JPS5856142B2 - Online recovery control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an online recovery control system, and particularly when a data cut failure occurs, for example, a read or write error occurs in a storage device that stores data, a program error, a parity error, etc. When this occurs, the latest data set can be quickly restored based on the entire dump data that has been read in advance, the log data set number and the log data block number that have been written at the beginning of the dump data. Concerning recovery control method.

For example, online recovery when processing online business such as inventory management business or banking business (4) has conventionally been performed as follows.

That is, FIG. 1 is a block diagram of a system that executes a conventional online recovery method, in which 1 is a magnetic disk that stores, for example, an inventory management data group, a bank account balance data group, etc. (hereinafter simply referred to as a data set); A magnetic tape device stores all contents read from the magnetic disk (all dump data); 3-1 and 3-2 are processing means (hereinafter simply referred to as application programs) for executing predetermined processing by application programs; 4 is a processing means for executing predetermined processing by application programs; A magnetic tape device for storing log data (hereinafter referred to as a log data set), 5 a data set recovery circuit for restoring the data set after a failure occurs, 16 and 17 terminals, and 51 converting the contents of the magnetic disk 1 to the magnetic tape 2 Write control means for reading and writing; 52 is a write control means for writing log data into the log data set 4;

First, when the online service is stopped, the contents are read from the magnetic disk 1 under the control of the write control means 51, and all dump data is written to the magnetic tape device 2 to obtain original data.

This work may be performed, for example, after the online business has finished the previous day, or before the online business starts.

When a job is started and, for example, a message is transmitted from the terminal 16, the application program A prepared for the terminal starts, and the data up to that point regarding the job input from the terminal is downloaded. Read from the data storage area 61 of the magnetic disk 1 (for example, in the case of inventory management, the number of goods A in stock, in the case of banking business, Mr. A's deposit balance \)
The read data is processed based on the data input from the terminal (for example, Mr. A's deposit amount), and the updated data obtained as a result is stored again in the data storage area 6-1.

At the same time, the input data is separately stored in the log data set 4 as log data by the write control means 52.

After processing such an application program several times, an input is received from a certain terminal, and application program B is started in response to the input, and the magnetic disk 1 is
The data is read from the data storage area 6-2 of the terminal, processed based on the data input from the terminal, and when the processed results are stored again in the data storage area 6-2, the data is stored on the magnetic disk 1. It is assumed that a failure has occurred.

In this case, it is necessary to immediately restore the data set to the latest version in order to continue online operations.

At this time, in order to restore the data set, it is necessary to have the original data before the start of the day's work and log data indicating what kind of processing was being performed before the failure occurred.

Therefore, as mentioned above, the data set must be restored by repeating this process using the original data obtained by dumping all data from the magnetic disk 7 before the start of work and the log data processed up to the time of the failure. .

However, in such conventional methods, in order to obtain the original data necessary for data set restoration, it is necessary to stop the business process, and therefore, it is possible to obtain the original data required for data set restoration only in a very limited state such as before the start of business. It is only possible to obtain the original data.

Therefore, if a certain amount of time has elapsed between when the original data was obtained and when a failure occurs, a large number of application programs, that is, business processes, will have been processed by the time the failure occurs. Revival of that data set! This had the disadvantage of being time consuming.

Therefore, an object of the present invention is to dump a dataset at any point during online execution and reduce the amount of log data required to restore the dataset after a failure occurs.
Online recovery side (foundation) that solved these conventional problems
The purpose of this invention is to provide a data set storage means that is updated based on input data and stores the updated results, and a data set storage means that reads out the entire contents of the data set storage means when offline. a total dump data storage means for storing contents according to the input data, a log data storage means for storing contents according to the input data, and a data set storage means for storing contents of the total dump storage means and the log data storage means after a failure occurs. In an online recovery control unit method in an online system, the method includes a data set recovery means for recovering the contents of the log data, and a means for sequentially storing log data in the log data storage means and storing identification information for identifying each log data. , the entire contents of the dataset storage means can be read out to the total dump data storage means at any time during online execution, and when the contents are stored in the total dump data storage means, the control of the total dump data storage means ( The identification information at the time is stored in the iRItif information recording area, and after a failure occurs, the contents of the data set storage means at the time stored in the total dump data storage means and the log data identification information stored in the control information storage area are stored. This is achieved by using a data set recovery means to recover the contents of the data set storage means from the contents of the log data storage means stored between the time obtained using the method and the point of occurrence of the failure. .

