JPH05108272A - File controller device - Google Patents

Abstract

PURPOSE:To improve the reliability by wiring all of data, which is received by an OS(operating system) and should be written into a disk file, into the disk file at the time of detecting abnormality by abnormality detecting means and writing data in the synchronous write system after occurrence of the abnormal state. CONSTITUTION:In the steady state, the OS writes data in the disk file of a hard disk 6 by the delay write system when a data write command is generated by an application. If a power source device 2 free from service interruption, detects a service interruption, it is reported to a serial port 1a of a device main body 1, and as the result, data marked by delay write on a free block is written in the disk film. while the abnormality continues thereafter, data is written in the disk file by the synchronous write system at the time of generation of the data write designation from the application. Thus, data write is preferred to improve the reliability.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file controller device for writing data to a disk file.

[0002]

2. Description of the Related Art As a method of writing data in a disk file, a delayed writing method, a synchronous writing method, and an asynchronous writing method are known. In the delayed write method, even if the application issues a data write command to the disk file, the data is only written to the memory in the operating system, and the operating system immediately responds to the application that the operation is complete, and then, if necessary. The operating system is now writing the data to a disk file. In the synchronous write method, when an application issues a data write command to a disk file, the operating system writes the data to the disk file.
When the writing is completed, the operating system responds to the application that the operation is completed. Also, in the asynchronous write method, when an application issues a data write command to a disk file, the operating system starts writing data to the disk file, but the operation completion response is sent to the application without waiting for the completion of the write. Is becoming

Therefore, the delayed write method has an advantage that the processing speed is fast from the viewpoint of the application, but has a problem that it is unknown when the data will be written. Further, the synchronous writing method can surely confirm that the data has been written from the viewpoint of the application, but has a problem that the processing speed is slow.

[0004]

Therefore, when a file controller device that writes data to a disk file by a delayed write method with priority given to processing speed is used, if an abnormal state such as a power supply error occurs, the system is then restarted. There was a problem that even if you tried to recover it, data that was not written to the disk file occurred and the system could not be recovered normally.

Therefore, in the present invention, data is written to a disk file by a delayed write method with priority given to processing speed, and the data is surely written to the disk file even if an abnormal state such as a power failure occurs. It is an object of the present invention to provide a file controller device capable of improving reliability by prioritizing data writing to a disk file by a synchronous writing method after an abnormal state occurs.

[0006]

According to the present invention, an application issues a command to write data to a disk file, the operating system receives the command and responds to the application that the operation has been completed, and then the operating system sends the command to the disk as needed. In a file controller device that performs file control by a delayed writing method that writes data to a file, an abnormality detection unit that detects an abnormality such as a power supply abnormality, and when this abnormality detection unit detects an abnormality, writes it to a disk file received by the operating system. When data is written by the abnormal data writing means that writes all the data to be written to the disk file and this abnormal data writing means is executed, the data from the application to the operating system will be transferred to the disk file. The write command in which the operating system has provided a writing method switching means for switching the file control in synchronous write method that does not respond to operation completion to the application if it does not complete the writing of data to a disk file.

[0007]

In the present invention having such a configuration, when the abnormality detecting means detects an abnormality such as a power supply abnormality, all data to be written in the disk file received by the operating system is written in the disk file. After that, when the application issues a command to write data to the disk file to the operating system, the operating system immediately writes the data to the disk file, and when the writing of the data is completed, the operating system responds to the operation completion.

[0008]

Embodiments of the present invention will be described below with reference to the drawings.

In FIG. 1, reference numeral 1 is an apparatus main body, and 2 is an uninterruptible power supply for supplying electric power to the apparatus main body. The device main body 1 is provided with a serial port 1a, and when a power failure occurs, the uninterruptible power supply 2 notifies the serial port 1a of the occurrence of a power failure. That is, the uninterruptible power supply 2 constitutes an abnormality detecting means.

As shown in FIG. 2, the apparatus main body 1 is provided with a microprocessor 3, and a memory 5, a hard disk drive 6 having a disk file provided on the microprocessor 3 via a bus line 4, a display device 7, an input device. 8,
An RS-232C interface 9 or the like having the serial port 1a is connected. The memory 5 is provided with a write control flag 5a.

In the apparatus main body 1, when an application issues a data write command, the operating system receives the data write command and writes the data to the disk file of the hard disk device 6 by the delayed write method.

The delay writing method will be described here. FIG. 5 shows the data on the memory 5 managed by the operating system. Each number is the record number of the disk, and the actual data for the record number of which the remainder of record number / 4 is 0, and the remainder of record number / 4 is 1 Actual data for record number, record number ÷ 4, actual data for record number with remainder 2 is record number / 4
The actual data for the record numbers with the remainder of 3 are stored in the memory 5, respectively. Each data block is connected in a chain.

The record number is as follows: "(a) The data is not completely aligned while the disk is currently being read or written. (B) In the past, the data having the record number is transferred from the disk to the memory. Read, the contents of the disk are the same as the contents of the record in the memory, but this block in the memory may be used to read or write another record number. (C) Data of that record number in the past Was instructed to write to the disk.
The contents of the disk and the contents of the record in memory are different, and the data in memory is correct. Once the data in the memory has been written to the disk, this block may be used to read or write another record number. There are different types.

