JPH07114444A - File movement control system - Google Patents

Abstract

PURPOSE:To allow other disk controller to accept a data reference update request from a host computer during file movement or copy processing by storing file processing control information in a drive so as to improve the reliability of the file against a fault of a file processing unit. CONSTITUTION:A disk controller 100 receives a storage location of a movement source file 150 and a movement destination file 151 and a movement command between them from a host computer 109, then allocates a destination track in the unit of tracks, copies data on a source track to the destination track sequentially and connects the source track and the destination track by means of an alternate track pointer. A reference update request from the host computer 109 is executed to data on the source track when the data of the source track are not copied. When the data of the source track are copied, the request is executed to data on the destination track pointed out by the alternate track pointer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk control device, and more particularly, to a data reference update request from a host computer to a file to be moved during movement of a file storage location in the same drive or copy processing. By executing, it relates to a method for improving file processing throughput of a disk subsystem.

[0002]

2. Description of the Related Art In a conventional disk controller, for example,
As disclosed in Japanese Patent Laid-Open No. 1-230147, when a data update request is made from the host computer to the copy source file during the copy process from the copy source file to the copy destination file, the update data is copied to the copy source file. Write to file Furthermore, the update data address is stored, and the update data address is compared with the copy pointer that stores that the copy to the copy destination file has been completed, and when it is determined that the address has been copied, the copy destination file When it is determined that the data has been written to the copied portion of the file and is not copied yet, the data update request from the host computer may be executed during the file copy processing without writing to the copy destination file.

However, in the above-mentioned conventional technique, control information used for copy processing, such as a copy pointer for storing a copied portion, is held only in the file processing device,
A copy process was realized.

[0004]

In the above-mentioned prior art, during the file processing, a file processing device holding control information (file processing control information) such as a copy pointer for processing a file causes a power failure, a processor failure, a bus failure. When a failure or the like occurs, the file processing control information held before the failure occurs is lost, the consistency of the file cannot be obtained, and the reliability of the file is lowered.

Further, when a plurality of file control devices share a file, in the above-mentioned conventional technique, a common control memory for storing control information used for copy processing such as a copy pointer is provided between the plurality of file processing devices. I had to have it.

The object of the present invention is to improve the reliability of a file against a failure of the file processing device by holding the file processing control information in the drive, and to transfer the data from the host computer during the file moving or copying process. This is to enable the reference update request to be accepted by any disk control device in the plurality of disk control devices.

[0007]

In order to solve the above-mentioned problems, a disk controller according to the present invention comprises: (1) A disk controller transfers a source file (copy source file) and a destination file (from a host computer). Copy destination file) storage locations and instructions to move between them are received, destination tracks (copy destination tracks) are assigned to each track, and source tracks (copy source tracks)
A first means for sequentially copying the above data to a moving destination track (copying destination track) and then connecting the moving source track (copying source track) and the moving destination track (copying destination track) with an alternate track pointer.

(2) If the replacement track pointer does not exist by checking the existence of the replacement track pointer, it is determined that the source track (copy source track) has not been copied, and the replacement track is determined. If the pointer exists, it is judged that copying has been completed, and if the moving source track (copying source track) exists in the buffer, the second means for judging that copying is in progress.

(3) When the reference source update request from the host computer indicates that the source track (copy source track) has not been copied, the source track (copy source track).
Executed on the above data, and when it has been copied,
Third means for executing data on a destination track (copy destination track) pointed to by the alternate track pointer and executing data on a cache slot when a copy process is in progress.

(4) Transfer source file (copy source file)
A fourth means for cutting the replacement track pointer after the copy processing of the transfer destination file (to the copy destination file) is completed.

It is characterized by comprising:

[0012]

The file can be copied by the first means. During this time, when a reference update request is made to the file from the host computer, the second means determines whether or not there is a replacement track pointer, thereby determining that the copy state of the source track (copy source track) has not been copied. It can be defined as either already copied or being copied. As a result, the third means can execute the reference update request that matches the data copy state. As a result, the reference update request from the host computer can be executed consistently during the file move or copy process. Furthermore, the fourth
By cutting the alternate track pointer by this means, execution of a file move or copy command from the host computer can be completed.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 is a block diagram of a disk subsystem with a cache which is a target of an embodiment of the present invention.

As a prerequisite for the present invention, the drive 111
Is shared by a plurality of disk control devices 100 and 162. When the drive 111 is being used by the disk control device 100, the drive 111 gives a busy notification to the access from the disk control device 161.
62 is excluded. Alternatively, it has information that excludes the disk control device 100 and the disk control device 162.

