JP3139635B2 - Duplex file control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling a duplicated file in a disk subsystem, and more particularly to a method capable of operating at high speed in response to a write / read command from a host computer and having a failure in track units. And a dual file control method capable of managing double files.

[0002]

2. Description of the Related Art Conventionally, a method of improving the reliability of a disk drive by writing the same data to two disk drives in response to a write instruction from a host computer is often used. Are known. As this double writing method, two methods are used, one of which is described in, for example, JP-A-55-3475.
As in the method described in Japanese Patent Application Laid-Open No. 6-206, there is a method in which data is simultaneously written to two disk drives in synchronization with each other. Then, the data is written to one of the disk drives and the cache memory built in the disk controller, and the writing to the other disk drive is performed asynchronously later as a stage process from the cache memory. Is the way.

[0003]

In the conventional dual write file control method, if a failure occurs in one of the disk drives or the cache memory and the contents of the two disk drives cannot be matched with each other. However, the disk drive that cannot be updated due to a failure or the like was disconnected, and the file was forced to be reduced to a single file.
As a result, in the case of a track failure of the disk drive or a failure of the cache memory, the duplication function is stopped for all data even though the data which cannot be duplicated is a part of the entire data. Duplicate files are gone. If the fault range is only one track, or a fault in only a partial area such as several tracks, or a fault in the cache memory, the other tracks and the other disk drive can operate normally as a matter of course. Therefore, if only the faulty part is not used, duplication is possible, and it is desirable from the viewpoint of reliability. SUMMARY OF THE INVENTION An object of the present invention is to solve such a conventional problem, and even when a failure occurs in a disk subsystem, minimize the influence of the failure and write double data for data that can maintain redundancy. It is an object of the present invention to provide a dual file control method capable of continuously increasing the availability of dual writing and improving reliability.

[0004]

To achieve the above object, according to an aspect of the duplexing file control method of the present invention, when the host computer has a write instruction, the disk control device 2 disk drives designated as spare drives writing the same data, when there is a read instruction from the host computer, the duplicated file control method for a disk subsystem for reading data from one of the spare drives, to the disk control unit, the spare drives each stores the accessibility of the track unit for, if any one of the arbitrary track of the pair drive is inaccessible, 該De disk controller corresponds to the normally accessible the trucks Only the disk drive that has Performing read and write operations of data, performs unification degeneracy of track units, for accessible track both the spare drives, have rows duplicated control, the above-mentioned disk controller further key
Cache memory and write from host computer
When the drive is instructed to
Data to the disk drive and the cache memory.
Write, write to the other disk drive
Later, asynchronously from the cache memory
File control method to perform
The disk control unit is used for either one of the paired drives.
The track that is stored as inaccessible for the disk drive
When a write command is issued to the
For the corresponding track and cache memory of the live
After writing data, it is memorized as above
When a track is repaired,
Change inaccessible to accessible and change the above cache memory
Is written to the track of the disk drive.
Performs destage processing and resumes redundant control
It is characterized by doing.

[0005]

According to the present invention, in the disk controller, 2
Accessible for each track of each disk drive,
A memory map for managing impossibility is provided. When a write / read command is input from the host computer, writing / reading is performed on an accessible track by referring to this memory map to hold normal data. When a failure occurs, access is set to a track in which the latest data cannot be written in any of the disk drives due to its influence. As a result, only the tracks that cannot be updated by any one of the disk drives are reduced to a single file, and the other tracks continue to be duplicated. Thereafter, in order to resume the duplicated file in the entire drive again, only the file including the track degraded to the singlexedness is subjected to the duplication recovery process by the file copy. In the meantime, the duplication of the other files can be continued, so that a highly reliable and highly available dual writing file control method can be realized.

