JP2003330624A - Extra disk assignment system and its method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an extra disk assignment system for dividing a disk group into a sub-disk group and an extra disk group, and for executing input/output load decentralization by the dynamic assembly and restoration of a disk according to the load condition of the sub-disk group. <P>SOLUTION: This preliminary disk assignment method is provided with an assignment means for assigning one physical disk among a plurality of physical disks configuring a work group to which each inputted job is belonging to the job, a means for judging whether or not the mean access frequency of all physical disks in the work group to be accessed by all the inputted jobs belonging to the work group is higher than the upper limit value of the work group for each work group, and a means for adding the extra disk as a physical disk configuring the work group whose access frequency is judged to be higher than the upper limit value. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk allocation method for jobs, and more particularly to a disk allocation method for jobs that dynamically changes allocation according to the frequency of input / output to / from a disk.

[0002]

2. Description of the Related Art An example of a system for allocating a temporary file from a disk device having a performance according to a priority value of a user program is described in Japanese Patent Laid-Open No. 07-084840.

In the disk allocation method for creating a temporary file described in this publication, a table holding electronic disks and magnetic disks grouped into disk groups according to their performance and holding a correspondence table between priority values and disk groups. Since then, the disk of the disk group is assigned to the user program.

[0004]

However, this conventional technique has the following problems. In this mechanism, after determining the disk group from the correspondence table, the optimum disk in the sub disk group is assigned to the user program, but when many user programs use one sub disk group, the sub disk group Since the number of volumes in the disk is fixed, one disk is used among a plurality of user programs, which causes a problem of waiting for input / output and allocation of the disk. The cause of this problem is that the disk configuration in the sub-disk group can use only the preset configuration and the configuration of the sub-disk group cannot be changed during operation.

According to the present invention, a disk group is divided into a sub disk group and a spare disk group, and spare disk allocation in which the input / output load is distributed by dynamically installing and retrieving disks according to the load status of the sub disk group. The purpose is to provide a scheme.

[0006]

According to the present invention, for each job that is submitted, one physical disk out of a work group to which the job belongs and which is composed of a plurality of physical disks is assigned. Means, and for each work group, the average access frequency for all physical disks in the work group that are accessed by all submitted jobs belonging to the work group is higher than the upper limit value for the work group. Means for determining whether or not, and the work group in which the average access frequency is determined to be higher than the upper limit,
And a means for adding a spare disk as a physical disk constituting the work group, and a spare disk allocation method is provided.

In the above-mentioned spare disk allocation method, the average access frequency for all physical disks in the work group that are accessed by all jobs that belong to the work group for each work group is the work group. A means for determining whether the average access frequency is lower than the lower limit value for the group, and a work group for which the average access frequency is determined to be lower than the lower limit value,
And means for deleting the added spare disk.

The above spare disk allocation method may further comprise means for updating the upper limit value.

The above spare disk allocation method may further comprise means for updating the lower limit value.

In the above-mentioned spare disk allocation method, for the work group whose average access frequency is judged to be higher than the upper limit value, when the number of spare disks added to the work group reaches a predetermined value. ,
It is possible to further include a unit that does not add a spare disk as a physical disk configuring the work group.

In the above-mentioned spare disk allocating method, the allocating means defines a means for receiving a job submission request described in a job description statement, and a workgroup corresponding to the type of the job for which the submission request is made. A means for searching a table storing the correspondence between types and work groups; and a means for searching the table storing the correspondence between each physical disk and the work group for the physical disks constituting the searched work group. A means for retrieving the disk name with the lowest I / O count among the retrieved physical disks, and a means for allocating the physical disk with the lowest I / O count for the job that has been submitted. , May be provided.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION [First Embodiment] The present invention relates to a temporary work file (sort,
In a computer system in which a work disk (for work files used for merging) is composed of multiple physical disks, the work disk is set to multiple work disk groups, and one of them is set as a spare disk group. The work disk group's I / O load status (the number of I / Os per hour) is monitored, and disks are dynamically installed from the spare disk group to the work disk group and dynamically restored to the spare disk group. It is possible to improve the operability of the system by distributing the load to the work disk used for the job.

In the embodiment shown in FIG. 1, a job description statement 1 for performing a task such as sorting using a temporary file, a job execution control unit 10, jobs 11, 12, 13, 14 and an input / output mechanism unit 20. And the disk input / output information table 30
A work disk group 40, a work disk group management control unit 50, and a work disk control unit 60 which is a new control unit in the present invention.

The work disk group 40 is a work disk group 4 composed of physical disks 41.
2, 43, 44 and a spare disk group 45 are provided.

The work disk control unit 60 includes a dynamic configuration setting unit 61 and a work disk group management table 62.
And a work disk information table 63.

