JPS58168155A - Recovery system in duplex system - Google Patents

Abstract

PURPOSE:To make the recovery of data set easy, by recording career log data on a portable storage medium after revision, and writing a special control log data on a log file after the start and end of the operation. CONSTITUTION:Log environment, job program and resource group are defined in an environment definition file CIRP1. A host computer H performs jobs by using resource groups A, B. The data after revision to the resource groups A, B are stored in a log file LOGP1 as carrier log data after revision. The log file LOGP1 is a file recording recovery data when a failure takes place in the resource groups A, B. Special control log data are written on the log file LOGP1 at the start and end of operation.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to a recovery method in a duplex system that operates as one system during certain time periods and as two systems during other time periods. be.

[Conventional technology and problems]

When operating a duplex system, one
It has been proposed to perform online processing as a system, have two systems at night, and have each system perform batch processing with each data set van. Generally, when updating a dataset, the dataset is updated after acquiring the post-update history log data, but if this method is applied to the operation of the duplex system as described above, , one updated layer in case of l system! Post-update log data is written to the IIa log data file, and in the case of two systems, each system has its own post-update history file and writes post-update history log data to the corresponding post-update history log data file. I'm here.

For example, when two systems are operating together, if an abnormality occurs in the ship, data set, or data storage medium KIO failure, and the corresponding data set is to be recovered, first the system configuration of the Dublemax system and the backup data as the system are recovered. Recovery is performed using the update history log data file that was collected during the file and the system, and then
It was necessary to have a system configuration of two systems, and each system to perform recovery using the corresponding updated history log data file Y. but,
The recovery method in such a deubremax system has the disadvantage that the recovery procedure is extremely troublesome.

[Object of the Invention] The present invention eliminates the above-mentioned disadvantages and provides a recovery method for a duplex system that facilitates the recovery of 4CV of data in a duplex system. [Structure of the Invention] Therefore, the recovery method in the duplex system of the present invention is as follows.

In a Dublemux system that is operated so that data is stored in different log files, the control log data is stored in the log file to indicate normal order at the start and end of each time period. When a person records the historical log data on a portable recording medium after updating the log file and performs data set recovery, the user records the historical log data on a portable storage medium after updating the log file, and then records the log data related to the cough data set or the storage medium V.
The input source for IJ recovery is ytlf/# mold.

[Embodiments of the invention]

The present invention will be explained below with reference to vWAmv.

FIGS. 1A and 1B are diagrams showing an example of the operation of a duplex system, and FIG. 2 is a diagram showing one example of the collection of updated history log data.

In the 11i1, H and H' are host computers, A and B
are resource groups, LOGPl, LOGP2, and LOGP2″ are log files, CIRPI, CI RP 2, and CI RP
2' indicates the environment definition file t. Note that the log file LOGP1. LOGP2 and LOGP2
' is configured on DASD.

The 11th EI B) shows the system configuration at time period PI. The environment definition file CIRPI defines the log environment, business program, resource group, etc. for the 1 time period P1. In the 0 time period PI, as shown in the figure, the host computer H performs business operations using the resource groups A and Bt. There is. The updated data for resource groups A and H are stored in the log file LOGPI as updated layer jlag data. The log file is a file that records data Y for recovery when a resource group A and BK failure occurs.

FIG. II1 ρJ shows the system configuration in time period P2. In time period P2, the duplex system is operating as two systems operating independently. The host computer H is performing business Y using the resource group AV, and the host computer H' is operating the resource group AV. I am doing work Y using group BY. The updated data for resource group AK is stored in LOGP2 of the log file as updated history log data.

The updated data for the resource group BK is written to the log file LOGP2' as updated history log data. 1st
The symbols in the figure indicate an online processing station that processes resource groups A and BY simultaneously during the day, and Figure 1 p) shows an online processing station that processes data separated by Y for each resource group A and B during the night. It represents the disposition.