[0006]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram of a disk subsystem showing an embodiment of the present invention. The disk subsystem according to the present embodiment includes a host computer 1, a disk controller 2, and disk drives 31 and 32. The disk control device 2 has two routes as interfaces with the host computer 1, and the control processors 21 and 22 are in charge of the respective routes. Each of the control processors 21 and 22 has a disk drive 3
Both 1 and 32 can be accessed. further,
The disk controller 2 is provided with a common control memory 23 shared by the control processors 21 and 22. The control processors 21 and 22 respond to a write / read request from the host computer 1 by using the common control memory 23. Control is performed by referring to the contents. Common control memory 23
Stored therein are information on a drive access permission flag 231 and a track unit access permission map 232 for each disk drive. Drive access permission flag 2
A flag 31 indicates whether or not the corresponding drive is normally accessible. When ON, the flag indicates that access is possible, and when OFF, it indicates that access is not possible. The track unit accessibility map 232 is assigned one bit for each track in the drive, and indicates whether each track is normally accessible. In this case, when ON, access is permitted, and when OFF, access is disabled. First, all information in the common control memory 23 is initially set to ON so that all tracks can be accessed.

FIG. 2 is a processing flowchart of the control processor in FIG. 1, and shows a processing procedure when a failure occurs in a disk drive. First, the cause of the obstacle is identified. That is, it is determined whether the failure is related to the entire drive or only a specific track (step 40). If a failure has occurred in the entire drive, the access permission flag 231 of the drive is reset (step 41), and the host computer 1 is notified that the dual writing function has been stopped in the entire drive (step 42). ). If the track failure is found as a result of the identification, the bit corresponding to the track in the track unit access permission / inhibition map 232 is reset (step 43), and the host computer is reset.
Then, it reports to the data 1 that the track has been unified and reduced (step 44).

FIG. 3 is a processing flowchart of the disk controller when a read instruction for a certain track is received from the host computer. First, disk drive #
1 (31) drive access permission flag 231 is ON
Is determined (step 50). If it is ON, it is next determined whether or not the bit corresponding to the read instruction track of the track unit access permission map 232 of the disk drive # 1 is ON (step 51).
Next, with respect to the disk drive # 2 (32), the bit corresponding to the drive access permission flag 231 and the read instruction track of the track unit access permission map 232 is determined (steps 52 to 55). The above four bits (that is, the disk drive access permission flag 23
1 is 2 bits, and 2 bits of 2 bits of the designated track corresponding bit of the track unit access permission / inhibition map 232 are 4 bits in total. Data is read from either one of them (step 56). If both bits are ON for the disk drive # 1, and if any bit is OFF for the disk drive # 2, data is read from the normal disk drive # 1 (step 57). ). Further, conversely, for disk drive # 1, any bit is OFF
If so, data is read from the normal disk drive # 2 (step 58). If either bit is OFF in both drives, the read of that track has failed, and the read failure is reported to the host computer 1 (step 59).

FIG. 4 is a processing flowchart of the disk controller when a write instruction for a certain track is received from the host computer. First, in the same manner as when the read instruction is received, the bit corresponding to the write-instructed track of the drive access permission flag 231 and the track unit access permission map 232 is determined for each of the disk drives # 1 and # 2 ( Step 60-
65). If all of the above four bits are ON, data is written to both disk drives # 1 and # 2 (step 66). In addition, both bits are ON for the disk drive # 1, and the disk drive # 1 is ON.
Regarding 2, if any bit is OFF, data is written only to the disk drive # 1 (step 67). And vice versa, disk drive #
If any one of the bits is OFF with respect to 1, data is written only to the disk drive # 2 (step 68). If either bit is OFF for both drives, the writing of that track has failed, and the writing failure is reported to the host computer 1.
(Step 69). By such control, when a track failure occurs, only that track is excluded from the target of the double-writing file, but the duplication can be continued for the other tracks.

