JPH0235529A - Check point dumping acquisition system - Google Patents

Abstract

PURPOSE:To dump only information with a flag and to decrease the number of data which are acquired by providing an update/nonupdate flag corresponding to a table to be recovered which is obtained by check point dumping. CONSTITUTION:A recovery object area control table 11 has an object table kind 11a, an area relative location 11b, and key information 11c as a control information for the recovery object table. In its entry information, the flag relative location 1d for determining which part of the table the actual update/ nonupdate flag for the table is present at and a flag setting area are provided. The flag setting area is set by table kinds by recovery information setting processing in initial setting processing and information regarding a terminal control table is also set in the recovery information setting processing.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a checkpoint dump acquisition method, and in particular to a checkpoint dump acquisition method that reduces the amount of data acquired by checkpoint dump in order to perform online program recovery processing. This relates to the dump acquisition method.

[Conventional technology]

Traditionally, in online processing of data processing systems,
When a failure occurs, recovery processing must be performed immediately. Therefore, checkpoint dump information and journal information for full recovery processing are acquired. In the full recovery process of the online control program, checkpoint dump information and recovery journal information are used to recover information on the virtual memory. In order to prevent full recovery processing from taking a long time, when the amount of acquired journals exceeds a certain value, checkpoint dump acquisition processing is performed to acquire all data related to one recovery area. . Furthermore, the recovery journal acquisition process acquires journal data when the recovery target area is actually updated. In full recovery processing, after initializing the area to be recovered, the information obtained from the latest checkpoint dump is overlaid on the area to be recovered, and the information from the latest checkpoint dump of this area is used for subsequent recovery. Reflect the information in the journal one by one. This recovers the information on the virtual memory.

In this way, recovery processing using checkpoint dump information and recovery journal information is performed.
There is a "forced check point processing method" described in Japanese Patent Publication No. 62-56535.

[Problem to be solved by the invention]

By the way, in the conventional checkpoint dump processing as described above, all data related to the recovery area is acquired in the checkpoint dump. in this case,
It is not taken into account that information to be recovered that has not been updated many times remains in its initial state.In other words, regardless of whether or not there has been an update, all the information to be recovered is lost during the checkpoint dump acquisition process. is being obtained. Therefore, since full recovery processing acquires information that is not necessarily required, there is a problem in that a large amount of information is acquired through checkpoint dumping.

The present invention has been made to solve the above problems.

An object of the present invention is to provide a checkpoint dump acquisition method that reduces the amount of data acquired in a checkpoint dump for performing online program recovery processing.

The above and other objects and novel features of the present invention will become clear from the description of this specification and the accompanying drawings.6 [Means for Solving the Problems] In order to achieve the above objects, the present invention .

In a data processing system that fully recovers information on virtual memory using checkpoint dump information and recovery journal information, an actual update flag is set for the recovery target table obtained by checkpoint dump, and actual update is performed. A feature of this method is that only information about recovery target tables whose presence/absence flag is on is acquired by checkpoint dump.

[Effect]

According to the above means, an actual update presence/absence flag is provided corresponding to the recovery target table to be checkpoint dumped. Then, in the 5-ro recovery target area at the time of checkpoint dump acquisition, only the information of the recovery target table for which the actual update flag was turned on and the actual update was performed is acquired by the checkpoint dump. Therefore, among the 1 recovery target areas, information is not acquired by checkpoint dump for areas where the actual update presence/absence flag is not turned on and no actual update has been performed. This is because the state before 1 recovery processing remains the initial state, and even in full recovery processing,
This is because the information of the later checkpoint dump is not required.

The area for actual update information is specified at the same time as the recovery target area is declared during system initialization processing.

In this process, the relative location and bit position of the actual update information area are stored in the recovery target area information.1 When the recovery target area is updated, the 1 recovery journal acquisition routine is called, within this routine, the recovery target A flag indicating that there has been an actual update is set in the actual update presence/absence information in the area information. During checkpoint dump acquisition processing, the relative location and bit position of the actual update presence/absence flag are determined from the actual update presence/absence information in the recovery target area information for the corresponding area, and the actual update presence/absence flag determines whether or not there has been an actual update. Obtain only information about tables that have been updated but are subject to recovery using a checkpoint dump. This makes it possible to reduce the amount of data acquired by checkpoint dump. In full recovery processing, the latest checkpoint dump information after creation of the recovery target table is reflected, but those that have not been updated remain in their initial state. Even in the method of acquiring all information in the recovery area through checkpoint dump acquisition processing, the checkpoint dump information remains in its initial state for items that have not been updated. The recovery process can be performed in the same way even if the checkpoint dump is not acquired. In other words, for items that have not been updated, the same information as the initial state is acquired in the checkpoint dump, and based on that information, in full recovery processing, the initial state information is overlaid with the latest checkpoint dump information for recovery. Since processing is performed, similar recovery processing will be performed.