If operation 1 in Figure 1 (b) is performed following operation 1 in Figure 1 B), the log data for recovery in 2 will be as follows.
- Resource group A...LOGPI+LOGP2 resource group B
...LOGP1+LOGP2' FIG. 2 is a diagram showing an example of acquiring history log data after update. In Figure 2, BPA and BPB are backup files, LO
GPI-1 and LOGP2-1 and LOGPI-2 and LOG
P2'-1 is a log file, LOGBPI-1 and LOG
BP2-1, LOGBPI-2, and LOGBP2'-1 are shown in the log files as backup files, respectively. Note that system 1 is a system on the host computer H side, and system 2 is a system on the host computer W side.

Also, log7eye/kLOGP1-1. LOGP2-1
゜LOGPl-2 and LOGP2'-1 are configured on DASD, and backup LOGBPI-1, LOGB
P2-1. LOGBPI-2, and LOGBP2'-
1 is constructed on a magnetic tape.

At the start of the first time NPl, backup data to the resource group is taken and backup file B is created.
PAK is stored, and backup data of resource ##B is also taken and stored in backup file BPH.

The updated data for resource groups A and B in the first time period PI is stored in a log file k as updated history log data.
Stored in LOGP 1-1. Log file LOG
The updated history log data of Pl-1 is stored in the backup 7-aisle LOGBPI-IK at an appropriate timing, for example, at the end of the time period PI. Backup file kLOGBP1-1, LOGBP2-1, LOGB
PI-2 is constructed on magnetic tape.

In the first time period P2, the updated data for resource group A is stored as updated log data in log file LOGP2-1, and the updated data for resource group BK is stored as updated history log data in log file LOGP2'-1. IK stored 0 For example, at the end of time period P2, the updated history log data of log file LOGP2-1 is stored in backup LOGBP2-1, and log file LOGP2
The updated history log data '-1 is stored in the backup file LOGBP2'-IK.

In the second time period P2, the updated data for the resource groups A and H are stored in the log file PI-2. For example, at the end of the second time period P2, the log file LO
The updated history log data of GPL-2 is stored in the backup file LOGBPI-2.

In FIG. 2, when a failure occurs in the user file at time i1, the input resources for recovery are as follows.

File person backup BPA + LOGPl-1 or LOGBPI
-1+LOGP2-1 or 1,0GBP2-1+LOG
Pl-2 or LOGB Pl-2 file B Backup BPB + LOGP1-1 or LOGBPI
-10LOGP2'-1 or LOGBP210LOG
Although the normal order of log data is guaranteed within each log file of Pl-2 or LOGBPI-2, the normal order is not guaranteed at the end of operation. Therefore, special control log data can be written to a log file at the start and end of operation according to the user's instructions. Correctness is checked at the end of operation by checking the consistency of control log data, and if they do not match, the decision on recovery is left to the user. Recovery is performed using only either host computer H or host computer H'.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, the post-update history log data acquired by each system is finally recorded on a portable recording medium, and the data is recorded by percentage at the start and end of operation. Since the control log data is written to the sound log file, even if a duplex system is operated in an online format with one system during the day, and in a batch processing format where data is shared among multiple systems at night. Even in such cases, data set recovery can be easily performed.

[Brief explanation of the drawing]

Figures 1 and 2 are diagrams showing an example of the operation of a duplex system, and Figure 2 is a diagram showing INt' of acquisition of updated history log data. H and H' are host computers, A and B are resource groups, LOGPI
and LOGP2 and LOGP2' are log files, CI R
PI, CIRP2 and CI RP2' are environment definition files, BPA and BPB are backup files, LOG
PI-1 and LOGP2-1 and LOGP1-2 and LOGP
2'-i are log files, LOGBPI-1 and LOGB
P2-1, LOGBPl-2 and LOGBP2'-1 are backup files that can be used as log files. Patent Applicant: Fujitsu Ltd. Representative Patent Attorney Kyotani Shibujo' 1 Figure (A) Page 1 Figure (B)

Claims (1)

[Claims] In a duplex system that is operated in such a way that data is stored in different log files, control for indicating normal order at the start and end of each time period is stored in the DADAME V log file, and the time period After updating the log file at the end of the log file, the history log data can be stored in a portable memory and the data set is recovered IJ% - Row 5III%
A recovery method in a duplex system that uses log files or recording media associated with a data set as an input source for V4I recovery.