JPH11143646A - Control method for external storage sub-system and external storage sub-system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the throughput of data set erasure in the storage device of redundant storage constitution emulating plural logic devices. SOLUTION: In a sub-system where a controller 12 is arranged between a host device 10 and the storage device 13, and the storage device emulates the plural logic devices, an erasure holding flag is provided in the common memory 123 of the controller 12, and control logic holding the execution of whole erasure when the whole erasure command of logic tracks in the logic devices are received from the host device 10, recording the holding state in the erasure holding flag, resetting the erasure holding flag at the time of receiving a system write request for setting the later new data set and executing an erasure processing to a logic track end so as to avoid the writing of overlapped wasteful erasure data is provided.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control technique for an external storage subsystem and an external storage subsystem. The present invention relates to a technology that is effective when applied to an external storage subsystem or the like that stores data in a logical drive constructed on a storage device.

[0002]

2. Description of the Related Art Takeuchi Shoten Shinsha Co., Ltd., 19
As described in the literature such as “Magnetic Disk Techniques Collection = Programming Handbook =”, published on April 30, 1983, P45-P67, a disk device that records a variable-length data format is generally used. When the format write is executed, the data is stored in a predetermined area, and data subsequent to the record is erased to the end of the track to invalidate the data.

When erasing data of a logical device stored in an external storage device through a host device, the following two means are generally used. One is to invalidate the position information of the data set recorded at the head of the logical device by writing the format.

However, in this method, a program which does not refer to the position information of the data set may refer to the data of the already invalidated track. For this reason,
As another means, it is conceivable to invalidate all data of the data set to be erased by the logical device by format writing.

[0005] A command chain for invalidating the user data of the entire track issued by the host device issues a format write command (Write R0) to a logical record number 0 which is not normally used by the user, or a logical record. It is reasonable to execute a format write (End Of File record) in which the key length and the data length are zero for number 1. This is because the processing for invalidating the data of all the tracks is desirable because a data length that minimizes the data transfer time with the host device is desirable.

The conventional erasing process after the end of the format writing is often performed immediately after the format writing. Therefore, when the format writing is further performed on the track, the erasing process after the format writing process is executed again. As a result, redundant and useless erasure is performed. As a method for solving this technical problem, for example, there is a technique disclosed in Japanese Patent Application Laid-Open No. 61-241824. According to the technique disclosed in Japanese Patent Application Laid-Open No. 61-241824, a table is provided in a memory in a disk device for each logical track with respect to an erase operation after completion of format writing, and the format is dynamically referred to at the time of format writing to complete format writing. By eliminating unnecessary erasing processing from the erasing processing at the time, format writing is efficiently performed.

[0007]

According to the above-mentioned prior art, it is very difficult to omit unnecessary erasing processing after format writing performed to add a new logical record to the format-written logical track. If the data set is valid but is re-created after the format writing that invalidates the entire logical track, the erasing is executed again. Further, in a disk array device emulating a plurality of logical devices, it is physically impossible to store the control table in the disk device.

In the disk array device, when a logical volume is invalidated for a plurality of logical devices, the data transfer time with the host device is minimized as described above. If the logical device conflicts with a disk device in the array group, there is a concern that performance may be a significant technical problem because the data that is normally erased is also included in the data.

It is an object of the present invention to provide an external storage subsystem capable of realizing an improvement in performance by optimizing an erasing process without deteriorating the reliability of operation, and a control technique therefor.

Another object of the present invention is to provide an external storage subsystem capable of improving the throughput of erasing a data set in a storage device having a redundant storage configuration emulating a plurality of logical devices, and a control technique therefor. It is in.

Another object of the present invention is to provide an external storage subsystem capable of shortening the required time in the initialization format of a logical device and realizing the management and confirmation of the initialization state, and a control technique therefor. Is to do.

[0012]

According to the present invention, a means for recognizing a command chain to be used for erasing a data set is provided by a control device, and a management flag of, for example, one bit is set for each logical track. I do. If it is recognized that the data set is deleted, the data set is constructed again (for example, logical record number 1).
Erasure processing for the relevant track is suspended until the execution of this management flag, and the validity / invalidity of the data of the relevant logical track is determined by the setting and release status of this management flag.
It determines invalidity and controls access to the logical track from the host device.

When the data of the invalidated track is reflected from the cache memory to the disk device, the data itself is not recorded, only the management flag information is recorded, and the data in the track is invalid. Used as a means to ensure that

Further, at the time of initializing and formatting a logical medium (device), only the management flag is set for each logical track, and the actual writing of data for formatting is suspended. If necessary, the status of the management flag in the initialization format is visualized and output to a terminal such as a service processor.