In FIG. 1, a disk control device 100
Is connected to the host computer 109 via the channel controller 101 on the upper side, and is connected to the drive 111 which is a magnetic storage medium on the lower side. Drive 111
The source file 150 and the destination file 151 are included in the
And the failure recovery information 1000 is stored. A specific description of the failure recovery information 1000 will be given later. The disk control device 100 reads and writes data in the drive 111 in response to a request from the host computer 109. The control processor 102 built in the disk control device 100 controls the data transfer between the drive 111 and the host computer 109. The control processor 102 is connected to the host computer 109 and the drive 111 via the channel control device 101 and the drive interface 107. The control memory 104 is
It is a common memory that can be accessed by all the control processors 102, and the disk controller 100 is a drive 111.
The common control information for accessing is stored. The common control information will be described later. Cache memory 10
5 is also a memory that can be accessed by all control processors 102.
It is used to temporarily store the data read by. The control processor 102 is also connected to the service processor 120. When the operation panel 121 instructs the service processor 120 to update the common control information in the common memory 104, the service processor 120
An update request is sent to the control processor 102, and the selected control processor 102 updates the common control information in the control memory 104.

The disk controller 162 is connected to the host computer 160 on the upper side, and is connected to the channel controller 1
61, and the configuration is the same as the disk control device 100.

Next, the common control information will be described. The common control information includes a drive control block and a file move request table. Hereinafter, they will be described in order.

FIG. 2 shows a drive control block 200 (De
vice Control Block: hereinafter referred to as DCB).
Seven data are stored in the DCB 200. A drive number 201 (Device Connection Address: hereinafter referred to as DCA) for the disk control device 100 to identify each drive, a drive type 202 indicating drive information such as the track length of the drive, and a movement destination end track number of a movement target file. 203, moved track number 204, move source track pointer disconnected track number 205, host reference update request flag 2 for the drive
06 and the drive lock flag 207 are stored.
The host reference update request flag 206 is set by the control processor 102 when the host computer 109 issues a reference update request to the drive 111 that is in the process of moving a file.
Set to'on '. When there is no reference update request,
Set it to'off '. The drive lock information 207 is used by the control processor 102 when exclusively controlling drive access from another control processor, set to “on” during lock acquisition, and when unlocked (unlocked). Is set to'off '.

FIG. 3 shows the file movement control information. The file move request table 300 and the move processing work area 310 are stored in this control information. In the file move request table 300, the drive number 301 of the drive having the move target file, the move source start track number 302 indicating the storage position of the move target file, the move source end track number 303, and the move destination of the move target file. The destination start track number 304 indicating the number is stored.
The movement processing work area 310 is an area used for the movement processing of a track, and stores a movement source track number 311 and a movement destination track number 312.

Next, the operation of the control processor 102 of the disk control device 100 in this embodiment will be described.

FIG. 4 shows a file storage location move process 400.
Is the main flow of. File storage location move processing 40
0 indicates that the disk controller 100 is the host computer 1
It is activated when a file move command is received from the service processor 120 through the operation panel 121, or when a file move command from the server 09 is received.

First, the file move request table 300 is read (410). Specifically, the drive number 301 having the file to be moved and the source track number 3 of the movement source
02, the source end track number 303, and the destination start track number 304 are read (410). Then, the lock information 207 of the drive is judged (415), and if it is already'on ', the drive busy is returned to the host computer (425), after which the process is terminated (44).
0). If the lock information 207 is'off ',
The acceptance of the file move command is reported to the host computer (420), and the move process 500 is performed. Subsequently, post-processing 800 after the moving processing is performed, and the host computer 10
Report the completion of file transfer to 9 (43
0), the process ends (440).

FIG. 5 is a flow chart of the movement processing 500 in FIG.

First, the movement source start track number 302 is set to the movement source track number 311 (505). next,
Zero the moved track number 204 (510),
The lock of the drive having the drive number 301 having the file to be moved is acquired (515). Then, the copy processing 60 of the source track data onto the destination track is performed.
0 is executed to release the lock of the drive having the file to be moved (520). Next, the host reference update request flag 206 is checked to see if there is a reference update request to the drive from the host computer 109 (525). If the request flag 206 is'on ', the reference update processing 700 is executed. Request flag 20
If 6 is'off ', it is determined whether the moving process is completed (530). In step 530, the moved track number 204 and the movement source end track number 303 are compared, and if they are equal, it is determined that the movement process has ended. When it is determined in step 530 that the moving process is completed, the moving destination end track number 203 is set in the DCB 200 of the target drive (540), and the moving process 500 is performed.
Ends, and the process returns to the file storage location move processing 400 (545). If the moving process is not completed, the source track number 311 is incremented (53
5) Return to step 515 and repeat the movement process.