Therefore, even if the checkpoint dump is used to acquire only the information of the recovery target table that has been actually updated, the recovery process is similar to the conventional recovery process.

In this way, when acquiring checkpoint dump information, only the information necessary for full recovery processing is acquired, so the amount of checkpoint dumps is reduced and the overhead of acquiring checkpoint dumps can be reduced. Furthermore, the time required for full-scale recovery processing can be shortened.

〔Example〕

Hereinafter, one embodiment of the present invention will be specifically described using the drawings.

FIG. 1 is a diagram showing the relationship between management tables of main parts in an online control program for a data processing system according to an embodiment of the present invention. In FIG. 1, 10 is a system management table, 11 is a recovery target area management table, and 12 is a terminal management table. Reference numeral 13 denotes each terminal device, which is managed by control data in an area corresponding to each terminal in the terminal management table 12.

The system management table 10 is a table for managing online processing control tables, and the recovery target area management table 11. It has the address of a management table for controlling individual devices such as the terminal management table 12. In the initial setting process performed at the time of starting online, these management tables are created and online processing is started. In this way, in the data processing system, the online control program performs online processing based on the data in each management table created at the time of initialization. During this time, information is obtained by checkpoint dump for failure recovery, Get the recovery journal. When a failure occurs, the checkpoint dump information and the recovery journal information are used to immediately perform a full recovery process of the information on the virtual memory 1, for example, to recover the terminal management table and the like.

Hereinafter, a case where a predetermined recovery target area of the terminal management table is recovered will be explained as an example. The recovery target area management table 11 includes target table type 1, which is the entry information of the table, as management information of the recovery target table.
1a, recovery target area relative location 11b, and key information lie.

In this entry information, the actual update presence/absence flag relative location lid is added so that it can be determined where in the table the actual update presence/absence flag of the table exists.
An actual update presence/absence flag setting area is provided for setting the actual update presence/absence flag bit position lie. The actual update presence/absence flag setting area is set for each table type by the recovery information setting process in the initial setting process, and information regarding the terminal management table is also set in the recovery information setting process.

FIG. 2a is a flowchart showing the recovery information setting process in the initial setting process. Further, FIG. 2b is a flowchart showing one-recovery journal acquisition processing, and FIG. 2c is a flowchart showing checkpoint acquisition processing.

The checkpoint dump acquisition method according to this embodiment will be described with reference to FIGS. 2a, 2b, and 20.

First, in the recovery information setting process shown in FIG. 2a,
Information is set in each management table shown in FIG. That is, in step 21, 1 recovery target area information is set, and then in step 22,
Sets the relative location and bit position of the actual update information.

When updating the recovery target area 14 in the terminal management table 12, the online control program calls a recovery journal acquisition processing routine as shown in FIG. 2b. In the recovery journal acquisition processing routine of FIG. Next, each terminal table of the corresponding terminal management table 12 is obtained using the terminal name 15 as a key in the data of the target table type 11a, and the recovery journal of the recovery target area 14 is obtained. Step 23
When the recovery journal is acquired in the journal acquisition process, at the same time, based on the actual update presence/absence flag relative location lid and the actual update presence/absence flag bit position 1i11e stored in the recovery target area management table 11, in the next step 24, The actual update flag is set in the terminal management table 12.

Next, in step 25, it is determined whether the number of records in the acquired journal is greater than a predetermined specified value, and if the number of records in the journal is greater than the predetermined specified value, a check is made in the next step 26. Point dump acquisition processing is performed, and this recovery journal acquisition processing is ended. If the number of records in the journal is not greater than the predetermined specified value in step 25, the one-recovery journal acquisition process is immediately terminated.