FIG. 2 shows an example of the work disk group management table 62 and the work disk information table 63. The work disk group management table 62 is
Work disk group name 101 and job type 102
And the maximum number of input / output 103 to the work disk group per unit time and the minimum number of input / output 104, which are referred to / updated by the dynamic configuration setting unit 61. The work disk information table 105 includes a disk name 105, a purpose 106, and a work disk group name 1
07 and an allocation counter 108, which are referenced / updated by the dynamic configuration setting unit 61. In the work disk group management table 62, work groups are assigned for each job type. For example, a priority for each job type is used as the allocation standard.

A work disk group name 101, a job type 102, a maximum input / output count 103, and a minimum input / output count 104 in the work disk group management table 62 are set as environment definitions before job execution. Furthermore, the disk name 10 of the work disk information table 63
5, use 106, and work disk group name 107
Is set in the same way. There are two types of applications 106, “work” and “spare”. When “work” is selected, it is necessary to set a work disk group name.

The job execution control unit 10 issues a disk reservation request to the work disk group management control unit 50 at the start of job execution, allocates the physical disk 41 to each job, and at the end of job execution, the work disk group management control unit 50. It is equipped with a function for making a disk release request.

The input / output mechanism section 20 has a function of performing input / output to / from the physical disk 41 and updating the disk input / output information table 30 when the input / output is performed.
The disk input / output information table 30 is a physical disk 41.
It holds information on the number of inputs and outputs for each.

Work disk group management controller 50
When a disk allocation request is received from the job execution control unit 10, the job allocation information is notified to the dynamic configuration setting unit 61 along with the job type information, and the physical disk information acquired from this is returned to the job execution control unit 10. In addition, when a disk release request is received from the job execution control unit 10, it has a function of notifying the dynamic configuration setting unit 61 of the request together with the disk name information of the disk used in the job.

When a dynamic disk setting request is received from the work disk group management control unit 50, the dynamic configuration setting unit 61 searches the work disk group management table 62 and the work disk information table 63 to find the optimum physical disk information. When the disk is returned to the work disk group management control unit 50 and a disk release request is received from the work disk group management control unit 50, the work disk information table 63 is updated based on this request. Further, the dynamic configuration setting unit 61 operates at regular time intervals, calculates the average number of input / output times of each work disk group based on the disk input / output information table 30, and sets the maximum value set in the work disk group management table 62. It has a function of updating each work disk group information set in the work disk information table 63 by comparing the input / output count and the minimum input / output count.

Next, the operation at the time of executing a job executed in the environment configured by FIGS. 1 and 2 will be described with reference to the sequence chart of FIG. 3 and the flowchart of FIG. FIG. 3 shows the flow of processing from job execution to end, and FIG.
Indicates a processing procedure during job execution.

First, the operation from job execution to completion will be described.

Referring to the sequence chart of FIG. 3,
The job execution control unit 10 that has received the job submission request (step A1) described in the job description 1 (step A2) passes the job type information to the work disk group management control unit 50 to request a disk allocation request. (Step A3), and the work disk group management control unit 50 that has received the request passes the request to the dynamic configuration setting unit 61 (Step A4). The dynamic configuration setting unit 61 searches the work disk group management table 62 using the job type as a key, and obtains the work disk group name corresponding to the job type (step A5). Further, after a plurality of disk names are searched in the work disk information table 63 by using the work disk group name obtained in step A5 as a key (step A6), the disk input / output information table among the disk names searched in step A6 The disk name with the lowest I / O count is searched from 30 (step A7), and the allocation counter of the work disk information table 63 is incremented by 1 (step A8). Thereafter, the work disk group management control unit 50 is notified of the disk name retrieved in step A7 (step A9), and the work disk group management control unit 50 notifies the job execution control unit 10 of the disk name (step S9). A10). The job execution control unit 10 notified of the disk name allocates the physical disk 41 having the notified disk name to the job (step A11) and starts executing the job (step A1).
2).

After the job is executed (step A13), at the end of the job (step A14), the job execution control unit 10 that recognizes the end of the job manages the work disk group together with the disk name used by the job. A disk release request is issued to the control unit 50 (step A1
5) Upon receipt of the request, the work disk group management control unit 50 passes the request to the dynamic configuration setting unit 61 (step A16). The dynamic configuration setting unit 61 decrements the disk allocation counter used by the job in the work disk information table 63 by one (step A17).

Next, the operation during job execution will be described.

The input / output mechanism section 20 operates when the job inputs / outputs to / from the disk and updates the disk input / output information table 30. The dynamic configuration setting unit 61 operates at regular time intervals, monitors the disk input / output information table 30 and performs the following processes (A) and (B) for all work disk groups.