Hereinafter, the present invention will be explained in detail with reference to the drawings.

FIGS. 2a to 2c are block diagrams illustrating the present invention in detail, in which 6 and 7 are magnetic disks that store data sets, 8 and 9 are application programs that execute predetermined processing, and 10.11 14 is a control table, and 15 is provided corresponding to each application program. Part of entry for storing control information, 16.1
7 is a terminal, 18 is a magnetic tape device that stores all dump data, 19 is a control information storage section that stores control information, 2
0-1 to 20-4 are data storage units for storing data, 2
1.22 is a log data set.

Note that a plurality of log data sets (2L22, . . . ) are provided, each of which is distinguished by a data set number, and each log data set is composed of a large number of blocks.

Now, in the present invention, as shown in FIG. 2a, an entry is made in the control table 14 for each application program.
- IJ part 15 is allocated and the entry part 1
5 contains history control information, that is, the number of the log data set in which the latest log data is to be recorded (log data set number), and the number of the block of that log data (block number).
and a status display signal (STATUS) 15-1, etc. are stored.

This status display signal contains log data during processing based on the application program.
(21, 22) In the middle of displaying whether or not the log acquisition display M1 has been recorded, there is a start display ST that displays the start of processing by the application program, and an end display ED that displays the end of processing by the application program.

Assuming that the application program 8 is currently executing processing in response to a request from the terminal 16, log data to store input data from the terminal 16 is stored in the entry part 15 corresponding to the application program 8. The log data set number of the set and the block number of the log data are read from a control information storage section of the log data set, which will be described later, and are recorded.

Moreover, the status signal is also recorded according to the status.

When the processing based on the application program starts, the status display signals (ST.M, ED) are (1,0
, 0), if log data is recorded in the log data set during processing (1, 1, 0), if processing is completed (1, 1
,1).

On the other hand, all dump data is transferred to the magnetic tape 18 from the data set 6 or 7 under the control of the write control means 51, as shown in FIG.

Note that a control information storage section 19 is provided at the beginning of the data sections 20-1 to 20-4, and the control information storage section 1
9, when acquiring all dump data onto the magnetic tape 18,
Part of the entry corresponding to the application program being executed at the time of the dump data acquisition request, or if there is no running program, part of the entry corresponding to the application program whose processing was completed most recently based on the time of the dump data acquisition request. The log data set number and log data block number read from the log data are recorded.

Further, as shown in FIG. 2C, the log data set 2L22 includes input data (log data) input from the terminal through the processing of the application program, for example, in the case of banking, account numbers, deposits or payment amounts, deposits or Change item data consisting of types of payments, etc. are recorded one after another.

Note that the log data sets 21 and 22 are a control information storage section 21-1.21-2 and a data storage section 21-2.2, respectively.
2-2, and the log data set number, block number, and entry start address (entry pointer) are recorded in the control information storage section.

That is, the data set number and block number in which the latest log data should be recorded are always recorded in the control information storage unit, and once the writing of the log data is completed, the block number of the log data changes, for example, by one step. Also, when one log data set becomes full with log data, the log data set number incremented by one step and the log data block number /i are updated to the control information storage section of the next log data set.
61 is recorded, and thereafter, the block number is updated by 1 each time writing of log data is completed until the log data set is full.

Further, in this control information storage section, an entry pointer indicating an entry corresponding to an application program currently being executed or whose processing has been completed most recently is written as shown below.

That is, this writing of the entry pointer is carried out by the control means (not shown) reading out the entry pointer stored in the task control means corresponding to the application program when a predetermined application program is activated in response to a request from the terminal. This is done by writing to the control information storage section.