In this way, in the checkpoint dump acquisition process that performs a checkpoint dump, the amount of journal acquisition is
Acquisition is requested and acquired when the value exceeds a predetermined value given in the system definition at the time of initial setup. In the checkpoint dump acquisition process in the recovery journal acquisition process, as shown in FIG. 2C, first, in step 27, all types of recovery target tables are checkpointed based on the information in the recovery target area management table 11. It is determined that the data has not been acquired by dumping.Next, in step 28, all entries of the relevant table are checked to see if the last entry of the relevant table has not been exceeded, and further, in step 29, the actual update is performed. In the next step 30, a checkpoint dump is acquired. In this way, only the update area of the corresponding table in which the recovery target area has been actually updated is acquired as a checkpoint dump.

FIG. 3 is a diagram specifically explaining the checkpoint dump acquisition process performed as described above.

FIG. 3 shows changes in the contents of the recovery target table acquired by checkpoint dump at each point in time. That is, the first checkpoint dump acquisition process 26a is performed on the recovery target table 31 in the initial state at time t1. Thereafter, the first actual update of the A table is performed and the state becomes the 1 recovery target table 32. A second checkpoint dump acquisition process 26b at time t2 is further performed on this recovery target table 32. Thereafter, a second actual update is performed on the A table, and a first actual update is performed on the C table, resulting in the state of the recovery target table 33. Then, at time t3, the third
The checkpoint dump acquisition process 26c is performed for the recovery target table in this way, so that when a failure occurs, the recovery process is performed immediately using the checkpoint dumped data.6 As shown in Figure 3. Based on a specific example of the recovery target tables 31 to 33 when a checkpoint dump is acquired in , the amount of data acquired by a checkpoint dump is calculated as follows.

That is, in the recovery target table 31 in the initial state,
There are four table elements (A, B, C, and D) to be acquired in the checkpoint dump, and the actual update presence/absence flags in the initial state are all O.

In this state, even if there is a checkpoint dump acquisition request, all the actual update presence/absence flags are O, and no actual update of each table element is performed.

In the first checkpoint dump acquisition process 26a,
Checkpoint dump data is not acquired, and the amount of data acquired is 0. In the second checkpoint dump acquisition process 26b for the recovery target table 32 at time t2, only the data of the A table that has been actually updated is acquired by checkpoint dump, so the amount of data acquired is 1. In addition, in the third checkpoint dump acquisition process 26c for the recovery target table 33 at time t3, the data of the A table that has been actually updated,
Since only the data of the C table is acquired by checkpoint dump, the amount of data acquired is 2. Therefore, the amount of data acquired by a total of three checkpoint dumps is 3 in total.

On the other hand, in the conventional checkpoint dump acquisition processing method, checkpoint dumps of four tables are acquired each time, so the amount of data acquired is 12 in total.

In this way, in the checkpoint dump acquisition method of this embodiment, information on only the updated A table and C table is acquired in the checkpoint dump, so the amount of data acquired in the checkpoint dump can be significantly reduced. Can be done.

The present invention has been specifically explained above based on examples, but
It goes without saying that the present invention is not limited to the embodiments described above, and can be modified in various ways without departing from the spirit thereof.

〔Effect of the invention〕

As explained above, according to the present invention, only the information necessary for full recovery processing is acquired from the data acquired by checkpoint dump, and the amount of data acquired by checkpoint dump is greatly reduced. can be reduced. In this case as well, the full recovery process can be performed using the same process as the conventional method, so there is no need to develop a new recovery process program. Furthermore, by reducing the amount of data acquired by checkpoint dump, it is possible to reduce the overhead of checkpoint dump acquisition processing and at the same time shorten the full recovery processing time.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the relationship between management tables of main parts in an online control program for a data processing system according to an embodiment of the present invention. FIG. 2a is a flowchart showing the @return information setting process in the initial setting process. FIG. 2b is a flowchart showing the 1-recovery journal acquisition process, FIG. 2c is a flowchart showing the checkpoint dump acquisition process, and FIG. 3 is a diagram specifically explaining the checkpoint dump acquisition process. In the figure, 10... system management table, 11...
Recovery target area management table, 12...terminal management table. 13...Terminal device, lla...Target table type,
llb...Recovery target area relative location, ilc
...Key information, 11d...Actual update presence/absence flag relative location. 11e: Actual update presence/absence flag bit position. Shape 1

Claims (1)

[Claims] 1. In a data processing system that performs full recovery processing of information on virtual memory using checkpoint dump information and recovery journal information, a real update flag is set for the recovery target table obtained by checkpoint dump, A checkpoint dump acquisition method characterized in that only information on recovery target tables whose actual update presence/absence flag is on is acquired by checkpoint dump.