JPH05257617A - Magnetic disk subsystem - Google Patents

Abstract

PURPOSE:To secure the safety and reliability of data by speeding up a data recovery processing in case of occurring of frequent read errors and switching a magnetic disk before a fatal error occurs to a magnetic disk device. CONSTITUTION:This subsystem is provided with a means which reads data in the track of the magnetic disk on a magnetic disk device 12 in a data buffer 1 while there is no access from a host device and checks the data, a recovery means judging section 7 which judges which of the off-track read process, defective substituting track linking process, and home-address reformatting process of respective recovery means is applied to an error-occurring track at the time of frequent data error occurrence, an error information holding part 9 which logs error information on the error-occurring track and reports it to the host device, and a medium defective state reporting part 10; and the read to the magnetic disk device 12 is automatically checked and the magnetic disk subsystem judges one of the data recovery processes by itself from the tendency of the error occurrence and makes it applicable.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing system of a magnetic disk subsystem in an electronic computer system,
Especially, it relates to a data maintenance method.

[0002]

2. Description of the Related Art In a conventional magnetic disk subsystem,
If a read error occurs frequently when reading data from the magnetic disk device, if the magnetic head is misaligned, etc., the software saves the data on the magnetic disk to a magnetic tape, etc. for recovery, and then a read error occurs. The reformatting process of reformatting the magnetic disk in which the error occurred and restoring the data is being performed.

If the cause is a scratch on the medium surface of the magnetic disk, the software similarly saves the data of the track in which the read error has occurred to an alternative track, and then makes the track defective. Processing for linking defective replacement is being performed.

Further, in order to determine whether or not the cause of the read error is due to the displacement of the magnetic head, an off-track read process is performed in which the position of the magnetic head is deviated by a specified width from the center of the track for reading according to a software instruction. Is being done.

Which of the reformatting process, the link process for defective replacement, and the off-track read process to be applied to the medium is determined by the error information logged when the read error occurs and the number of I / Os from the host device. The maintenance personnel make judgment based on the error occurrence frequency, etc.

[0006]

In the above-mentioned conventional magnetic disk subsystem, when many read errors occur,
A maintenance staff checks the error log and the like due to the occurrence of a read error, estimates the cause of the error, and determines what kind of recovery means should be applied to the magnetic disk subsystem.

For example, when a read error occurs only at a specific head or a higher-order head address, a maintenance person determines that the cause is a deviation of the magnetic head or the like, and when a job is input from software, the track is recorded. The home address is being reformatted. Also,
When a read error occurs at a specific same head address of several consecutive cylinders, the maintenance personnel determines that it is caused by scratches, etc. After the retreat, the relevant track is regarded as a defective track, and a defective substitute link process is performed. If the cylinder or track in which the error occurred has no characteristics, an off-track read is performed in which the magnetic head position is read by shifting the position of the magnetic head from the center of the track by a specified width in order to distinguish the abnormality of the medium. It is processing.

However, it is said that the judgment of the trigger for applying the recovery means such as checking the error log is complicated, and the appropriate processing depending on the tendency of error occurrence may not be applied due to the technical ability of the maintenance staff. There is a problem.

Also, after the host device accesses the track in question and a read error occurs, the maintenance staff applies recovery means by referring to the error log and the like until the recovery processing of the track is completed. There is a problem that it takes time.

Particularly, regarding the reformatting process, since the data is saved to the magnetic tape or the like during the reformatting process, it takes more time than other recovery means.

Further, when the maintenance interval is long, the above-mentioned recovery means cannot be applied in a timely manner, and there is a problem that read errors frequently occur during that time.

Further, when the recovery means has not repaired, the defect of the disk can be judged for the first time by accessing the track from the host device after the recovery processing is applied and causing a read error. Therefore, there is a problem that it is not possible to determine a disk defect before access from the host device, and access from the host device after the recovery process is wasted.