(A) Dynamic incorporation This is a process of incorporating a spare disk into a work disk group, and is performed as follows.

Referring to the flowchart of FIG. 2, the dynamic configuration setting unit 61 acquires a work disk group from the work disk group management table 62 (step B1), and the average number of input / output times of the work disk group including the spare disk. Calculate (total number of I / Os per unit time of each disk / number of disks) (step B2),
This number is compared with the maximum input / output number set in the work disk group management table 62 (step B3).
If the average input / output count is higher than the average number of input / output times, the work disk information table 63 is searched for a disk name for which the usage is set to “spare” and a work disk group name is not set, and a search is performed. The work disk group name is set to the created disk name (step B4).
FIG. 5 shows a disk image example and a work disk information table transition example in this dynamic incorporation.

(B) Dynamic recovery This is a process for recovering the spare disk incorporated in the work disk group to the spare disk group, and is performed as follows.

If it is determined in step B3 that the average I / O count is not greater than the maximum I / O count, the dynamic configuration setting unit 61 determines the average I / O count of the work disk group excluding the spare disk (each disk (Total number of input / output times per unit time / number of disks) is calculated (step B5), and this number is compared with the minimum number of input / output times set in the work disk group management table 62 (step B6). If the average I / O count is lower than the average, the work disk information table 63 is set with the work disk group name as a key, the usage is set to "spare", and the work disk group name is set. The disk name is searched for, and the work disk group name is deleted from the searched disks whose allocation counter is 0 (step B7). FIG. 6 shows an example of a disk image and a transition example of the work disk information table in retrieving.

At the time when the processing of (A) or (B) is finished, if there is an unprocessed work disk group (NO in step B8), the processing is continued for the remaining work disk groups. Yes (Steps B1 to B
7).

[Second Embodiment] In the second embodiment, the maximum number of input / output operations 103 of the work disk group management table 62 is 103.
The difference from the first embodiment is that the value of the minimum input / output count 104 can be dynamically changed during job execution.

In this embodiment, referring to FIG. 7, a schedule table 64 is added to the work disk controller 60. Referring to FIG. 8, the schedule table 64 has a work disk group name 109, a change time 110, a maximum input / output count 111, and a minimum input / output count 1
12 is held, and the change time according to the system operation schedule is set as the environment definition. The dynamic configuration setting unit 61, when the change time of a certain work disk group arrives, the work disk group management table 62.
The maximum I / O count and minimum I / O count of the work disk group are updated to the maximum I / O count and minimum I / O count held in the schedule table 64. As a result, it is possible to perform more precise load adjustment.

[Third Embodiment] In the third embodiment, referring to FIG. 9, in the work disk group management table 62,
The difference is that information on the number of embeddable disks 113 is added, and based on this, the maximum number of spare disks to be installed is limited for each work disk group.

FIG. 10 shows a processing procedure during execution of a job according to the present embodiment. As is apparent from comparison between FIGS. 10 and 4, this processing procedure according to the present embodiment is different from that according to the first embodiment. Steps B9 and B10 are added. By referring to the work disk information table 63, the number of built-in spare disks of the work disk group is obtained (step B9), and this number is compared with the number of embeddable disks 113 (step B10) to install the work disk group. Only when the number of completed spare disks does not reach the number of mountable disks 113, step B4.
Process. As a result, it is possible to prevent the spare disk from being occupied by the work disk group with a high load, and it is possible to achieve a more balanced load distribution.

The job execution unit 10 and the input / output mechanism unit 2
0, the work disk group management control unit 50, and the dynamic configuration setting unit 61 can be realized by the computer reading and executing a program for causing the computer to function as these units. Further, the computer receives the program for causing the computer to function as the job execution unit 10, the input / output mechanism unit 20, the work disk group management control unit 50, and the dynamic configuration setting unit 61 from the network, and executes the program. It can also be realized by

[0038]

EFFECTS OF THE INVENTION In a work disk group composed of a plurality of disks, the I / O status of a disk that changes every moment is monitored, and a spare disk is dynamically allocated to a job to wait for disk allocation or I / O wait. Has been realized. As a result, T of each work
It becomes possible to shorten AT (Turn Around Time).

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a spare disk allocation system according to first and third embodiments of the present invention.

FIG. 2 shows a configuration of a work disk group management table shown in FIG. 1 according to a first embodiment and data examples stored in the table, and a configuration of a work disk information table shown in FIG. 1 and data examples stored in the table. It is a figure.

FIG. 3 is a sequence chart showing an operation from job input to job end in the spare disk allocation method according to the embodiment of the present invention.

FIG. 4 is a flowchart showing a regular operation of the dynamic configuration setting unit shown in FIG. 1 according to the first embodiment.