Therefore, the control information storage unit 21-1 or 21-2 of any one of the log data sets contains the log data set number and log data block number indicating the position where the latest log data should be written, and The start address (entry pointer) of the entry corresponding to the application program whose processing was completed most recently is stored.

Next, the operation of the present invention will be explained in two parts: (1) data set update and log data accumulation processing, and (2) data set dump processing.

(1) Data set update and log data accumulation processing ■
At the initial stage, the log data set number indicating the position where the latest log data should be written and the block number of the log data are recorded in the control information storage section of the predetermined log data set 21-22.

(2) Next, when an online job request signal and input data are input from the terminal 16, for example, the task control means 10 instructs the application program 8 to start up.

As a result, the application program (a) reads the data stored in the data storage area 6-1 of the magnetic disk 6, (b) performs predetermined processing based on the input data, and (/→ writes the input data to the log data set). writing process,
b) A process of storing the processed result in the storage area 6-2 is executed.

Note that the data processing results are also answered parked in the terminal device 16.

■ As mentioned above, when a predetermined application program starts running in response to a request from the terminal device 16 and starts executing processing, the control means (not shown) in parallel sets an entry pointer from the task control means 10 corresponding to the application program. The data is read and recorded in the control information storage section of a predetermined log data set.

As a result, an entry pointer indicating the start address of the entry corresponding to the application program being executed has been written in the control information storage section of the data set.

Note that even if the processing of the application program ends, this entry pointer is held until the next application program is started.

Therefore, it can be said that a pointer to an entry corresponding to an application program that is currently being executed or has completed processing most recently is written in the control information storage section.

(2) Next, the control means (not shown) reads the log data set number and log data block number that have been written in advance in the control information storage section of the log data set, and writes them to the entry section corresponding to the application program. It's crowded.

As a result, the log data set number into which log data should be written and the log data block number are stored in the entry section corresponding to the application program being executed.

At the same time as recording these numbers, the status signal recording section of the entry section also records the status according to the processing of the application program, that is, processing start (1, 0, '0), writing of log data to the log data set. Completed (1, 1, 0),
Process end (1, 1, 1) is written.

■ When log data is written to a log data set based on the processing of the application program, the block number of the log data in the control information storage section of the log data set is updated by 1, and the block number in which the next log data should be stored is instruct.

■ Application program processing ends.

Repeat operations from ■■ to ■.

As a result, the data set on the magnetic disk 6.7 is updated one after another, and the log data from the terminal 16.17 is accumulated in the log data set.

The data set update process and log data accumulation process have been described above, and next, the process of dumping the data set onto the magnetic tape 18 will be described.

(2) Data set dump processing ■ When a request to dump all the dump data recorded on the magnetic disk 6.7 to the magnetic tape 18 is generated, the control means (not shown in the drawings) processes the log data to which the latest log data should be written. The entry pointer written in the control information storage section of the set is read, and the entry in the control table 14 pointed to by the entry pointer is accessed.

As mentioned above, this entry section is an entry corresponding to the application program that is being executed at the time the dump request is generated or has completed execution most recently based on the dump request generation time, and the latest log data is stored in this entry. The log data set number to be written or written, the log data block number, and the status signal are written.

(2) Next, the status signal of the encoder 1-IJ- is read, and the processes (a) to (c) are performed depending on the status.

(G) When the status signal is (1, o, 0)...In this case, the log data has not yet been written to the log data set, and the data set is set based on the log data. has not been updated.

Therefore, since the data set before the update is dumped, this log data is necessary for recovering the data set after a failure occurs.

Therefore, the log data set number and log data block number in which the log data is to be stored, that is, the log data set number and log data block number written in the entry, are stored in the control information storage section of the magnetic tape 18. write to.

(C]) When the status signal is (1, 1, 0)...
In this case, the log data has already been included in the log data server H, but the data set has not been updated yet.