FIG. 6 is a flow chart of a copying process 600 of the source track data on the destination track in FIG.

Prior to the description of the copying process 600, the track format in this embodiment will be described with reference to FIG. In FIG. 9, 900 indicates a movement source track, and 905 indicates a movement destination track. 901, 90
Reference numeral 6 denotes a header portion of each track, and 903 and 90
Reference numeral 8 indicates an alternate track pointer. Also, 902, 9
Reference numeral 07 indicates the data portion of each track.

First, the destination track is assigned as the destination of the source track, and the assigned track number is set to the destination track number 312 (605). At this time, the allocation is started from the track having the destination start track number 304. Next, the source track number 3
Data 902 is loaded into the cache memory 105 from the track having 11 (610). For the track having the destination track number 312, the cache memory 1
The data loaded in 05 is written (615). A pointer 903 pointing to the destination track is written in the header section 901 of the source track (620). Next, a pointer 9 pointing to the source track is added to the header section 906 of the destination track.
08 is written (625). DCB of target drive
Set the moved track number 204 to 200 (63
0), and returns to the movement process 500 (635).

FIG. 7 shows a flow of the reference update processing 700 for executing the reference update request from the host computer 109 during the file move processing in FIG.

First, the lock of the drive is acquired (70
5). The presence or absence of the pointer 903 of the header portion 901 of the reference update request track (moving source track) on the drive 111 is checked (710). Destination track pointer 9
If there is 03, the reference update request is executed for the data on the move destination track (720), and if there is no pointer 903, it is executed for the data on the reference update request track (the move source track) (725). . In this way, to determine whether or not the source track has been copied, the presence or absence of the pointer 903 in the header section 901 of the source track may be checked. Then unlock the drive (7
30), the host reference update request flag 206 to the drive
Is set to'off '(735), and the process returns to the moving process 500 (740). By acquiring and releasing the lock of the drive having the file to be moved in units of one track, it is possible to accept the reference update request to the file from the host computer 109.

FIG. 8 shows a post-processing flow 800 after the moving process in FIG.

After the movement is completed, the process 800 is a replacement track pointer 90 between the movement source track and the movement destination track.
This is processing for cutting 3,908.

First, the track number 205 which has been disconnected from the source track pointer is cleared to zero (805), and the lock of the drive having the file to be moved is acquired (81).
0), the destination track number 312 is set to the destination start track number 304 (815). Next, the source track pointer 908 of the header part 906 of the track having the destination track number 312 is cleared to zero (82).
0). Then, the destination track pointer 903 of the header part 901 of the track having the source track number 312
Is cleared to zero (825). As a result, the replacement track pointer 90 between the source track and the destination track 90
3, 908 are separated, and the movement of the movement source track is completed. Next, the DCB 20 of the drive to be moved
0 is the track number of the source track pointer and the track number is 20
5 is set (830), and the track number 205 of the source track pointer disconnection and the track number 203 of the destination end track are compared (835). If the results of the comparison are equal, the lock of the drive having the file to be moved is released (84
5) The process returns to the file storage location move process 400 (850). If the comparison results are not equal, the destination track number 312 is incremented (840),
Until the replacement track pointers 903 and 908 are completely disconnected, the replacement track pointer disconnection process (820, 82)
Repeat 5).

While the process 800 is being executed, the drive is locked because the drive is locked in response to a request from the host computer 109, and the host computer 1
There is a waiting time for the reference update request from 09. However, this waiting time is limited to the conventional host computer 10
It is shorter than that of 9-led file move processing.

FIG. 10 is a diagram showing the structure of the fault recovery information 1000.

In the above-described file moving process, the disk controller 100 is responsible for updating the header portion 901 of the source track or the header portion 906 of the destination track.
Therefore, the failure recovery information 1000 must be held in case of a failure (such as a power failure) of the disk controller 100.

FIG. 10 shows the contents of the fault recovery information 1000. The fault recovery information 1000 is the source track number 3
11, the destination track number 312, the execution step number 1003 of each processing flow, the header section 10 of the source track
04 (corresponding to 901 in FIG. 9) and the header section 1005 (corresponding to 906 in FIG. 9) of the destination track. These can be used as needed during the file move process.
It is written in the failure recovery information 1000 of the drive to be moved. If the file is being moved, the control processor 10
2 or 103 causes a failure and loses the header section 901 of the source track and the header section 906 of the destination track, the failure recovery information 1 retained in the drive 111
By reading 000, it is possible to know the header portions 901 and 906 of the track at the time of failure occurrence, so that the interrupted processing can be restarted and the reliability is improved.