FIG. 5 is a conceptual diagram showing an operation of dynamically incorporating a spare disk allocation method according to an embodiment of the present invention.

FIG. 6 is a conceptual diagram showing a dynamic recovery operation of the spare disk allocation method according to the embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of a spare disk allocation system according to a second embodiment of the present invention.

8 is a diagram showing a configuration of a schedule table shown in FIG. 7 and an example of data stored in the table.

FIG. 9 is a diagram showing a configuration of a work disk group management table shown in FIG. 1 according to a third embodiment and an example of data stored in the table.

10 is a flowchart showing a regular operation of the dynamic configuration setting unit shown in FIG. 1 according to the third embodiment.

[Explanation of symbols]

1 job description 10 Job execution control unit 11-14 jobs 20 I / O mechanism 30 Disk input / output information table 40 work disk group 41 physical disk 42-44 Work Disk Group 45 Spare disk group 50 Work disk group management control unit 60 Work disk controller 61 Dynamic configuration setting section 62 Work disk group management table 63 Work disk information table 64 schedule table

Claims (13)

[Claims] 1. An assignment unit that assigns, to each submitted job, one physical disk of a work group to which the job belongs and which is composed of a plurality of physical disks; and for each work group, A means for determining whether or not the average access frequency for all physical disks in the work group accessed by all submitted jobs belonging to the work group is higher than the upper limit value for the work group; A spare disk allocation method comprising: a means for adding a spare disk as a physical disk constituting the work group, for a work group determined to have an average access frequency higher than the upper limit value. 2. The spare disk allocation method according to claim 1, wherein, for each work group, all physical disks in the work group that are accessed by all submitted jobs belonging to the work group A means for determining whether or not the average access frequency is lower than the lower limit value for the work group; and a spare disk added to the work group for which the average access frequency is determined to be lower than the lower limit value. A spare disk allocation system, further comprising: 3. The spare disk allocation system according to claim 1, further comprising means for updating the upper limit value. 4. The spare disk allocation system according to claim 2, further comprising means for updating the lower limit value. 5. The spare disk allocation system according to claim 1, wherein, for a work group whose average access frequency is determined to be higher than the upper limit value, the number of spare disks added to the work group is predetermined. A spare disk allocation system, further comprising means for not further adding a spare disk as a physical disk configuring the work group when the value is reached. 6. The spare disk allocating method according to claim 1, wherein the allocating unit receives the job submission request described in the job description statement, and includes the submission request. A means for searching a work group corresponding to the type of the job from a table storing the correspondence between the job type and the work group; and physical disks constituting the searched work group, each physical disk and the work group. Means for searching the table storing the correspondence relationship with the above, a means for searching the disk name having the lowest I / O count among the searched physical disks, and the most input for the job for which the submission request has been made. A spare disk allocation method comprising: a means for allocating a physical disk having a low output frequency. 7. An allocation step of allocating, to each submitted job, one physical disk of a work group to which the job belongs and consisting of a plurality of physical disks, and for each work group, Determining whether the average access frequency for all physical disks in the work group that are accessed by all submitted jobs belonging to the work group is higher than the upper limit value for the work group; For a work group whose average access frequency is determined to be higher than the upper limit value, a step of adding a spare disk as a physical disk constituting the work group is included, and a spare disk allocating method is provided. 8. The spare disk allocation method according to claim 7, wherein for each work group, all physical disks in the work group that are accessed by all submitted jobs belonging to the work group Determining whether the average access frequency is lower than the lower limit value for the work group, and deleting the added spare disk for the work group for which the average access frequency is determined to be lower than the lower limit value. And a spare disk allocation method. 9. The spare disk allocation method according to claim 7, further comprising the step of updating the upper limit value. 10. The spare disk allocation method according to claim 8, further comprising the step of updating the lower limit value. 11. The method of allocating a spare disk according to claim 7, wherein, for a work group whose average access frequency is determined to be higher than the upper limit value, the number of spare disks added to the work group is predetermined. A spare disk allocation method, further comprising the step of not further adding a spare disk as a physical disk that constitutes the work group when the value has been reached. 12. The spare disk allocating method according to claim 7, wherein the allocating step includes a step of accepting a job submission request described in a job description statement, and A step of searching a work group corresponding to the type of the job from a table storing the correspondence relationship between the job type and the work group; the physical disks constituting the searched work group, each physical disk and the work group. A table storing a correspondence relationship between the physical disk and the physical disk having the lowest I / O count among the searched physical disks; Allocating a physical disk having a disk name with a low output count, and Spare disk allocation method to. 13. A program for causing a computer to execute the spare disk allocation method according to any one of claims 7 to 12.