Therefore, the unupdated data set based on the log data will be dumped, and since this log data will be required when recovering the data server after a failure, it will be stored in the entry as in (1). The current log data set number and log data block number are written in the control information storage section of the magnetic tape 18.

(c) When the status signal is (1, 1, 1))...
In this case, the log data has already been written to the log data set, and the update of the data set has been completed based on the log data.

Since the updated data set is dumped, this log data is not necessary when recovering the data set after a failure occurs.

Therefore, the block number obtained by adding 1 to the block number of the log data stored in the entry and the data set are written to the control information storage section of the magnetic tape 18.

(2) Next, the entire data set on the magnetic disk 6.7 is dumped onto the magnetic tape 18 under the control of the write control means 51.

After that, the updating of the data set and the log data accumulation process ■~■ are repeated, and the data set is updated one after another.
Also, log data is accumulated in a log data set.

Therefore, if a failure occurs at a certain time, all log data from the time of the full dump to that time will be stored in the log data set, and the log data will be written after the full dump due to sequence ■. The first position is written in the control information storage section of the magnetic tape 18.

■ When a failure occurs and the data set is recovered after the failure, all the dump data written on the magnetic tape 18 and the position information (log data set) written in the control information storage section of the magnetic tape 18 are recovered. The data set is recovered by a predetermined recovery program using the log data stored from the log data set position indicated by the log data set position (number, block number) to the log data set position immediately before the failure occurred.

As described above, according to the present invention, all datasets can be dumped at any time during online execution, so the time from when all dumped data is obtained to when a failure occurs can be shortened. In other words, data sets can be recovered quickly. Since less log data is required, the time required to recover a data set can be significantly reduced.

In the above description, the control table 14 is provided separately, but the control table may be removed and its function may be taken over by a task control means provided corresponding to each application program.

Further, although the case where a plurality of log data sets are provided has been described, it is also possible to provide only one log data set.

Furthermore, although the case has been described in which the block number of log data is updated by 1 each time log data is written, the present invention is not limited to this, and it is sufficient to be able to indicate the location where the next log data should be stored. be.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram for explaining a conventional online recovery control method, and FIG. 2 is a block diagram for realizing the online recovery control method according to the present invention. 6.7... Magnetic disk for storing data sets, 8, 9... Processing device operated by application program, 10, 11... Task control means, 14... ...Control table, 1
5... Entry section, 18... Magnetic tape device for storing all dump data, 19... Control information storage, 21... Log data set (magnetic tape device).

Claims (1)

[Scope of Claims] 1. Data set storage means that is updated based on human data and stores the updated results; full dump data storage means that reads and stores the entire contents of the data set storage means when offline; a log data storage means for storing contents corresponding to input data; and a data set recovery means for restoring the contents of the data set storage means based on the contents of the entire dump storage means and the contents of the log data storage means after a failure occurs. In an online recovery control method for an online system equipped with a system, log data is sequentially stored in a log data storage means, and means for storing identification information for each log data is provided, and all of the data set storage means are stored in the entire dump data storage means. The contents can be read at any time during online execution, and when the contents are stored in the total dump data storage means, the identification information of the log data at the time is stored in the control information storage area of the total dump data storage means. , after the failure occurs, the period from the time to the time of failure, which is obtained using the contents of the data set storage means at the time stored in the total dump data storage means and the log data identification information stored in the control information storage area; An online recovery control method, characterized in that the data set recovery means recovers the contents of the data set storage means from the contents of the log data storage means stored in the log data storage means. 2. The online recovery processing method according to claim 1, wherein position information indicating the position of the log data storage means in which the log data is to be stored is used as the identification information of the log data. 3. Provide a state storage means for storing whether writing of the log data to the log data storage means has been completed and whether updating of data to the data set storage means has been completed; If the writing has been completed but the updating has not been completed, the location information is stored in the control section clear information storage area, and if both the writing and updating of log data have been completed, the location information is stored. 3. The online recovery control method according to claim 2, wherein position information obtained with predetermined modification is stored in the control information storage area.