In the present embodiment, the cache memory 105 is used to temporarily store the track data read by the control processor 102 from the drive 111.
The file moving process can be similarly realized by using a buffer capable of storing track data for one track instead of the cache memory 105.

Next, as a second embodiment, an example using the cache memory 105 and the backup memory 110 will be described with reference to FIG. In the first embodiment, the update data on the track is stored in the cache memory 105, but since the cache memory 105 is volatile, the data is lost if a power failure occurs. In the second embodiment, in order to deal with this, the non-volatile backup memory 110 is used for the purpose of backing up the cache memory 105 and the update data is stored.

FIG. 11 is a flow chart of the file move processing 1100 in this embodiment. This process corresponds to the move process 500 in the file storage location move process 400.

First, since the file to be moved is divided into a plurality of track units including at least one track and loaded into the cache memory 105, the lock of the drive to be moved is acquired (1105). Next, the source track divided into a plurality of tracks is loaded into the cache memory 105 for each track (1110). At this time, a location on the cache memory 105 for storing one track of data is called a cache slot. afterwards,
The lock of the drive to be moved is released (1115). It is judged whether or not all the movement source tracks in the track unit have been loaded into the cache memory 105 (1120), and if not moved, the next movement source track is set in the cache memory 1
When it is loaded in 05 and has been moved, the following processing is performed. During this period, the reference update request to the transfer source track loaded on the cache memory 105 is executed on the cache memory 105, the update data is stored in the backup memory 110, and the reference update request to the track not on the cache memory 105 is executed. Execute on a track on the drive 111. On the cache memory 105 (updated if there is an update request from the host computer 109)
In order to copy the data of the source track to the destination of the drive 111, the lock of the target drive is acquired (1
125), the track copy processing 1200 using the cache memory 105 and the backup memory 110 is executed. Then, unlock the drive to be moved (1
130). By releasing the cache slot in the cache memory 105 corresponding to the moved source track, the reference update request to the data in the unit of a plurality of uncopied tracks is executed in the source track on the drive 111 and copied. The reference update request to the data in the already completed plural track units is executed in the movement destination track pointed by the alternate track pointer on the drive 111. The reference update request to the transfer source track in the cache memory 105 which has not released the cache slot is issued by the cache memory 1
05 is executed in the cache slot and the updated data is stored in the backup memory 110. Next, it is judged whether or not there is a movement source track which has not been moved yet in the cache memory 105 (1135), and if there is a movement source track, the next movement source track is moved. If not, it is judged whether all the files to be moved are copied to the destination file (1
140), if it is not copied, the unmoved track of the file to be moved is copied. If it has been copied,
The process returns to the file storage location move processing 400 (114
5) The post-process 800 after the moving process for disconnecting the replacement track pointers 903 and 908 is executed to report to the host computer 109 that the file to be moved is completed (430), and the process ends.

FIG. 12 is a track copy process 1 using the cache memory and the backup memory in FIG.
It is a flow of 200.

First, the lock of the cache slot storing the data of the source track on the cache memory 105 is acquired (1205). The source track data is assigned to the destination track on the drive 111 on a track-by-track basis, the data on the source track is written to the destination track, and connected by the alternate track pointer (12
10). Next, the slot of the cache memory 105 storing the movement source track data is released (121
5) The cache slot of the source track is released from the cache memory 105 (1220). At this point, the next reference update request to the source track is executed on the drive 111. After that, the file move processing 122 using the cache memory and the backup memory is performed.
It returns to 00 (1225).

In the case of a failure in this embodiment, the first processing is
The processing is the same as that of the embodiment described above.

As in the second embodiment using the cache memory 105 and the backup memory 110 described above, the cache memory 105 and the backup memory 11
Even if 0 is used, the file moving process described in the first embodiment can be realized similarly.

The two examples have been described above. The first embodiment can be easily realized by only preparing a cache memory or a buffer capable of storing track data for one track. The feature of the second embodiment is that the first
By providing the cache memory having a larger capacity than that of the above embodiment, a plurality of tracks can be loaded on the cache memory. As a result, the reference update request from the host computer can be executed on the cache memory, and the time to acquire the drive lock becomes shorter than when the reference update request is executed on the drive as in the first embodiment. The number of lock acquisitions is also reduced, and the reference update request is not kept waiting for a long time. In addition, by providing a backup memory that is a non-volatile memory,
The data in the cache memory can be retained, and the reliability against failures is improved. Further, since the update data is sequentially written to the tracks on the drive by using the alternate track pointer, it is not necessary to store the update data on the cache memory and the backup memory.