Of these data blocks, (b) and (c) are connected by a chain of free blocks,
In addition, as for (c), it is displayed with a delayed write mark * as a delayed write block.

When an application issues a write command to the disk, it is converted into a corresponding record number,
The application is terminated normally by attaching the delayed write mark at the end of the free block chain together with the actual data. That is, the processing is completed in a short processing time without actually writing data. Further, the data block with the delayed write mark * is written by the operating system about every 30 seconds. If the system power is turned off within these 30 seconds, the delayed write mark *
Data marked with will be lost.

Therefore, when the microprocessor 3 receives a notification of the occurrence of a power failure from the serial port 1a, it performs the interrupt processing shown in FIG. This first adds 1 to the write control flag 5a.

Then, it is checked whether the value of the write control flag 5a is 6 or more. Write control flag 5a
If the value of is 5 or less, the data with the delayed write mark * on the free block is written to the disk file. (Abnormal data writing means) Since there is a block in the transition period which is currently in use, the schedule is determined so that this routine will operate again after 5 seconds.

The microprocessor 3 checks the write control flag 5a in the write processing routine as shown in FIG. 4, and the write control flag 5a is turned on.
That is, when it is 1 or more, the synchronous write processing is performed, and when the write control flag 5a is OFF, that is, when the record is 0 and the record is not the synchronous write instruction, the delayed write processing is performed.
(Writing method switching means) Even if the writing control flag 5a is 0, if the record is a synchronous writing instruction, synchronous writing processing is performed.

In the embodiment having such a configuration, the value of the write control flag 5a is "0" when the uninterruptible power supply 2 does not detect a power failure such as a power failure.
At this time, the data writing to the disk file is performed by the delay writing method. That is, when the application issues a data write command to the disk file, the operating system receives the command and writes the data in the memory in the operating system, and immediately responds to the application that the operation is completed. The data written to memory is then written to the disk file by the operating system as needed.

As described above, since the operation completion response is immediately issued when the application issues a data write command to the disk file, the processing speed seen from the application is fast.

In the delay writing method, the memory 5
As the above data, when the disk is currently being read or written, the data is not completely complete yet, or in the past, the data of that record number was read from the disk into the memory 5, and the contents of the disk and the memory The contents of the record on the memory 5 are the same, or there is an instruction to write the data of that record number to the disk in the past. The contents of the disk and the contents of the record on the memory 5 are different. The above data produces the correct situation.

When the uninterruptible power supply 2 detects a power failure while writing data to a disk file by such a delayed writing method, a power failure occurs from the uninterruptible power supply 2 to the serial port 1a of the apparatus main body 1. Be notified. As a result, "1" is added to the write control flag 5a.

When the write control flag 5a becomes "1", the data with the delayed write mark * on the free block is written in the disk file. As a result, the disk file data is rewritten to the correct data in the memory 5.

The interrupt processing is scheduled to be executed again after 5 seconds. This interrupt process is executed again after 5 seconds, and this interrupt process is repeated 5 times. Then, when the write control flag 5a becomes "6", the repetition of the interrupt process ends. Since the interrupt process is repeated 5 times in this way, even if the data with the delayed write mark * has a block in the transition period during data alignment, the interrupt process can surely write to the disk file.

When the write control flag 5a becomes 1 or more, after that, when a data write command is issued by the application, the data write to the disk file is performed by the synchronous write method. Therefore, thereafter, when the application issues a data write command to the disk file, the operating system writes the data to the disk file, and when the writing is completed, the operating system responds to the application that the operation is completed.

Although the response of the operation completion to the application is delayed by this, the reliability is high because the application can surely know that the data is written in the disk file.

In the above embodiment, the disk file is provided in the hard disk device, but the invention is not limited to this, and the disk file may be provided in the floppy disk device.

[0028]

As described above in detail, according to the present invention, when data is written to a disk file by a delayed write method with priority given to processing speed, even if an abnormal state such as a power failure occurs. It is possible to provide a file controller device capable of surely writing data to a disk file and further improving reliability by prioritizing data writing to the disk file by a synchronous writing method after an abnormal state occurs. ..

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a device body in the embodiment.

FIG. 3 is a flowchart showing interrupt processing of the microprocessor in the embodiment.

FIG. 4 is a flowchart showing a data writing process of the microprocessor in the embodiment.

FIG. 5 is a diagram showing an example of a data configuration on a memory in the embodiment.

[Explanation of symbols]

1 ... Device body, 2 ... Uninterruptible power supply device, 3 ... Microprocessor, 5 ... Memory, 5a ... Write control flag, 6 ... Hard disk device.

Claims (1)

[Claims] 1. An application issues a command to write data to a disk file, the operating system receives this command and responds to the application that operation has been completed, and then the operating system writes data to the disk file as necessary. In a file controller device that performs file control by a delayed write method, an abnormality detecting unit that detects an abnormality such as a power supply abnormality, and when the abnormality detecting unit detects an abnormality, all the data to be written to the disk file received by the operating system is described above. When the abnormal data writing means for writing to the disk file and the data writing by the abnormal data writing means are performed, the application writes a data write command to the disk file thereafter to the operating system. The file controller device further comprises a writing method switching means for switching the file control to a synchronous writing method in which the operating system does not respond the operation completion to the application unless the operating system finishes writing the data to the disk file.