An object of the present invention is to read the data of the track on the magnetic disk of the magnetic disk device 12 into the data buffer 1 while there is no access from the host device, to perform a data check, and when a data error occurs frequently. ,
A recovery means determination unit 7 for determining which of the recovery means of off-track read processing, defective alternative track link processing, and home address reformatting processing is applied to the error-occurring track, and error information of the error-occurring track are logged. However, it has an error information holding unit 9 and a medium defect status reporting unit 10 that are means for reporting to a higher-level device,
The magnetic disk device 12 is automatically read-checked, and the magnetic disk subsystem independently determines the data recovery processing according to the error occurrence tendency, and by applying the data recovery processing, the above-mentioned drawbacks are eliminated and recovery processing is performed. It is an object of the present invention to provide a magnetic disk subsystem capable of speeding up and exchanging a magnetic disk before a fatal error occurs in the magnetic disk device and ensuring the safety and reliability of data.

[0014]

A magnetic disk subsystem according to the first aspect of the present invention is a data buffer for storing data of tracks on a magnetic disk of a magnetic disk device for writing / reading data, and a data buffer. Read control means for controlling the data read operation to the magnetic disk, check correction means for checking and correcting the error of the data read from the magnetic disk by the read control means, and the magnetic head of the magnetic head during the read operation of the data from the magnetic disk. Off-track read means for reading data by shifting the position from the center of the track, reformatting means for reformatting the home address of the designated track, and check correction means save data on the track with many data errors to the alternative track Tracks with lots of bad tracks Error information including a recovery means including at least one of link means for performing defective replacement, a cylinder number of a track in which an error has occurred by the check correction means, a head number, an error occurrence frequency, and an error history of the applied recovery means. And holding means for holding the error information, and analysis and reporting means for analyzing the content of the error generated from the error information held by the holding means and reporting the analysis result and the error information to the host device.

The magnetic disk subsystem according to the second aspect of the present invention is the magnetic disk subsystem according to the first aspect of the present invention, in which the recovery means to be applied is selected according to the analysis result of the error content analyzed by the analysis reporting means, and applied. The magnetic disk subsystem according to the third aspect of the present invention is provided with a means for repairing an error by performing a data correction by a check correction means in the magnetic disk subsystem according to any one of the first aspect or the second aspect. When there are many errors, it is provided with means for reporting to the host device that the magnetic disk or the magnetic disk device needs to be replaced.

The magnetic disk subsystem according to the fourth invention is the magnetic disk subsystem according to any one of the first invention, the second invention or the third invention, wherein
It is provided with means for monitoring whether or not an access from a higher-level device arrives and, if it does not arrive within a specified time, providing an opportunity to perform recovery processing by the subsystem alone.

[0017]

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

FIG. 1 is a block diagram of a magnetic disk subsystem according to an embodiment of the present invention.

In FIG. 1, the magnetic disk subsystem of the present embodiment comprises a data buffer 1, an access status management unit 2, a buffer control unit 3, an off-track read control unit 4,
Defective alternative track link control unit 5, reformatting control unit 6, recovery means determination unit 7, data check correction circuit 8, error information holding unit 9, medium defect status reporting unit 10, magnetic disk control device 11, magnetic disk device 12, It consists of

Next, the operation of the magnetic disk subsystem of this embodiment will be described.

First, the access status management unit 2 has a timer for a specified time, and if the next access does not come within the specified time from the last access from the host device, the buffer control unit 3 is caused to perform a read operation. Give instructions. Upon receiving the read operation instruction, the buffer control unit 3 controls to read the data of the track T1 from the magnetic disk device 12 into the data buffer 1 and instructs the data check correction circuit 8 to check the data error. If no data error occurs, the data check correction circuit 8 reports to the buffer control unit 3 that there is no error. Upon receiving the report, the buffer control unit 3 controls the track T2, in which the track number is incremented by the specified pitch, to perform the read operation in the same manner to perform the data error check.

If a data error is detected during the track T2 read check, the data check correction circuit 8
Corrects the data in the data buffer 1 and
It reports to the error information holding unit 9 that an error has occurred in the track T2. The error information holding unit 9 uses the track T2
At the same time as incrementing the number of data error occurrences at, the cylinder number and the track number are held. Further, the error information holding unit 9 checks whether or not a data error has occurred a prescribed number of times or more. If the specified number of times has not been exceeded, the buffer controller 3 is again instructed to perform a read check on the same track T2. If the specified number of times is exceeded, the error information holding unit 9 transfers all the held error information to the recovery means determination unit 7. The recovery means determination unit 7 analyzes the received error information, selects the recovery means to be applied to the track T2, and applies the application to each recovery processing control unit (off-track read control unit 4, reformatting control unit 6, defective replacement). Instruct the track link control unit 5).

