JPS622335A - Back-up processing system for file device - Google Patents

Abstract

PURPOSE:To back up the contents of a file of the same event even when a data is being processed, by copying an unupdated part of a file device to a back-up file, and copying an updated part to the back-up file from a journal buffer. CONSTITUTION:When a host CPU 1 is processing a data, if a data of a DASD file device is updated, the data comes not to be the same event. Accordingly, a data before updating a track in which the data is updated is stored in a journal buffer 4, and also a data showing a track in which the data is updated is stored in a journal directory 5. Accordingly, a track for reading out a data from the DASD 2 in order to execute a back-up processing is referred to the inside of the directory 5 in advance, and if the data is unupdated, the data is stored in a transfer buffer 11 through a transferring circuit 10 from the DASD 2. At the same time, the entry of the directory 5 of the track which has read out the data is eliminated, and the data is stored in a magnetic tape 3 through a transferring circuit 9.

Description

[Detailed Description of the Invention] [Summary] Even when the processing device is processing data using a file device,
In order to enable simultaneous backup, a journal buffer (17) indicating the updated area of the file device during the backup processing period and a journal buffer for storing data before updating are provided, so that the area to be backed up on the file device is If the data has been updated, a backup of the data is taken from the journal buffer.

[Industrial application field]

The present invention relates to a backup processing method for storing data in a file device such as a magnetic disk device to a backup device such as a magnetic tape device, and in particular makes it possible to back up data at the same time without interrupting data processing. This invention relates to a backup processing method for a file device.

In systems that require reliability, such as bank banking systems, it is common practice to take a copy of data in order to improve data integrity.

The conditions required by the button to perform this backup include that all data to be backed up be collected in the same state, and that the impact on data processing be minimized for backup processing.

[Conventional technology]

Traditionally, the target of backup processing is magnetic disk devices, and magnetic tape devices are used as backup devices, but the timing for performing backup processing is at night when data processing is not being performed. This was done by copying data stored in a magnetic disk device to a magnetic tape device.

[Problems with conventional technology]

In the conventional method, the data processing device tl [
The backup had to be performed at a time when no other processes were being performed.

This is because when the data on the magnetic disk device to be backed up is updated during backup processing, there will be a conflict with the part that has already been backed up. As a condition for this, as mentioned above, all the data stored in the file device must be backed up with timing information data of the same number of hours.

[Means to solve the problem]

FIG. 1 shows a block diagram of an original wholesaler according to the present invention, in which a file device 2 is connected to a data processing device [11]. The journal buffer 4 is a file storage r! ! It is connected to t2 and stores the unupdated data of the area where the data is updated during the backup processing period.

The journal directory 5 stores data indicating areas of the file device 2 where data has been updated during the backup processing period.

The backup file device [3 backs up data in the file device 2.

[Effect]

As shown in FIG. 1, by copying the data stored in the file device 2 to the backup file device 3,
When taking a backup, data is updated for the file device 2 during the backup processing period so that the data processing device 1 can take a backup of the file device 2 while processing data using the file device 2. is performed, by storing data indicating the updated area in the journal directory 5VC and storing the data before the update in the journal buffer 41C, the unupdated portion of the file device 2 is stored in the file device 2. Copy it to the backup file device 3,
By copying the updated portion of the file device 20 from the journal buffer 4 to the backup file device 3, it is possible to back up the contents of the file device 2 at the same time even during data processing.

〔Example〕

FIG. 2 is a configuration diagram of an embodiment of the present invention, and FIG. 3 is a backup processing flowchart of the present invention.
Components with the same symbols as those in FIG. indicate data transfer buffers, respectively.

The process flowchart shown in FIG. 2 will be explained with reference to the configuration shown in FIG. 1.

The control storage 6 stores programs for the processor 7 to operate.

The transfer circuits 9 and 10 transfer data between the magnetic tape device 3 and a DASD (Direct Access Storage Device) 2, and transfer data between the CPUI and the DASD 2.

Data transfer bar/7a 4 is magnetic tape device 3 and DASD
2 and a data transfer buffer for temporarily storing data.

Host CPU 1 performs backup via channel 8.

When a backup instruction is issued, this instruction is received by the processor 7, and the Danop mode program C (backup processing program) on the control storage CSS executes a step 75 to start up, and starts backup.

The magnetic tape device that is first instructed to start the backup (only one magnetic tape device is shown in the example, but it is assumed that there will be more than one magnetic tape device).K Indicates the file name... Write a record. Next, the gross data 7 starts reading data from the DASD 2 via the transfer circuit 10 track by track, stores it in the transfer buffer 11, and transfers the data via the transfer circuit 9 to the transfer buffer II.
The data stored in Vc is written to the magnetic tape device W13.