[0010] Next, the flow of the file recovery processing by the host computer when a certain track is unified and reduced.
It is a chart. When the host computer 1 receives a report from the disk controller 2 that a certain track has been reduced to a single track, the host computer 1 executes the next process in the middle of the job based on the order or the plan for executing each program. First,
The job related to the file operation including the unified degeneration track is interrupted (step 70). Next, a replacement track in place of the single degenerate track is secured in the pair drives # 1 and # 2 (step 71). That is, since only one disk drive has a faulty track, the replacement track may be only the disk drive having the faulty track. However, when an access instruction is issued from the host computer 1, a redundant function is achieved. Since it is necessary to write data on the same track, the same track is secured in the pair drives # 1 and # 2. If the replacement track has been secured, the data of the unified degeneration track is read (step 72). In this read processing, as shown in the flow of FIG. 3, data is read from a drive having a normally accessible track (step 5).
7 or 58). Next, the read data is written to a previously secured replacement track (step 7).
3). In this writing process, as shown in FIG.
Data is written to both drives # 1 and # 2 (step 66). Finally, the interrupted job is restarted (step 74). By this processing, the file recovery is completed, and the duplex control for the entire drive is restarted again. In the above embodiment, the host computer
Although the file recovery processing by the data controller 1 has been described, the processing procedure shown in FIG. 5 may be spontaneously performed by the disk controller 2. That is, when a certain track is unified and degenerated, the disk control device 2 notifies the host computer 1 of the occurrence of the unified and degenerated track to the paired disk drives # 1 and # 2. Is secured as a replacement track, and the data of the faulty track is duplicated by the replacement track according to the procedure shown in FIG. Then, it reports to the host computer 1 that the unification and degeneration have been resolved. After that, reading to the failed track,
In response to the write instruction, the track address needs to be read, and the disk control device 2 automatically accesses the alternate track.

Next, the operation in the case of a disk control device having a cache memory will be described. FIG. 6 is a block diagram of a disk subsystem showing another embodiment of the present invention. The overall configuration is the same as the configuration in FIG.
The difference is that a cache memory 24 is provided inside the disk control device 2 and a track unit non-reflection map 233 is provided in the common control memory 23 for each disk drive. The cache memory 24 stores write data from the host computer 1 and data read from the disk drives 31 and 32 as necessary. Track unit unreflected map 23
Reference numeral 3 indicates that one bit is assigned to each track in the drives 31 and 32, and indicates whether or not the latest data of each track is reflected in the disk drive.
In this case, it indicates that ON is not reflected, and OFF
Indicates that has been reflected. This information is all initially set to OFF. That is, initially, it is assumed that all information has already reflected the contents of the cache memory.

Next, the processing of the control processors 21 and 22 in the disk control device 2 will be described. FIG. 7 is a processing flowchart when an instruction to read a certain track is received from the host computer. First, the cache memory 2
4 is determined to be normal (step 80), and if so, it is determined whether or not the data of the read instruction track exists in the cache memory 24 (step 81). If the data exists in the cache memory 24, the data is read from the cache memory 24, transferred to the host computer 1, and the process is terminated (step 82). On the other hand, when the cache memory 24 is unusable, or when the data
If the data does not exist in the cache memory 24, the bit corresponding to the read instruction track of the drive access permission flag 231 and the track unit access permission map 232 is determined for each of the disk drives # 1 and # 2 ( Steps 83 to 85). For each of the disk drives # 1 and # 2, the bit corresponding to the drive access permission flag 231 and the read instruction track of the track unit access permission map 232 is ON, and corresponds to the read instruction track of the track unit non-reflection map 233. If this bit is OFF, it indicates that the data of the designated track can be normally read from the disk drive. As described above, if both of the disk drives # 1 and # 2 satisfy the above-mentioned condition, that is, the condition that all the tracks have no failure and the contents of the cache memory 24 are reflected on all the tracks, one of the conditions is satisfied. Data is read from one drive (step 86). If only one of the disk drives satisfies the above condition, data is read from the drive that satisfies the condition (step 8).
7,88). If neither of the disk drives satisfies the above condition, the reading of that track has failed, so that the reading failure is reported to the host computer 1 (step 89).