As described above, in the present invention, in consideration of the feature of the command chain that would normally be used for erasing the data set, in such a case, the command chain for constructing the data set is again issued from the host device side. Since the erasure of the track is suspended until the data is issued, the processing time can be reduced by optimizing the erasure of the data set. Since the amount of data transfer processing with the host device handled by the command chain used for erasing the data set is clearly smaller than that when the format writing is performed again, erasing is performed by erasing the data set. Also,
It is clearly more reasonable to perform erasing in the remaining area of normal data writing in the formal writing for the construction of a data set to be subsequently processed, since the amount of data for erasing is reduced.

[0016]

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

FIG. 1 is a block diagram showing an example of a configuration of an information processing system including an external storage subsystem according to an embodiment of the present invention. FIG. 2 is a block diagram showing an external storage subsystem according to the present embodiment. FIG. 3 is a conceptual diagram showing an example of management information used in the embodiment, FIG. 3 is a conceptual diagram showing an example of a data storage method in the external storage subsystem of the present embodiment, and FIG.
It is a conceptual diagram showing an example of the visualization output of management information used in the external storage subsystem of this embodiment. Also,
FIGS. 5 and 6 are flowcharts showing an example of the operation of the control method of the external storage subsystem according to the present embodiment.

The information processing system of the present embodiment shown in FIG. 1 comprises a host device 10, a channel 11 for controlling data input / output thereof, and a subordinate external storage subsystem.

The external storage subsystem according to the present embodiment comprises:
Storage device 13 composed of a plurality of disk devices 13a
And a control device 12 for controlling data transfer interposed between the storage device 13 and the channel 11.

In the case of the present embodiment, as illustrated in FIG. 3, as an example, a plurality of disk devices 13a of the storage device 13 constitute a disk array, and a plurality of logical devices 13b are set. Individual logical device 13b
Includes a plurality of logical tracks 13c. further,
Each logical track 13c has a plurality of disk devices 13
A unit is composed of a plurality of unit data 13d distributed and stored in a and redundant data 13e such as parity generated from the unit data 13d, and adopts, for example, a RAID 5 redundant storage configuration. Thus, when a failure occurs in one unit data 13d in each logical track 13c, it is possible to restore the failed data from the other sound unit data 13d and the redundant data 13e. In the case of the present embodiment, each logical device 1
For the logical track 13c of 3b, for example, CK
Data is stored in a D (variable length recording) system.

The controller 12 mainly performs a process when an input / output request is received from the channel 11.
A processor D122 for controlling storage and readout of data with respect to the subordinate storage device 13;
A data transfer unit H124 for transferring data from the storage device 13 and a data transfer unit D125 for transferring data from the storage device 13.
And data and the like transferred between the storage device 13 and the channel 11 are accessed from the cache memory 126 temporarily stored and the plurality of processors H121 and D122, for example, in units of logical tracks 13c. The shared memory 123 stores management information and the like for performing exclusive control between the two.

In this embodiment, the track management table 127 downloaded from the storage device 13 is stored in the shared memory 123 as needed.

As exemplified in FIG. 2, the track management table 127 of the present embodiment stores, for each logical track, track management information 127a related to the logical track, for example, for controlling a disk array. To
1-bit erase hold flag 127 for identifying whether or not an erase operation is held for the logical track
b is stored.

The control unit 12 includes a service processor 14 having a user interface such as a display 14a, a keyboard 14b, and a mouse 14c, as necessary.
Is connected, so that a maintenance manager or the like can monitor and control the operation of the control device 12 from outside.

Hereinafter, an example of the operation of the present embodiment will be described. First, as in the present embodiment, when a disk array is configured by a plurality of disk devices 13a of the storage device 13 and a plurality of logical devices are emulated, the individual devices must be emulated before starting to be used by the host device 10. Logical device 1
It is necessary to execute the initialization process 3b.

In this initialization processing, the initialization format of the logical track 13c of the emulated logical device 13b is executed. At this time, in the case of the present embodiment, the erase is performed on all the logical tracks of the logical device 13b. Only the setting of the hold flag 127b omits the writing of the initialization data. If there is an input / output request from the host device 10, the erase pending flag 12
By referring to 7b, it can be determined at one time that data does not exist (initialized).

If necessary, the display 14a of the service processor 14 may be used as shown in FIG.
Then, by visualizing and outputting the initialization state (the state of the erasure hold flag 127b) of each logical device 13b, the system maintenance manager can visually confirm the state.

The control device 12 according to the present embodiment performs a process when a format write for invalidating the entire track is issued according to the procedure illustrated in FIG.