Continuous output of data from the DASD 2 and writing of data to the magnetic tag stirrup 3 are generally performed using well-known controls, so they will be omitted.0The transfer from the DASD 2 to the magnetic tape device 3 will be omitted from the DASD 2. cylinder 10
1I, head 101I/cylinder, 011. Het 11
”・1・Noyo5 in order.

The extent to which data on the DASD has been backed up can be recognized by storing the cylinder and head numbers in the control storage 6.

Next, operations using the journal buffer 4 and journal directory 5 will be explained.

During this backup process, the boss CP'UI is processing data, so there is a possibility that data will be updated on the DASD2.

When this data update is performed, the data on the DASDZ no longer has the same number of hours.

Therefore, according to a method to be described later, additional data 11r@ of the track whose data has been updated is stored in the journal buffer 4, and data indicating the track whose data has been updated is stored in the journal directory 5.

Therefore, for the backup process, refer to the journal directory 5 in advance for the track from which data is to be read from the DASD 2, and if the data has not been updated,
Data is read from the DASD 2 via the transfer circuit 10 and stored in the transfer buffer 11, and the end l) of the directory 5 of the track concerned is deleted, and then the data is stored in the magnetic tape device 3 via the transfer circuit 9. do.

The extent to which the backup process has been performed is stored in the control storage 6 when the backup amplifier data is stored in the data transfer buffer IIK.

The data transfer buffer 11 has an area capable of storing at least two tracks of data.

Next, referring to the data on the control storage 6 in which the track number for which the backup process has been completed is stored, the track for which the backup process has been completed is the last track h.
-fCheck.

If it is not the last track, the above process is repeated.

Note that the data is read from DASD2 and transferred to the transfer buffer 1.
Since the timing or transfer time for storing data in 1 and the timing or transfer time for writing data from the transfer buffer 11 to the magnetic tape device 3 are different, the transfer bar/7
The data storage area 11 can sequentially store data for a plurality of tracks on the DASDZ at the same time, and enable sequential reading.

When the backup of the data on the DASDZ is completed until jlM's failure, the flag indicating Dump Mode is reset from the control storage C86, and the data remaining on the transfer bar I7A11 is transferred to the magnetic tape. Give a write instruction to device 3,
When this is completed, a trailer record is written and the backup process is completed.

The above operations are performed at a predetermined processing level of the processor, but normal host processing is performed at a higher processing level that takes priority over this processing. Update (write command processing will be explained.

Figure 4 shows the immediate flowchart of this write command.
When the processor 9 receives a write command from the host 1 via the channel 1' channel interface 8,
Check whether the DASD 2 that received the write command is in dump mode or not based on the flag set in the control storage 6, and if it is in dump mode, check whether the track to be processed by the write command is larger than the track being backed up.
That is, it is checked whether the backup acquisition has been completed by comparing it with the last track of the backup adapter stored on the control storage 6, and in the case of a track for which the backup acquisition has not been completed, the journal directory is further checked. Referring to step 5 above, if data has not been updated on the track during dump mode, if the data has not been updated, the data on the track will be transferred from the DASD2 to the journal buffer. 4 and register the corresponding tor)9 in the journal directory lJ5.

Upon completion of this process, the write command from the host 1 is executed.

On the other hand, in the case of a write command for a track for which backup acquisition has been completed or a write command for a track for which data has already been updated during dump mode, the write command is executed as is.

When the execution of the write command is completed in this way, the processor 7 returns to the backup process.

In the single-unit embodiment, commands from the host 1 other than write commands have been explained using a magnetic tape device as an example of a file device for storing backup data, but other file devices such as a magnetic disk device may be used.

〔effect〕

As described above, according to the present invention, data can be backed up for the same amount of time without interrupting host processing.

[Brief explanation of drawings]

FIG. 1 is a principle block diagram of the present invention. FIG. 2 is a detailed explanatory diagram of the present invention. FIG. 3 is a backup processing flowchart of the present invention. FIG. 4 is a write instruction processing flowchart of the present invention. ≧ Indicate each item. In the figure, numeral 1 indicates a data tab installation (host), 2 indicates a file device, 3 indicates a backup file device, 4 indicates a journal buffer, and 5 indicates a journal directory. Hearts Fat 7'' Fino Vhe J Ri Quan's Principle Pro Nfro 1 Figure 8゛° Sofa Soft 6'X!L Rifu D- + Eato
3T5 τ■1 included rehe") L/Write command, pqfu υ-+Tart Aoi 4Tf
J

Claims (1)

[Scope of Claims] A file device connected to a processing device and used for data processing; a backup file device for backing up data stored in the file device; When data is updated on the file device, the processing device is provided with a journal buffer for storing the data before the update and a journal directory indicating that the update has been performed, and the processing device is configured to update the data on the file device during the backup processing period. When data is updated, the data before the update is stored in the journal buffer, and
A file device backup process characterized in that when data in an updated area is stored in a journal directory and the data is backed up from the file device to a backup file device, the updated area is backed up from a journal buffer. method.