In the above embodiments, the operation of the disk control device has been described. However, since the alternate track pointer is held in the drive, the disk control device updates the data of the disk control device during file processing. The request can be fulfilled by following the alternate track pointer. In the embodiment, the file moving process is described, but the file copying process can be performed in the same manner.

According to the embodiment described above, the disk control device receives an instruction to move the file storage location from the host computer, and independently executes the file moving process,
Even when the file is being moved, the reference update request to the file can be executed from the host computer, and the file processing throughput of the disk subsystem can be improved. In addition, even if the processing is interrupted due to a failure that occurred during processing, the interrupted processing can be continued from the failure recovery information on the drive, and the information in the header of the track is not lost The property is improved. Further, since the drive lock is acquired and released for each track unit, when the drive lock is released, the drive lock can be obtained for the reference update request from the service processor to the file. The request and the request from the host computer 109 can be executed in parallel. Further, when the file is moved, the file movement control information such as the alternate track pointer is held in the drive, so that the reliability of the file against the failure of the disk control device is improved. Further, another disk control device can receive a data reference update request from the host while the disk control device is moving a file in the drive or executing a copy process.

[0048]

According to the present invention, the file movement processing control information such as the alternate track pointer is held in the drive, so that the reliability of the file with respect to the failure of the disk controller is improved. Furthermore, in response to a file storage location move instruction from the host computer, while one disk controller is executing a file move process, another disk controller can issue a reference update request from the host computer to the file being moved. It can be executed and the file processing throughput of the disk subsystem is improved.

[Brief description of drawings]

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

FIG. 2 is a configuration diagram of a drive control block.

FIG. 3 is a configuration diagram of file migration control information.

FIG. 4 is a flowchart of a file storage location moving process in the first embodiment.

FIG. 5 is a flowchart of a moving process.

FIG. 6 is a flowchart of a process of copying the source track data to the destination track.

FIG. 7 is a flowchart of a process of executing a reference update request from a host computer.

FIG. 8 is a flowchart of post-processing after moving processing.

FIG. 9 is a format diagram showing a track format.

FIG. 10 is a diagram showing the contents of failure recovery information.

FIG. 11 is a flowchart of a file moving process according to the second embodiment.

FIG. 12 is a flowchart of a track data copying process.

[Explanation of symbols]

100 ... Disk controller, 102, 103 ... Control processor, 1
04 ... Control memory, 105 ... Cache memory, 107, 108 ...
Drive interface, 109 ... Host computer, 1
10 ... Backup memory, 111 ... Drive, 120 ... Service processor, 130, 131 ... Signal line between channel controller and control processor, 132, 133 ... Signal line between service processor and control processor, 150 ... Source file, 151 …
Destination file, 1000 ... Disaster recovery information.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Mitsuaki Nita 2880 Kozu, Odawara-shi, Kanagawa Stock Company Hitachi Storage Systems Division

Claims (3)

[Claims] 1. A disk subsystem comprising a disk controller and a drive which is a magnetic storage medium, and a file storage including a storage location of a file to be moved or copied from a host computer and a movement or copy destination of the file. Alternatively, the disk control device that receives the copy instruction assigns the move destination track or the copy destination track in track units corresponding to the move source track or the copy source track in which the file to be moved or copied is stored, After moving or copying the data on the moving source track or the copying source track to the moving destination track or the copying destination track, a replacement track between the moving source track or the copying source track and the moving destination track or the copying destination track Connect with a pointer to the copied data By performing a reference update on the data on the destination track or the copy destination track indicated by the alternate track pointer, and performing a reference update on the uncopied data on the data on the source track or the copy source track. A file move control method characterized in that a reference update request to a target file of the move or copy instruction is executed from a host computer. 2. The disk control device comprises a memory capable of storing at least one track of data, and moves from the source track or copy source track to the destination track or copy destination track via the memory. 2. The file movement control method according to claim 1, wherein a copy process is executed. 3. The disk control device further comprises a non-volatile backup memory for storing data stored in the memory by the moving or copying process and not stored in the moving destination track or the copying destination track. The update is executed to the memory, the update data and the information for identifying the track to be updated are stored in the backup memory, and the update process for the data stored in the move destination track or the copy destination track is performed.
3. The file moving control system according to claim 2, wherein the data is moved to the data of the moving destination track or the copying destination track.