When an input / output request is issued to a certain logical track, the control device 12 executes step 300 for ascertaining the existence of data on the track. Then, it is determined in step 310 whether or not the track management table 127 of the target logical track exists in the shared memory 123.
In the case of N, a step 315 of copying the track management table 127 to be processed to the shared memory 123 is executed. Thereafter, it is determined in step 320 whether or not the command is a format write command.

In the case of Y in step 320, a determination is next made in steps 330, 340, and 350 as to whether or not the access is such that the data of the logical track is erased. In step 330, Loc
ate Record or Locate Record
d The operation specified by the Extended command is performed after the data portion 8 bytes of the standard R0 record is transferred.
After the data transfer, a determination is made as to whether or not a Write Tracks operation for executing erasure up to the end of the track is performed. It is determined whether the data of the count part received from the channel 11 by the format writing (Write CKD command or Erase command) for the record number 1 has the key length of 0 byte and the data length of 0 byte. These steps 33
If 0, Y in steps 340 and 350, it can be recognized as a format write command that invalidates the data of the entire track. Therefore, the process of turning on the erase hold flag 127b is executed in step 360.
And the processing of the Write command is executed in step 37.
Perform at 0. At step 370, whether the format write command is chained or not is determined at step 330
Channel 1 except for te Track operation
Since it is not specified by the pattern of the command chain issued from No. 1, only the processing of the record processed by the command is executed, and the erasing of the data thereafter is not performed.

FIG. 6 shows the case where the determination at step 320 is N, that is, the command is not a format write command, and the case where the result of the determination performed at steps 330, 340, and 350 is N even if the command is a format write command, is shown in FIG. The determination of step 400 is performed.

FIG. 6 shows the flow of command processing other than the format write command for erasing the data of the entire track.

In step 400, it is determined whether or not the erase hold flag 127b is ON for the track to be accessed. If N in step 400, that is, if the erase hold flag 127b is not ON, normal command processing may be executed, and subsequent processing is omitted.

If step 400 results in Y, step 410 determines whether the command is a format write command.
If it is a command other than the format write, the command processing is performed in step 415 assuming that there is no data of the track. This is because, depending on the type of command, there is a command for the data of the next track. Although not shown in FIG. 6, a Read HA command or a Read
HA part and R0 like Record Zero command
It is clear from the spirit of the present invention that a command is normally accepted when a record is to be processed. Also, R
The key length and data length are 0 even if one record is to be processed
In the case of bytes, the processing is continued as End Of File.

If step 410 is Y, that is, if a format write command for constructing a new data set is issued, the erase pending flag 127b of the logical track is turned off in step 411, and in step 412 Execute the format write process. In the format write process of step 412, a process of writing specific erase data to the end of the logical track is executed.

When erasure is performed on a logical track on the cache memory 126, the operation of reflecting (ejecting) the logical track on the logical device 13b on the storage device 13 involves erasing the logical track. Only the hold flag 127b is stored in the storage device 1.
The operation of reflecting the data of the logical track on the logical device 13b on No. 3 is not performed. When an access request to the logical track reflected on the logical device 13b is issued from the host device such as the channel 11, the erase pending flag 127b relating to the logical track is read out from the storage device 13 to read the track. A process is performed in the management table 127 to respond that no data exists as an erased logical track.

As described above, according to the present embodiment, the erasure suspension flag 127b is stored in the track management table 127.
In the format write process for invalidating the data of the entire logical track, the erase pending flag 127 is provided.
Since only the setting of b and the data transfer process performed with the channel 11 need be performed, the process of writing duplicate erase data is avoided, the speed of the data set erasing process is significantly increased, and the throughput of the data set erasing is improved. , The upper level device 10 for the track recognized as erased by referring to the erase hold flag 127b.
, It is possible to accurately respond to the access request from the user that there is no data, so that the same reliability as the conventional control method can be secured.

As a result, for example, a stable response time can be obtained even when a similar job is executed in multiplex.

In particular, for example, when the storage device 13
In a system in which a plurality of logical devices 13b are set and used by technology or the like, when erasure of a logical track 13c is performed simultaneously by a plurality of logical devices 13b, an access conflict occurs at the level of a physical storage device as it is. Then, the required time of the erasing process varies depending on the logical address. In the case of the present embodiment, the erasing process is performed with a constant response time regardless of the logical address of the logical track 13c to be erased. Can be.

Although the invention made by the present inventor has been specifically described based on the embodiments, the present invention is not limited to the above-described embodiments, and can be variously modified without departing from the gist thereof. Needless to say, there is.

[0041]

According to the control method of the external storage subsystem of the present invention, it is possible to realize the performance improvement by optimizing the erasing process without impairing the reliability of the operation.
The effect is obtained.