Next, a processing procedure when a write instruction for a certain track is received from the host computer 1 will be described.
FIG. 8 is a processing flowchart of the control processor for a write instruction when the cache memory is normal. First, for the disk drive # 1, the bit corresponding to the write-instruction track of the drive access flag 231 and the track unit access map 232 is determined (step 90). If both of the above two bits are ON, data is written to the cache memory 24 and the disk drive # 1 (step 9).
1). Next, regarding the write instruction track corresponding bit of the track unit non-reflection map 233, the disk drive # 1 is turned off, and the disk drive # 2 is turned off.
N is set (steps 92 and 93). This indicates that the latest data exists only in the disk drive # 1. As a result of determining the drive access permission flag 231 and the bit corresponding to the write instruction track of the track unit access permission map 232 for the disk drive # 1 (step 90), one of the bits is OFF.
If it is determined that writing to the disk drive # 1 cannot be performed, then the disk drive # 2 corresponds to the drive access permission flag 231 and the write-instructed track of the track unit access permission map 232. The bit is determined (step 94). If both bits are ON, data is written to cache memory 24 and disk drive # 2 (step 9).
5). Then, regarding the bits corresponding to the write-instructed track of the track unit non-reflection map 233, the disk drive # 2 is turned off and the disk drive # 1 is turned on (steps 96 and 97). If it is determined in step 94 that one of the bits is OFF and writing to the disk drive # 2 cannot be performed, it is determined that writing of the track has failed, and the host computer determines that writing has failed. 1 (step 98). Here, the processing for the write instruction when the cache is normal has been described. However, when the cache is unusable, data is not written to the cache in steps 91 and 95 except that it is not performed. Then, the same processing as in FIG. 8 is performed. When a write instruction is received from the host computer 1, a write process for a disk drive in which data has not been written is performed as a stage process.

FIG. 9 is a flowchart of the stage processing according to the present invention. The destaging process is executed by the disk controller 2 only when the cache memory 24 is normal and at a time when an access instruction from the host computer 1 does not come. Here, the cache memory 2
4 cannot be used due to a failure or the like, so that data is written only to disk drive # 2, and the same data has not yet been written to drive # 1. Since the cache memory 24 has been recovered, the control processors 21 and 22 start the destaging process to continue the duplexing control. First, the drive access permission flag 231 of the disk drive # 1 is determined (step 1000). Flag 231 is OFF
If, the destage processing ends. Flag 2
When 31 is ON, one track whose corresponding bit of the track unit non-reflection map 233 is ON is selected next (step 1010). If there is no corresponding track, the destage processing ends. If one track is selected, the bit corresponding to the selected track in the track unit access permission / inhibition map 232 is determined (step 1020). If the bit is OFF, the track cannot be accessed, so that the process returns to the original process and returns to the process of selecting another track (step 101).
0). If the bit is ON, it is next determined whether or not the data of the selected track exists in the cache (step 1030). At that time, when the cache has failed or has just recovered from the failure, the data of the selected track is not present in the cache, and the data is normally stored in the disk drive # 2. Determine if it exists. That is, it is determined whether the drive access permission flag 231 of the disk drive # 2 and the selected track corresponding bit of the track unit access permission map 232 are ON and the selected track corresponding bit of the track unit non-reflection map 233 is OFF. (Step 1040). When the above condition is satisfied, that is, the corresponding bit of the flag 231 and the map 232 is ON, and the corresponding bit of the unreflected map 233 is O.
In the case of FF, the data of the selected track is read from the disk drive # 2 onto the cache (step 105).
0). If the cache fails during the destaging process, the destaging process is stopped at that point.
If the above condition is not satisfied, since the latest data of the selected track does not normally exist anywhere, the duplication of the track is abandoned, and the process proceeds to another track selection processing (step 1010).

If the data of the selected track exists in the cache or is normally read from the disk drive # 2 into the cache, the data is then transferred from the cache to the disk drive # 1. Write (step 1060), and then turn off the selected track corresponding bit of the track unit non-reflection map 233 of the disk drive # 1 (step 1070). The above is the procedure of the stage processing in the disk drive # 1. The same applies to the stage processing of the disk drive # 2. Thus, the cache memory 2
In the duplex control of the disk control device 2 provided with
When there is no failure in the cache memory 24, when there is a write instruction from the host computer 1, data is written to the disk drive # 1 and the cache memory 24, and thereafter, the disk is read from the cache memory 24 as a stage processing. This is realized by writing data to drive # 2. Further, when a failure occurs in the cache memory 24, the host computer reports that the unified degeneration of a track in which the track unit non-reflection map 233 of one of the disk drives # 1 and # 2 is ON. Report to 1. When the cache memory 24 is restored, the disk control device 2 automatically resumes the destaging process, and normalizes the data of the track that has been unified and reduced according to the procedure shown in FIG. Is copied from the disk drive to which the data has been written to the disk drive to which the data has not been reflected. Thereby, the duplex control of the track can be restarted.