The processes from the determination of the application of each recovery process to the read check of the next track T3 are performed as follows.

For example, when the read error exceeds the specified number of times only for a specific track or a higher-order track number, the recovery means determination unit 7 determines that the magnetic head shift or the like is the cause, and reformatting control is performed. The unit 6 is instructed to reformat the track T2.
The reformatting control unit 6 that has received the instruction moves to the track T2.
After the reformatting of the home address of the above is performed on the magnetic disk device 12, the buffer controller 3 is instructed to control the rewriting of the corrected data in the data buffer 1 to the track T2. Under the control of the buffer control unit 3, the corrected data in the data buffer 1 is written in the track T2. After writing, the buffer control unit 3
Performs a read check again on the track T2, and if no error occurs, a flag indicating that the track T2 has been reformatted has been set in the error information holding unit 9 so that the access status management unit 2 sends it from the host device. Until the access is detected, the read check operation of the next track T3 incremented by the specified pitch is continuously performed. If an error occurs in the read check after the reformatting process, the error information holding unit 9 makes the track T2 a defective track and transfers the error information of the track T2 to the medium defect status reporting unit 10. The medium defect status reporting unit 10 checks whether the number of defective tracks exceeds the specified number, and if it exceeds the number, reports the medium defect of the magnetic disk device 12 to the upper device, and if there is a request from the upper device, All the error information stored in the error information storage unit 9 is transferred to the host device. If the number of defective tracks does not exceed the specified number, the counter of the number of defective tracks of the medium defect status reporting unit 10 is incremented and the buffer control unit 3 is instructed to check the next track T3.

Next, when a read error occurs at a specific same track address of several consecutive cylinders, the recovery means determination unit 7 causes the magnetic disk device 12 to operate.
The defective substitute track link control unit 5 is instructed to perform the defective substitute track link process on the track T2 on the assumption that the medium surface is damaged. Upon receipt of the instruction, the defective substitute track link control unit 5 confirms with respect to the magnetic disk device 12 whether or not an alternative track remains, and if not, the defective substitute track link control unit 5 uses the track T.
It reports to the error information holding unit 9 that 2 is a defective track. Upon receiving the report, the error information holding unit 9 sets the track T2 as a defective track, and transfers the error information of the defective track T2 to the medium defect status reporting unit 10. After the error information is transferred to the medium defect status reporting unit 10, the same processing as is performed. If the alternative track remains, the defective alternative track link controller 5 instructs the buffer controller 3 to save the corrected data in the data buffer 1 to the alternative track T2 ′, and then the track T2. The defective track is set, and the defective track and the alternative track are linked. After the link processing is completed, the buffer control unit 3 performs a read check on the track T2 again. If no error occurs, the buffer control unit 3 stores the track T2 in the error information holding unit 9.
A flag indicating that the defective alternative track link processing has been completed is set, and the read check operation of the next track T3 incremented by the specified pitch is continuously performed until the access status management unit 2 detects an access from the host device. If an error occurs in the read check after the link processing, the error information holding unit 9 regards the track T2 as a defective track, and the medium defect status reporting unit 10 notifies the track T2.
Transfer error information of. After the error information is transferred to the medium defect status reporting unit 10, the same processing as is performed. Also,
The defective alternative track link processing may be applied to the repeated alternative track T2 ′.

When the cylinder or track in which the error has occurred has no characteristic, the recovery means determination unit 7 causes the off-track read control unit 4 to track the position of the magnetic head in order to separate the abnormality of the medium. It is instructed to perform the off-track read processing in which the lead is shifted from the center of the track by the specified width. Upon receiving the instruction, the off-track read control unit 4 applies the off-track read process to the magnetic disk device 12, and again controls the buffer control unit 3 to instruct the track T2 to perform a read check. If no error occurs after the read check, the off-track read control unit 4 instructs the reformatting control unit 6 to apply the reformatting process to the track T2, and causes the error information holding unit 9 to track T.
The flag 2 indicates that the off-track read processing has been completed. This is the same as after applying the reformatting process. If an error occurs after the read check, a flag indicating that the track T2 has been subjected to off-track read processing is set in the error information holding unit 9 and at the same time, the track T2 is regarded as a defective track and the medium defect status reporting unit 10 is tracked. Transfer the error information of T2. Medium defect status reporting unit 10
After the error information is transferred to, the same processing as is performed.

