1
INFORMATION PROCESSING SYSTEM
HAVING BACKUP/RECOVERY FACILITIES
This application is a continuation of application Ser. No. 08/109,600, filed Aug. 20, 1993, now abandoned. 5
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an information processing 10 system having backup/recovery facilities, more particularly relates to a backup/recovery technique of data introduced in a main storage in that system.
In information processing systems, generally use is made of the technique of backing up data by loading data of a 15 direct access storage device (DASD), for example, a magnetic disk unit, one type of external storage unit, in another external storage unit, such as a magnetic tape unit, to prepare against a fault or the like and recovering data by reading out the data from that other external storage unit when a fault 20 occurs in the DASD.
So as to carry out such a backup/recovery, there is a method in which the data is physically read out and transferred from the DASD in certain units of volume in the order 25 of the recordation and then is loaded in an external magnetic tape unit etc. In this method, however, the data (including a program) read out from the DASD to a main storage or a system storage unit (memory for expanding the main storage, abbreviated as SSU) is processed at a central processing 3Q unit (CPU), and therefore it is not the same as the original data. For this reason, so as to reliably perform the backup/ recovery of the data, it is necessary to fetch also the contents loaded in the main storage (or SSU) into the magnetic tape unit etc. 35
2. Description of the Related Art
Conventionally, where backup/recovery is carried out, as will be explained later with reference to the drawings, to back up the data of the DASD after the data of the main storage is reflected in it, it is necessary to write the data in 40 the main storage back to the DASD, so a long time is taken for performing the logical processing, for example, converting the address in the main storage to an address in the DASD. Further, after the data is written in the DASD, the data is further written in the magnetic tape unit etc. via a 45 buffer of the main storage, and therefore there arises a problem in that the processing load for the backup was large and a long time was taken. Also, even when recovering the contents after backup, after the data is loaded in the DASD from that magnetic tape unit etc., it is necessary to further 50 restore the state before the fault occurred by writing the same data as that immediately before the related backup is carried out into the main storage from the DASD, and therefore a problem of a long processing time occurred.
55
SUMMARY OF THE INVENTION
Therefore, an object of the present invention is to provide an information processing system having backup/recovery facilities of data introduced in the main storage, which can go perform the backup of the data introduced in the main storage without writing data back to the DASD and which can efficiently perform the processing when restoring the contents after backup to the main storage.
To achieve the above-mentioned object, the system of the 65 present invention is constituted in that it first is provided with a control table showing the region of the data intra
2
duced from the first external storage unit in the main storage in the first external storage unit. In the backup processing, if data which has not been introduced in the main storage is discriminated by that control table, that data is written in the magnetic tape unit etc. from that first external storage unit in certain units of volume, while the data which has been introduced in the main storage is read out from the main storage and written in the magnetic tape unit etc. In the recovery processing, the data read out from the magnetic tape unit etc. is written in the first external storage unit. At the same time, when it is discriminated that there is data which should be introduced in the main storage, that data is written at a corresponding address in the main storage. Thus, it is possible to perform backup of the data which has been introduced in the main storage without writing data back to the first external storage unit and thus to efficiently perform processing when restoring the contents after backup to the main storage.
BRIEF DESCRIPTION OF THE DRAWINGS
The above object and features of the present invention will be more apparent from the following description of the preferred embodiments with reference to the accompanying drawings, wherein: FIG. 1 is an explanatory view of the related art, FIG. 2 is a structural view of the principle of the present invention,
FIG. 3 is a structural view of hardware of an embodiment, FIG. 4 shows the flow of processing in the case of backup, FIG. 5 shows the flow of processing in the case of recovery, and
FIG. 6 is a view explaining a conversion of a data format in the main storage.
DESCRIPTION OF THE PREFERRED
EMBODIMENTS
Before describing the embodiments of the present invention, the related art and the problems therein will be described with reference to the related figures.
FIG. 1 is an explanatory view of the related art. In the system shown in FIG. 1, reference numeral 1 represents a DASD, which loads database sets; 2 is a main storage or SSU (hereinafter these will be simply referred to as the "main storage"), which takes the data out of the DASD 1 and performs the processing on it; 11 is a CPU; 8 is a magnetic tape unit, which is an external storage unit for the backup/ recovery; and 12 is a bus.
Where the backup is performed in certain units of volume, physical records in the DASD 1 are sequentially read out from the head and loaded in the magnetic tape unit 8. In this method, in the case of the data (represented by DS) introduced in the main storage 2 among the data in the DASD 1, the contents have been changed by the processing of the CPU 11 and no longer coincide with the contents of the corresponding original data on the DASD 1 (which has not become the newest content). When backing up the DASD 1, including also the data DS which has been introduced in the main storage 21, conventionally the following has been carried out.
The data DS loaded in the main storage 2 is processed in units of the data set, and the data is written once into the DASD 1 for every data set (route of "1" of FIG. 1). When this writing is ended, the data of the DASD 1 is transferred to the main storage 2 for each unit of the data set, and the
3
operations of loading data to the buffer (expressed by BF) therein (route of "2" of FIG. 1) and of writing data from the buffer BF into the magnetic tape unit 8 (route of "3" of FIG. 1) are repeatedly carried out. Also, when the data is recovered, the data is written into the DASD 1 from the magnetic 5 tape unit 8 (inverse route to "2" and "3" of FIG. 1), then the necessary data is logically fetched from the DASD 1 to the main storage 2 and is written in the same (inverse route to "1" of FIG. 1).