Further, according to the method of controlling the external storage subsystem of the present invention, it is possible to improve the throughput of erasing a data set in a storage device having a redundant storage configuration emulating a plurality of logical devices. Can be

Further, according to the method of controlling the external storage subsystem of the present invention, it is possible to shorten the required time in the initialization format of the logical device, and to manage and confirm the initialization state accurately. Is obtained.

According to the external storage subsystem of the present invention,
The effect is obtained that the performance can be improved by optimizing the erasing process without impairing the reliability of the operation.

Further, according to the external storage subsystem of the present invention, it is possible to improve the data set erasing throughput in a storage device having a redundant storage configuration emulating a plurality of logical devices.

Further, according to the external storage subsystem of the present invention, it is possible to shorten the required time in the initialization format of the logical device, and to manage and confirm the initialization state accurately. .

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an example of a configuration of an information processing system including an external storage subsystem according to an embodiment of the present invention.

FIG. 2 is a conceptual diagram illustrating an example of management information used in an external storage subsystem according to an embodiment of the present invention.

FIG. 3 is a conceptual diagram showing an example of a data storage method in an external storage subsystem according to an embodiment of the present invention.

FIG. 4 is a conceptual diagram showing an example of visualization output of management information used in an external storage subsystem according to an embodiment of the present invention.

FIG. 5 is a flowchart illustrating an example of an operation of a method of controlling an external storage subsystem according to an embodiment of the present invention.

FIG. 6 is a flowchart illustrating an example of an operation of a method of controlling an external storage subsystem according to an embodiment of the present invention.

[Explanation of symbols]

10: Host device, 11: Channel, 12: Control device, 1
3 storage device, 13a disk device, 13b logical device, 13c logical track, 13d unit data,
13e: redundant data, 14: service processor, 14
a display, 14b keyboard, 14c mouse, 121 processor H, 122 processor D, 1
23 ... shared memory, 124 ... data transfer unit H, 125 ...
Data transfer units D, 126 ... cache memory, 127 ...
Track management table, 127a: track management information,
127b: Erase hold flag

Continued on the front page (72) Inventor Hiroaki Konuma 2880 Kozu, Kozuhara, Kanagawa Pref., Ltd.Storage Systems Division, Hitachi, Ltd.

Claims (3)

[Claims] 1. A host device comprising: a plurality of rotary storage devices; and a control device including a cache memory for temporarily storing data transmitted and received between the rotary storage devices and a host device. A method of controlling an external storage subsystem that emulates a plurality of logical devices with respect to
When accessing the data of the logical device from the higher-level device is performed in units of logical tracks that are each divided into a plurality of unit data and distributed and stored in the plurality of rotary storage devices, the higher-level device Wherein when an erasure instruction is issued to the logical track, the erasure operation up to the end of the logical track is suspended. 2. A control device comprising: a plurality of rotary storage devices; and a control device including a cache memory for temporarily storing data exchanged between the rotary storage device and a host device. Emulates a plurality of logical devices with respect to the higher-level device, and accesses from the higher-level device to the data of the logical device are each divided into a plurality of unit data and distributed to the plurality of rotary storage devices. An external storage subsystem configured to be executed in units of logical tracks stored as data, comprising: a management flag for managing data in each of the logical tracks; When an erasure instruction is given to any of the logical tracks, all data in the logical track is invalid for the management flag corresponding to the logical track. External storage subsystem comprising the control logic sets the information indicating the bets, suspends the erase operation to the end of the logical track. 3. The external storage subsystem according to claim 2, wherein the control device refers to the management flag when the higher-level device accesses an arbitrary logical track, and the management flag is invalid. In the case of the above, an operation of regarding that there is no data in the logical track and processing is performed, and for the logical track for which the management flag is set, a format write for constructing a new data set is performed from the upper-level device. When executed, or when the power of the external storage subsystem is turned off, an operation of releasing the management flag and executing an erasing process up to the end of the logical track after the format writing process, wherein the management flag is set When flushing the invalidated data of the logical track and the management flag from the cache memory, Data corresponding to the logical track is not recorded in the rotary storage device, only the management flag is recorded on the rotary storage device, and when there is an access request to the logical track from the higher-level device, the management is performed. An operation of storing management information in which a flag is stored in the cache memory and executing a process as no data in response to the access request from the higher-level device, which is performed before starting use from the higher-level device In the initial format processing of the logical device, the management flag is set for all the logical tracks configuring the logical device, and writing of specific format data is suspended, and when the initial format is completed, Operation to visualize and output the initialized state of the logical device data to the outside External storage subsystem, characterized in that it comprises a control logic for performing one operation.