As described above, the magnetic disk device 1 is automatically used while the access from the host device does not come.
A read check is performed on No. 2, and the data recovery processing is determined by the magnetic disk subsystem alone according to the error occurrence tendency, and it can be applied.

[0028]

As described above, the magnetic disk subsystem of the present invention performs a read operation on a disk by submitting a software job, and if a read error occurs frequently, the magnetic disk subsystem alone causes an error. By analyzing the contents and reporting the analysis result and error information to the host device, maintenance personnel can easily grasp the state of the medium and can immediately select the recovery means to be applied. ..

In addition, the magnetic disk subsystem alone,
By selecting and applying an appropriate recovery means according to the analysis result of the error content, there is an effect that an appropriate recovery process can be performed in a short time regardless of the technical ability of the maintenance personnel.

When the number of defective tracks exceeds the specified number, it is necessary to replace the medium together with the cylinder number of the track in which the error occurred, the track number, the error occurrence frequency, and the error history information of the recovery means applied. By making it possible to report to the host device, it is possible to almost replace the disk before a fatal error occurs at the time of access from the host device, and it is possible to ensure the data safety.

Further, by providing means for managing the access status from the host device, it is not necessary to submit a job from software, and the free time of the magnetic disk subsystem that is not accessed by the host device is used, The magnetic disk subsystem alone automatically prevents the read error from occurring when a read command is issued from the host device by automatically performing the read operation, error cause determination, recovery means selection and application. Can
The recovery process does not depend on the maintenance interval and can be performed promptly in a timely manner, which further improves the reliability of the data and has the effect of eliminating wasteful read command retry processing due to a read error.

[Brief description of drawings]

FIG. 1 is a block diagram of a magnetic disk subsystem according to an embodiment of the present invention.

[Explanation of symbols]

 1 data buffer 2 access status management unit 3 buffer control unit 4 off-track read control unit 5 defective alternative track link control unit 6 reformatting control unit 7 recovery means determination unit 8 data check correction circuit 9 error information holding unit 10 medium defect status report Part 11 Magnetic disk controller 12 Magnetic disk device

Claims (4)

[Claims] 1. A data buffer for storing data of a track on a magnetic disk of a magnetic disk device for writing / reading data, and a read control means for controlling an operation of reading the data into the data buffer. Checking / correcting means for checking and correcting an error in the data read from the magnetic disk by the read control means, and the position of the magnetic head is shifted from the center of the track during the operation of reading the data from the magnetic disk to read the data. The off-track read means for reading out, the reformatting means for reformatting the home address of the designated track, and the check-correction means save the data of the track with a large number of data errors to the alternative track, and then make the track with a lot of data errors a defective track. Alternate link Recovery means including at least one of the link means, and holding means for holding, as error information, the cylinder number of the track in which an error has occurred by the check correction means, the head number, the error occurrence frequency, and the error history of the applied recovery means. And a analysis report unit for analyzing the content of an error generated from the error information held by the holding unit and reporting the analysis result and the error information to a higher-level device. 2. The magnetic disk subsystem according to claim 1, further comprising means for selecting a recovery means to be applied according to an analysis result of the error content analyzed by the analysis reporting means, and applying the recovery means to recover the error. A magnetic disk subsystem comprising: 3. The magnetic disk subsystem according to claim 1, wherein when the check correction means has a large number of data errors, the magnetic disk or the magnetic disk device needs to be replaced. Is provided to the host device, and a magnetic disk subsystem is provided. 4. Claim 1 or claim 2 or claim 3.
In any one of the magnetic disk subsystems described, it is provided with means for monitoring whether or not an access from the higher-level device is received, and providing an opportunity to perform recovery processing by the subsystem alone if the access is not made within a specified time. A magnetic disk subsystem characterized by the above.