Also, if a failure occurs in the disk drive,
In the state where the tracks are unified and degenerated, as shown in FIG. 8, data is written to the disk drive which can be accessed normally and, at the same time, data is written to the cache memory 24. Thereafter, when the disk drive is restored, the access permission flag 231 of the disk drive is restored.
And all the track unit access availability maps 232 are set to ON. As a result, for a single degenerated track,
As shown in FIG. 9, since the destaging process is performed from the cache memory 24 to the unreflected drive, the duplex control of the track can be resumed. In the embodiment,
Even when the cache memory 24 has a failure, only the tracks that have not been duplicated on the disk drive at the time of the failure are unified and degenerated, so that the range affected by the failure can be minimized. Further, when the faulty part is repaired, the single degenerate track can be automatically restored to the double track, so that the maintainability is excellent.

[0017]

As described above, according to the present invention,
The drive access permission flag and track unit access permission map in the disk controller can manage dual writing in track units, so only the tracks affected by the failure are reduced from double writing to single. No. Then, when the failure of the track is recovered, it is sufficient to perform the destaging processing only on the track, and the duplication control can be continued without being affected by other files or tracks at all. It is possible to improve the availability of the writing function and improve the reliability.

[0018]

[Brief description of the drawings]

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

FIG. 2 is a flowchart showing a processing procedure of a control processor when a drive failure occurs in the present invention.

FIG. 3 is a flowchart showing a processing procedure of a control processor in response to a read instruction in the present invention.

FIG. 4 is a flowchart showing a processing procedure of a control processor in response to a write instruction.

FIG. 5 is a flowchart showing a file recovery procedure of a host computer when a single degenerated track occurs in the present invention.

FIG. 6 is a block diagram of a disk subsystem showing another embodiment of the present invention.

FIG. 7 is a flowchart of a processing procedure of a control processor in response to a read instruction in FIG. 6;

FIG. 8 is a flowchart showing a processing procedure of the control processor in response to a write instruction when the cache is normal.

FIG. 9 is a flowchart showing a procedure of a stage processing of the control processor in FIG. 6;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Host computer 2 Disk controller 31 and 32 Disk drive 21 and 22 Control processor 23 Common control memory 24 Cache memory 231 Drive access permission flag 232 Track unit access permission map 233 Track unit non-reflection map

──────────────────────────────────────────────────の Continuing from the front page (72) Inventor Kenzo Kurihara 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Inside Hitachi, Ltd.System Development Laboratory Co., Ltd. (72) Inventor Yoshihiro Azumi 2880 Kozu, Kozu, Odawara City, Kanagawa Prefecture Hitachi, Ltd. Odawara Plant (56) References JP-A-3-94321 (JP, A) JP-A-58-175064 (JP, A) JP-A-2-37418 (JP, A) JP-A-3-268019 (JP, A A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 3/06 306 G06F 3/06 302 G06F 3/06 304

Claims (1)

(57) [Claims] When a write instruction is issued from a host computer, the disk controller writes the same data to two disk drives designated as a pair drive,
In response to a read instruction from the host computer, in a dual file control method of a disk subsystem for reading data from one of the paired drives, the disk control device includes an access unit for each of the paired drives in a track unit. Whether or not any of the tracks in the pair drive becomes inaccessible is stored, and the disk control device operates only the disk drive having a track corresponding to the track that can be normally accessed. Data read and write processing is performed on the tracks to perform singular degeneration on a track basis, and for tracks accessible to both of the pair drives, duplex control is performed. Host computer with cache memory When a write command is issued from the data storage device, data is written to either one of the paired drives and the cache memory, and writing to the other disk drive is performed later as a destage process from the cache memory asynchronously. A duplicated file control method, comprising: when the disk control device instructs writing to a track stored as being inaccessible to one of the paired drives, the corresponding track of the normal disk drive and the cache memory. After the data is written to the track, when the track stored as inaccessible is restored, the inaccessibility of the track is changed to accessible, and the data in the cache memory is written to the track in the disk drive. Especially Ri, performs a de-stage processing, duplicated file control method is characterized in that to resume the duplicated control.