In the related art shown in FIG. 1, as previously men- 10 tioned, after the data of the main storage 2 is reflected in it, the data of the DASD 1 is backed up. Therefore, when the data in the main storage 2 is written back in the DASD 1, it is necessary to perform logical processing, for example, convert the address in the main storage 2 to the address in 15 the DASD 1, and therefore a long time is taken. After the data is written into the DASD 1, further, the data is written into the magnetic tape unit 8 via the buffer BF of the main storage 2, and therefore there was a problem in that the processing load for the backup operation was large and a 20 long time was required. Moreover, also, when recovering the contents which were backed up, it is necessary to restore the state before the fault occurred by loading the data from the magnetic tape unit 8 into the DASD 1 and then further writing the data the same as the data immediately before the 25 related backup was performed from the DASD 1 into the main storage 2. Therefore, there arose a problem of a long processing time.
FIG. 2 is a structural view of the principle of the present invention. Generally speaking, the information processing 30 system according to the present invention comprises a main storage; a first external storage unit which holds a large amount of data partially selected from that main storage for use; a second external storage unit which saves the aforesaid large amount of data held in that first external storage unit; 35 and a backup/recovery means including therein a control table showing the region of data, which was read out from the aforesaid first external storage unit and has been introduced in the aforesaid main storage, if any, in the first external storage unit, which backup/recovery means is 40 operative, when backup processing is started, to write the data, which is discriminated to be not introduced in the aforesaid main storage with reference to the control table, from the aforesaid first external storage unit into the aforesaid second external storage unit in certain units of volume 45 and to write the data, which is discriminated to be introduced in the aforesaid main storage unit with reference to the aforesaid control table, from the main storage into the aforesaid second external storage unit, and which backup/ recovery means is operative, when the recovery processing 50 is started, to read out the aforesaid data held in the aforesaid second external storage unit for the backup purpose and write it into the aforesaid first external storage unit and, when finding data which should be introduced in the aforesaid main storage with reference to the aforesaid control 55 table, to write this data into the main storage.
In FIG. 2, reference numeral 1 represents a DASD, serving as the first external storage unit in which a large number of database (DB) sets are loaded; 2, a main storage (including an expanded memory such as an SSU); 3, a 60 backup/recovery processing means; 4, a backup processing part; 5, a recovery processing part; 6, a control table holding information indicating in which region in the DASD 1 the data introduced in the main storage 2 is loaded; 7, a database management means which manages the data of the main 65 storage 2 when backup/ recovery processing is carried out; la, an address conversion table which holds the information
4
indicating the correspondence between the logical address in the main storage 2 and the physical address in the DASD; lb, a format conversion part which performs the conversion of format between the format of the data in the main storage 2 and the format of the data in the DASD 1; and 8, a magnetic tape unit such as the second external storage unit. Note that, it is clear that the first external storage unit is not restricted to a DASD and that other external storage units can be used. Similarly, it is clear also that the second external storage unit is not restricted to a magnetic tape unit and that other external storage units can be used.
According to the present invention, when performing backup processing, physical backup is carried out in certain units of volume from the first external storage unit 1 to the second external storage unit 8 and the data introduced in the main storage 2 is written from the main storage 2 into the second external storage unit 8. When performing recovery processing, the data is directly transferred from the second external storage unit 8, in which the data which was backed up was written, to the first external storage unit 1 and, at the same time, the data which had been introduced in the main storage during the backup is directly written into the main storage 2.
When the start of a backup operation is instructed, the backup processing part 4 of the backup/recovery processing means 3 is started. The backup processing part 4 reads out the data from the head address of the first external storage unit 1 DASD and writes that data into a magnetic tape unit serving as the second external storage unit 8 (hereinafter simply referred to as a "magnetic tape unit"). In this case, the backup processing part 4 judges whether or not the read out position of the DASD 1 (physical address) corresponds to the region of introduction in the main storage 2 (logical address) with reference to the control table 6. Where that read out position is a region in which data has not been introduced in the main storage 2, it writes the data as is into the magnetic tape unit 8 from the DASD 1 via the buffer of the main storage. Conversely, where that read out position is a region in which data was introduced in the main storage 2, it fetches the related data from the main storage 2 via the database management means 7.
Namely, the database management means 7 converts the address of the DASD 1 to the address of the main storage 2 by the address conversion table la when the data of the DASD has been introduced in the main storage 2, reads out the data from the main storage 2 by that address, converts that data to the format used for the DASD in the format conversion part lb, and transmits the same. That data is written into the magnetic tape unit 8 by the backup/recovery processing means 3. The other data, which has not been written into the main storage 2, is physically sequentially written from the DASD 1 into the magnetic tape unit 8 in the order of the address via the buffer of the main storage for performing the backup. As described above, the data is read out from the main storage 2 and that data is converted to a format adapted to the DASD in the format conversion part lb and then written into the magnetic tape unit 8. By this, the necessity of performing different control depending upon whether or not the data to be recovered is the data which should be returned into the main storage 2 is eliminated. Thus, an improvement of the efficiency of processing at the time of a recovery operation can be achieved.
When the start of a recovery operation is instructed, the recovery processing part 5 in the backup/recovery processing means 3 reads the data of the magnetic tape unit 8 which has been backed up by the above-described operation and writes the data into the DASD 1 via the buffer of the main
|
