessed messages falls below a certain number, At this time, the next message on the file memory DASD 4 is restored onto the main memory buffer 31 and connected to the queue of messages waiting to be processed.

At this time, the main memory buffer 31 is designed so that the restoration process can be performed by a dedicated task provided in the data communication management system DCMS in parallel with the message processing in the application program Y2. Since the restored message is not occupied and the message restoration processing is performed in a dedicated task separate from the application program Y2, the application program Y2 can concentrate on processing the message, improving the processing efficiency of the entire system. There is an effect that can be done.

[Brief explanation of drawings]

Fig. 1 is a block diagram of the principle of message recovery processing of the present invention, Fig. 2 is a diagram schematically showing an embodiment of the present invention, and Fig. 3 is a diagram schematically showing the concept of communication between general application programs. FIG. 4 is a diagram schematically showing a conventional improved application program communication system. In the drawings, 1 is an application program X, 2 is an application program Y, 3 is a main memory (memory), 31 is a main memory buffer, 4 is a file memory (DASD),
5 is the data communication management system DCMS,...
indicates data (message), 'PUT' indicates a communication macro for sending data, and 'GET' indicates a communication macro for receiving data.

Claims (1)

[Claims] 1. Under the control of a data communication management system (DCMS), there are a plurality of application programs that process messages, and when one application program (X) sends a message, the message is Repeating the operation of writing the message into the memory buffer, notifying the other application program (Y) 2 of the writing, and causing the application program (Y) 2 to read and process the message on the main memory buffer 31. Therefore, in a data processing system that processes a series of messages, the above message is stored in the main memory buffer 31 in the form of a processing queue, and at the same time, a copy of the message is also stored in the file memory (DASD) 4 as a message queue. a means for storing messages in the processing queue in order to restore messages on the file memory (DASD) 4 to the main memory buffer at the start of system recovery after a system failure of the data processing system; The number of messages is monitored, and when the number of messages falls below a certain value, the file memory (DASD) is
4, it is equipped with a means for restoring the next message on the main memory buffer 31 and connecting it to the queue waiting for processing, a means for writing into the file memory (DASD) 4, and a means for writing from the file memory (DASD) 4. 1. A message recovery processing method, characterized in that a message recovery means performs system restart processing at the start of system recovery. 2. As the message recovery means, a message recovery task is provided in the data communication management processing system (DCMS), and the message is recovered by the task in parallel with the message processing of the application program (Y) 2. A message recovery processing method according to claim 1, characterized in that: