JPH0490034A - Data set allocating method for external storage device - Google Patents

Abstract

PURPOSE:To improve the archive performance even in the case of a high frequency in update of data sets by transporting saving only the data set area on a storage device, which resolves the cause of archive like a lack of space, at the time of archive of data sets. CONSTITUTION:A management register control part 8 of a data processing part 2 manages allocation areas of data sets on a higher-order storage device 9 in continuous area (extent) units. If the space is deficient at the time of allocating a new data set on the higher-order storage device 9, an extent having size enough to resolve this lack of space is selected, and only this extent is transported to a lower-order storage device 13. When transportation of the extent is completed, the extent state of a data set management register 15 is updated. Thus, the transportation of data sets is reduced even in the case of a high frequency in data set update, and data sets are efficiently allocated.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an efficient method of allocating data sets in an external storage device of a computer system that is compatible with high-speed data set movement.

[Conventional technology]

Conventionally, in computer systems, a collection of data (hereinafter referred to as a data set) according to the purpose and use is stored on an external storage device, and these data sets can be accessed directly from this external storage device. Is going.

A table for managing external storage devices is developed on the main storage device, and the program performs processing while referring to/updating the table. A plurality of external storage devices are provided in a hierarchical manner, and when a higher-order external storage device runs out of space, the lower-order external storage device is used.

In this way, as a program for managing a plurality of storage devices each having different memory capacity and cost and arranged hierarchically, for example, HARC (Hierarch
ical Storage Archive a
ndRetrieve Controller). With this HARC, when a new data set (collection of data) is allocated to a direct access device (upper storage device) with a small capacity and high cost, if a lack of space occurs, the A large-capacity, low-cost storage device such as MSS
(The storage capacity of one unit is 100 million bytes, with two magnetic tape data cartridges stored in a Mass Storage System niblastic container serving as one unit). In HARC, data sets are moved from the upper storage device to the lower storage device in order to cope with space shortages, etc., but moving from the upper to the lower storage device is called archiving, and data sets are saved to the lower storage device. Returning a data set to the upper storage device is called recall.

Conventionally, there is a "migration control method J" described in Japanese Unexamined Patent Application Publication No. 64-21554, which deals with such technology. In the method described in the above publication, recall from a lower storage device to an upper storage device is known. For datasets that have only been referenced without being updated, the archive operation is omitted by validating the previous archive information for re-archiving requests. Set an update flag on a delta set, turn it off when it is recalled, turn it on when the dataset is updated, and archive the dataset only if it is on, but not until it is turned on. This allows for faster archiving.

[Problems to be Solved by the Invention] In the technology described in the above-mentioned publication, when a re-archive request is made for a data set that has been recalled from a lower storage device to an upper storage device and only referenced, the archive cannot be re-archived. Speed-up can be achieved.

However, for datasets that have been updated after being recalled, the conventional dataset archiving operation is performed, so there is a problem in that archiving cannot be accelerated in systems where datasets are frequently updated.

An object of the present invention is to solve such conventional problems and to provide an external storage device that can reduce the movement of data sets and efficiently allocate data sets even when data sets are frequently updated. The object of the present invention is to provide a dataset allocation method.

[Means for Solving the Problems] In order to achieve the above object, the data set allocation method for an external storage device according to the present invention includes (a) managing the data set area on the external storage device in units of continuous spaces (extents); , when allocating a data set on an external storage device, if there is insufficient space on the external storage device, the data set can be moved by moving an extent with sufficient capacity to resolve the space shortage to another external storage device. The feature is that a new data set is allocated to the space area after (b) When allocating data sets on an external storage device, if there is insufficient space on the external storage device, continuous space for less frequently used datasets ( The feature is that by sequentially moving data sets from one extent to another external storage device, a new data set is allocated to the moved space area. Also,
(c) The extent to be moved to another external storage device is
In the case of an entire data set on an external storage device, the data set is moved by parallel processing in units of extents that constitute the data set.

In addition, (d) if the extent to be moved to another external storage device is a data set with multiple volumes, the unit of movement is a group of extents within the volume, and the movement is performed volume by volume in parallel. There is. Furthermore, (e) when moving a data set on an external storage device to another external storage device, the other external storage device is a storage device for eviction with lower performance and cost than the above external storage device; When it becomes necessary to access a data set having an extent that has been deleted, the feature is that only the extent is returned to the external storage device to access the extent.

[For production]

In the present invention, (a) the allocated area of a data set is managed in units of continuous areas (extents), and if a lack of space occurs when allocating a data set on an external storage device, the space shortage on that storage device is The amount of data to be moved is reduced by moving only extents large enough to solve the problem to another external storage device. If a space shortage occurs and a data set that is infrequently used on that external storage device is moved to another storage device, only the extents on that storage device are moved in parallel at the same time. This allows free space on the storage device to be quickly set up. (b) Also, when moving the entire dataset on an external storage device to another external storage device, the time required to move the dataset can be reduced by performing the migration process in parallel for each extent that makes up the dataset. can be done. Further, when moving a data set having a plurality of volumes, the time required to move the data set can be reduced by performing the movement process in parallel with the movement process for each volume.

(c) In addition, if the performance and device attributes of the destination external storage device are different from those of the original external storage device, re-migration of only the moved extents to the external storage device with the same performance and device attributes will save the data. Moving a set can be completed quickly, and the data set can be accessed in the same way as before. The same applies when the extent of part of a volume of a data set with a plurality of volumes exists in the external storage device for expulsion. (d) In addition, if a metal stent of a data set with multiple volumes exists in the expulsion external storage device, by simultaneously executing the remigration process to the original external storage device for each volume extent, Dataset migration can be completed quickly and dataset access can be maintained as before. In this way, data sets can be allocated efficiently.

[Example] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is an overall configuration diagram of a computer system showing an embodiment of the present invention.

In FIG. 1, 1 is a data processing device, 9 is an upper storage device, 13 is a lower storage device, 17 is a TSS (time sharing system) terminal, 18 is a batch input device, and 15 is a management device that stores an extent management table 16. It is a book file.

Both the upper external storage device 9 and the lower storage device 13 are connected to the data processing device l, and the management book file 1
5 is also connected to the data processing device 1.

The upper storage device 9 has volumes 10, 10a+10b, etc. to which files are allocated according to performance, capacity, and cost.
・ is stored. Also, in the lower storage device 13, there is a volume l for expelling files.4. 14a,
14 b... are stored. The extent management table 16 is a data set management list in order to manage the missing extent 12 of the data set 11 allocated to the volume 10 of the upper storage device 9, the extent (A, ) 12a, and the extent (B, ) 12b of the data set A11a. It is stored in 15. When a data set access request is received from the TSS terminal device 17 or the batch input device 18, the data processing device 1 refers to the extent management table 16 to efficiently use the upper storage device 9.

The data processing unit 2 of the data processing device 1 includes a file allocation processing unit 3 that allocates files, a space allocation processing unit 4 that allocates space in the upper storage device 9 to form an extent 12, and an upper storage device 9. An archive that moves the extent 12 of the dataset 11 to the lower storage device 13, and the lower storage device 13
an archive/recall processing section 5 for executing recalls to move data sets 11 from to the upper storage device 9, a migration processing section 6 for moving extents 12 of the data set 11 to another volume in the upper storage device 9, and extents between volumes. The data storage system is comprised of an extent migration processing unit 7, which performs migration, and a management book control s8, which manages a management book 15 that stores an extent management table 16 that manages data sets.

The management list control unit 8 of the data processing unit 2 manages the dataset allocation area on the upper storage device 9 in units of continuous areas (extents), so allocates a new delta set on the upper storage device 9. Sometimes, when a space shortage occurs, the system selects an extent large enough to eliminate the space shortage and moves only that extent to the lower storage device 13. When the extent movement is completed, it operates to update the extent status in the data set management list 15.

Also, when there is insufficient space on the upper storage device 9,
It searches for a data set on the upper storage device 9 that is used less frequently, and operates to move only the extent on the upper storage device 9 to the lower storage device 13 in the data set area. Furthermore, when there are multiple extents to be targeted, the migration process is performed in parallel for each extent.

When the extent movement is completed, the data set management list 15 is updated. This makes it possible to reduce the amount of data to be moved.

Furthermore, when moving the data set on the upper storage device 9 to the lower storage device 13, control is performed so that the migration process can be performed in parallel in units of extents that constitute the data set. Data set migration is recognized when all extents have been moved, and the data set management list 15 is updated. This makes it possible to shorten the data set moving time.

In addition, if the data set on the upper storage device 9 is to be moved to a data set with multiple volumes, the movement unit is controlled to be a group of extents within the volume, and each time the movement process for each volume is completed, the data set is moved. It operates to update the management list 15. The migration of the data set is recognized when all the migration of the target volume's negative stent is completed, and the data set management list 15 is updated.

Additionally, if there is insufficient space on the upper storage device 9,
When moving data set extents to eliminate the cause of lack of space, lower-level storage devices 13 with low performance are
By preparing for eviction, only extents of data sets that are infrequently used are preferentially evicted to the lower storage device 13 for eviction. When the extent movement is completed, the data set management list 15 is updated.

Also, in accessing the evicted dataset,
It operates to reload only the extent evicted from the upper storage device 9 for data set eviction onto the lower storage device 13, and operates to update the data set management list 15 when the extent movement is completed. do. Thereby, it is possible to realize conventional access by partially reloading the data set on the upper storage device 9.

FIG. 2 is a diagram of a management table for managing area allocation within a volume in FIG. 1.

FIG. 2(a) is a storage format diagram of volume 10,
Figure 2 (b) is a diagram showing the VTOC area in (a);
Figure (c) is a format diagram showing the format ID5CB in (b). Here, the VTOC area means Volume
It is a table of contents that collects information that manages datasets included in a disk volume (dataset name, dataset attribute, dataset location, etc.) and information about the volume (information on free space within the volume, etc.) It is something that

The standard volume label 10v includes the volume name (volume serial number) and VT, as shown in Figure 2(a).
This is a table for managing the OC area 19.

The VTOC region 19 is as shown in FIG. 2(b)G:
Manage VTOC area 19 format 4DSCB l 9a
1 Format 5DSCB 19 b for managing free space, Format ID 5C820 for managing allocated data sets
Consists of etc. Here, DSCB is Data
Abbreviation for Set Control Block, format ID5CB, format 4DSCB, format 5DSC
Each of the B formats has its own unique format. - As an example, the format of Type I DSCB is
Shown in Figure 0.

When allocating a new data set, the file allocation processing unit 3 determines the volume and calls the space allocation processing unit 4. Space allocation processing unit 4
is a format 5DSCB that manages free space within a volume.
l 9 b is input, and if there is a free area that satisfies the request, the free area is allocated to the data set, format 5DSC, B19b is updated, and format I DSCB 20 is created.

The format ID5CB20 for managing datasets has a second
As shown in Figure (C), data set name 20a, creation date 20b, last reference/update date 20C1 allocated extent information 20d, start address 20e, end address 20f, number of allocated tracks 20g, etc. are stored. The last reference/update date 20c in the format ID5CB is the date set when the data set is re-opened after creation.

FIG. 3 is a format diagram of the extent management table in FIG. 1.

In Figure 1, extent 1 of dataset 11
2. Extent (A,) of dataset (A) lla
12a and an extent management table 16 for managing extents (B,) 12b. This extent management table 16 is a central management table for data set movement in the present invention. This table is
Dataset name 16 for each assigned dataset
a, an extent number 16b, a volume name 16c in which the extent exists, an extent number 16d within the volume 1, and an extent status 16e. In response to extent movement in the archive/recall processing unit 5, migration processing unit 6, and extent movement processing unit 5, the volume name in which the extent exists, the extent number within the volume, and the status of the extent are updated, and the data Used for set access. The extent status 16e, ARCI VE, indicates that the extent has been evicted to a volume on the lower storage device 13. Here, data set AAA occupies three volumes and belongs to extent number 01 within each volume. Among these, volume name ARCIV
only has been evicted to the lower storage device 13.

Note that the size of each extent (that is, the start and end addresses of the extent, and the number of tracks) is managed by the format l DSCB 20 in the VTOC area 19 shown in FIG.

FIG. 4 is a format diagram of an extent migration parameter list in the present invention.

The extent migration parameter list is a table expanded on the main storage device, and the extent migration processing unit 7
This is an area for storing input information/output information.

Since the archive/recall processing unit 5 recognizes the volume/position within the volume of the data set to be moved, it stores that information (migration source and destination extent information) in this extent migration parameter list. , and is passed to the extent movement processing unit 7.

In this way, the extent migration parameter list 21 is a list used when the archive/recall processing section 5 and the migration processing section 6 call the extent migration processing section 7 for extent migration.

As shown in FIG. 4, the negative stent migration parameter list 21 includes migration source extent information 22, migration destination extent information 23, and return code 24. The migration source extent information 22 includes the volume name 22a where the extent exists and the extent position 23.
The migration destination missing stent information 23 is composed of a volume name 23a where the extent exists and an extent position 23b. The extent migration processing unit 7 executes extent migration according to the migration source extent information 22 and the migration destination extent information 23, and sets the processing result in the return code 24.

FIG. 8, FIG. 9, and FIG. 5 are flowcharts of space allocation processing when allocating a new data set, respectively, and show the processing when there is no free space in the volume that satisfies the request. Of these, FIG. 8 shows the space allocation flow for the allocated volumes by the space allocation processing unit 4 in FIG. 1.

In order to obtain free space information within the volume, the space allocation processing unit 4 first inputs the format 5DSCB 19b (step 41), and checks whether the requested free space exists based on the input format 5DSCB 19b. (Step 42). When the requested free space is available, update the format 5DSC319b to format ID 5.
CB20 is created (step 45), thereby creating the extent management table 16 shown in FIG.
Step 46), the process ends. In step 42, if there is no free space that satisfies the request, the archive/recall processing unit 5 is called (step 43).
. Check whether the archive/recall processing unit 5 has been able to secure free space that satisfies the request (step 4).
4). If a free area that satisfies the request can be secured, execution starts again from step 41. Also, if free space cannot be secured, an error will occur.

FIG. 9 is a flowchart of the archive processing of the archive/linyl processing section in FIG. 1 and the DSCBSC-
It is a connection diagram.

DSCBSC- is a queue temporarily secured in the main memory at the time of DSCBSC.

As shown in FIG. 9(a), the archive/recall processing section S first inputs the format ID5CB20 in order to secure a free space in the requested volume (step 50). It is checked whether the data set of the input format I DSCB 20 is in use (step 51). Whether or not a data set is in use is determined by exclusive control based on the data set name when the data set is used.

If the corresponding data set is not in use, it is registered in the DSCB queue 70 shown in FIG. 9(b) (step 52) 6. When connecting to the DSCB queue 70,
They are registered in the order of the oldest last reference/update date 20c, and in the order of the largest allocated area when the last reference/update date 20c is the same. This makes it possible to evict extents that are accessed infrequently and are useful for resolving space shortages.

It is checked whether all format ID5CB20 have been input, and if not, steps 50 to 53 are repeatedly executed (step 53).

After the input is completed, it is checked whether the DSCB queue 70 has been created (step 54). If the 6DSCB queue 70 has not been created, the process branches to step 64 and returns as there is no request available. On the other hand, DSCB
When the queue 70 is created, the corresponding format I
Based on the D5CB information 2°, the extent information 20d is extracted and stored in the source extent 22 of the extent migration parameter list 21 shown in FIG. 4 (step 55). Note that if it is composed of multiple extents, the source extent 22 is moved by the number of extents.
Store in. After setting the migration source extent information 22, the area corresponding to the migration source extent is transferred to the lower storage device 1.
3 (step 56). Allocating an area to the lower storage device 13 is performed in the same process as space allocation to the upper storage device 9. After allocating the area to the lower storage device 13, move the area allocated to the lower storage device 13 to the migration destination extent 2 in the extent migration parameter list 21.
3 (step 57).

After setting the extent migration parameter list 21 shown in FIG. 4, the extent migration processing unit 7 is called and archive processing is requested (step 58). After returning from the extent migration processing unit 7, the return code 24 of the extent migration parameter list 21 is checked (step 59). If the extent migration fails, the process branches to step 63 to proceed to the processing of the next entry.

Furthermore, if the extent movement is successful, the extent management table 16 corresponding to the relevant extent is updated (step 60). Additionally, format 5DSCB 1
9b, the format ID5CB20 is updated, and it is assumed that the corresponding extent has been deleted from the volume (step 61). Format 5DSCB19b, Format I DSCB2
After updating 0, a check is made to see if a free area that satisfies the requested amount has been secured (step 62). If free space that satisfies the requested amount cannot be secured, the next DSCB
Check whether there is an entry (step 63)
. If there is a next DSCB entry, step 55
Repeat steps 62 to 62. If there is no next DSCB entry, it is assumed that the requested free area could not be secured and the process returns (step 64). On the other hand, if the free space that satisfies the requested amount can be secured in step 62, it is determined that the requested free space has been secured and the process returns (step 65).

FIG. 5 is a flowchart of extent movement of a data set by the extent movement processing section shown in FIG.

The extent migration processing unit 7 extracts the migration target extent 22 stored in the extent migration parameter list 21 shown in FIG. 4 and the extent 23 of the migration destination storage device 13.
is recalled from the archive/recall processing section 5 using as input information.

When moving extents of a dataset, first, the number of parallel processing tasks is determined according to the number of target extents obtained from the extent movement parameter list 21, and processing tasks are started in parallel (step 32) and executed (step 33.33,). This execution is to perform the extent migration task. When the movement of each extent is completed, a return code is set in the return code area 24 of the negative stent movement parameter list 21 (steps 34, 34.). Extent movement completion management (step 3)
4) is recognized as the completion of all processing tasks, and the extent movement is completed.

FIG. 6 is an explanatory diagram showing the extent management table changing process according to the present invention. FIG. 6(a) is a diagram showing the operation of the extent migration processing unit 7 when there is a shortage of space, and FIG. 6(b) is a diagram showing the extent management table updating process before and after processing.

Here, a description will be given of how to change the extent management table 16 in the case where only some stents in the data set are moved. Assume that a space shortage has occurred in the data cell h allocation to the volume (B) 10b of the upper storage device 9.

In Figure 6(a), due to lack of space, the volume (B
When extent (B, ) 12b of ) 10b is targeted for archiving, archiving is executed as shown in the processing flow of FIG.

The archive/recall processing unit 5 requests the extent migration processing unit 7 to move only the extent (B, ) 12b of the data set (A) lla. The extent movement processing unit 7 moves the extent (B,) 12
b in the lower storage device 1 secured in step 56 of FIG.
The task 33 for moving the volume (a) to the area of the volume (a) 14a of No. 3 is activated.

After the migration is completed, as shown in FIG. 6(b), for the entry whose extent number 16b is 02 in the data set (A) 11a of the extent management table 16, the volume name 16c and the extent number 16d within the volume are changed.
, the extent state 16e, respectively <a><OD
Update <ARCIV> to complete the archive process.

FIG. 7 is an explanatory diagram of a migration operation of a data set having a plurality of volumes.

Here, all the niggestents of the data set (A)
2a, Volume (B) 10 b Nigestent (B
, ) 12b) are moved to volume (C) 10c and volume (D) lOd, respectively.

The migration processing unit 6 sends the extent (A,
) 12 a, volume (B) iob extent (
B) 12b, volume (C) 10C, volume (
D) Request to move to 10d. The negative stent movement processing unit 7 performs a negative stent processing task 33 on the extent (A,) 12a of the volume (A) 10a.
At the same time, extent processing task 33. is started for extent (B) 12b of volume (B) 10b. Start. In this manner, the extent migration processing unit 7 attempts to speed up migration by performing parallel migration processing.

The embodiments have been explained above with reference to the drawings, but in the present invention, (the allocation area of the internal data set is managed in units of extents, which are continuous areas, and when there is a shortage of space when allocating the data set on the external storage device) When this occurs, the amount of data to be moved is reduced by moving only extents large enough to resolve the lack of space on the external storage device to another external storage device. Allocation can be done efficiently.In addition, when a space shortage occurs, when moving a data set on an external storage device that is infrequently used, for example, a data set with an old last reference date in the VTOC of a volume, to another external storage device, By simultaneously moving only the extents on the external storage device in parallel in extent units, free space on the external storage device can be set at high speed.

(b) Also, when moving the entire dataset on an external storage device to another external storage device, the time required to move the dataset can be reduced by performing the migration process in parallel for each extent that makes up the dataset. can be done. Further, when moving a data set having a plurality of volumes, the time required to move the data set can be reduced by performing the move process in parallel for each volume.

(c) If the destination external storage device has different performance and device attributes from the original external storage device, the data set can be Moves can be completed quickly and data set access can be performed in the same way as before.

(d) Furthermore, even if only the extent of a part of a volume of a data set having a plurality of volumes exists in the external storage device for expulsion, the migration of the data set can be completed at high speed in the same manner.

(e) If all extents of a data set with multiple volumes exist in the external storage device for expulsion, the data set can be moved to the original external storage device simultaneously for each extent of the volume. Movement can be completed quickly.

〔Effect of the invention〕

As explained above, according to the present invention, when archiving a dataset, the entire dataset is not saved.
For example, by moving/evacuating only the data set area on the storage device that eliminates the cause of archiving such as lack of space, or by moving in parallel every two stents that are continuous spaces that make up the data set.
Even when datasets are updated frequently, archiving performance can be improved and datasets can be allocated efficiently.

[Brief explanation of drawings]

FIG. 1 is an overall configuration diagram of a computer system showing an embodiment of the present invention, FIG. 2 is a diagram showing a table in a volume of a storage device in FIG. 1, and FIG. 3 is a diagram showing a table in a volume of a storage device in FIG. 4 is a format diagram of the extent migration parameter list for the extent migration processing section of FIG. 1, FIG. 5 is an operation flowchart of the extent migration processing section of FIG. 1, and FIG. The figure is an explanatory diagram of a method for resolving space shortage by moving data set extents according to the present invention, 7th rgJ is an explanatory diagram of the case where a data set with a multiple volume configuration is moved to another external storage device according to the present invention, and Fig. Operation flowchart of the space allocation processing unit in the figure, No. 9r! 7J is a flowchart of the archive processing of the archive/recall processing section in FIG. 1, and FIG. 10 is the format I used in the present invention.
It is a format diagram of DSCB. 1: data processing device, 2: data processing unit, 3: file allocation processing unit, 4 space allocation processing unit, 5: archive/recall processing unit, 6: migration processing unit,
7: Extent movement processing unit, 8: Management list control unit, 9:
Upper storage device, 10° 10a, 10b Volume of two upper storage devices, lOv: standard volume label, 11. l
la: allocation data set, 12, 12a, 12b:
Extent of dataset, 13: Lower storage device, 1
4, l 4 a, 14 b: Lower memory bag f
(IJ tr IJ Neem, 15: Management book file,
16: Data set extent management table, 17
+TSS terminal device, 18: Batch input device, l 9
: VTOC area, 19a: Format 4DSCB, 19b:
Format 5DSCB, 20: Format ID 5CB, 20a: Data set name, 20b: Creation date, 20c: Last referenced/updated date, 20d: Extent information, 20e: Extent start address, 20f: Extent end address, 20g: Extent allocation track number,
21: Extent movement parameter list, 22: Source extent, 22a: Volume name, 22b: Extent position, 23: Destination extent, 23a: Volume name, 23b: Extent position, 24: Return code, 70: DSCB queue diagram (Part 1) (a) Figure (Part 2) (b) VTOC area diagram (Part 2) (b) ↓ Extent management table (after processing) Figure Space allocation diagram (Part 2) (b) DSCB Cu

Claims (1)

[Claims] 1. In a computer system comprising a central processing unit and a plurality of external storage devices connected to the central processing unit, a data set area on the external storage device is divided into continuous spaces (extents). When managing and allocating datasets on the external storage device, if a space shortage occurs in the external storage device, move extents of sufficient capacity to eliminate the space shortage to another external storage device. 1. A data set allocation method for an external storage device, comprising the step of allocating a new data set to a space area after the movement. 2. In the data set allocation method for an external storage device according to claim 1, when a data set is allocated to the external storage device, if a space shortage occurs in the external storage device, the space shortage is resolved. The feature is that data sets are sequentially moved from contiguous spaces (extents) of less frequently used data sets to another external storage device until the capacity is reached, and new data sets are allocated to the space area after the data have been moved. Data set allocation method for external storage devices. 3. In the data set allocation method for an external storage device according to claim 1 or 2, when the extent to be moved to the other external storage device is the entire data set on the external storage device, the data set is A data set allocation method for an external storage device, characterized in that data sets are moved by parallel processing in constituent extent units. 4. In the data set allocation method for an external storage device according to any one of claims 1 to 3, when the extent to be moved to the another external storage device is a data set with a plurality of volumes, the unit of movement is A data set allocation method for an external storage device, characterized in that a group of extents within the volume are moved by parallel processing in units of volumes. 5. In the data set allocation method for an external storage device according to any one of claims 1 to 4, when the data set on the external storage device is moved to another external storage device, the another external storage device is This is a storage device for eviction that has lower performance and cost than an external storage device, and if it becomes necessary to access a data set that has an extent that has been evicted, only that extent can be transferred to the external storage device again. A data set allocation method for an external storage device, characterized in that the extent is accessed by returning the extent.