JPH0675715A - Collective magnetic disk device - Google Patents

Abstract

PURPOSE:To provide a collective magnetic disk device which can improve the availability of a cache and also can shorten a format writing processing time. CONSTITUTION:An invalid data generating part 9 is provided to generate the invalid data 5b which become insignificant in a format writing execution state among those data included in a single logical track. A control part 3A stores only the valid data 5a that become significant in the format writing execution state into a cache 4 and then instructs the part 9 to generate the invalid data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention divides CKD data for one logical track, which is input from a computer main body, and stores the divided CKD data in a plurality of data disks. , Simply referred to as parity) in a parity disk, and in particular, the data that is temporarily stored in the cache memory (hereinafter referred to as the cache) only at the time of executing a format write is limited to valid data to improve cache utilization efficiency. The present invention relates to an aggregate type magnetic disk device which is improved and the format / write processing time is shortened.

[0002]

2. Description of the Related Art Conventionally, as a interface between a computer and a storage means such as a magnetic disk, a variable length format CKD (count key data) in which a recording format on a track can be arbitrarily defined by software.
A collective type magnetic disk device using a format is used.

In this type of collective magnetic disk device, the data length is arbitrarily set at the time of executing the format write, and the CKD format data transferred from the computer is temporarily stored in the cache and then divided into a plurality of units. Store on a data disk. Further, an array disk (RAID) device having a redundancy function is constructed by generating parity from the divided data and storing it in the parity disk.

FIG. 3 is a block diagram showing a conventional collective type magnetic disk device for storing CKD format data with a redundant function. In FIG.
Is a collective type magnetic disk device or disk subsystem connected to the computer main body 1. The disk subsystem 2 includes a control unit 3 that processes data input from the computer main body 1, a cache 4 that temporarily stores data 5 for one logical track under the control of the control unit 3, and a cache unit included in each data 5. It is composed of a parity generation unit 6 that generates parity based on data, a plurality of data disks 7 that individually store each divided data, and a parity disk 8 that stores the generated parity.

5a is valid data in the CKD format data 5 at the time of format write, that is, meaningful data, 5b
Is invalid data in data 5 which is meaningless at the time of format writing. The control unit 3 temporarily stores the data 5 transferred from the computer main body 1 in the cache 4 and divides the data 5 in the cache 4 into individual data. The parity generation unit 6 individually stores the divided data in each data disk 7, and also generates a parity from the divided data and stores the parity in the parity disk 8.

FIG. 4 is an explanatory diagram showing one logical track of variable-length format (CKD format) data 5 used in a general-purpose computer. In the figure, IDX1 and IDX2 are indexes of start and end points corresponding to the rotation direction of the disc, HA is management information of the entire track, R0 is a special record consisting of a count field C and a data field D, and R1 and R2 are These are records for normal data recording, each consisting of a count field C, a key field K and a data field D. Here, the case of having three records R0 to R3 is shown.
Any number of records after R1 are used.

The count field C records the position information of the track, its record number, the data length of the subsequent key field K and data field D, and the like, and the key field K records the key for retrieving the record. In the data field D, the data that the software wants to store is recorded.

Next, the operation of the conventional collective type magnetic disk device shown in FIG. 3 will be described with reference to FIG. First, when a write instruction is given from the computer main body 1 to the disk subsystem 2, the control unit 3
Get data from computer 1 and cache
Data 5 for one logical track is stored above.

Subsequently, at a certain timing in accordance with the processing sequence of the control unit 3, while dividing the data 5 on the cache 4 and generating the parity by the parity generation unit 6, the divided data and the parity are respectively divided into data disks. 7 and parity disk 8. This writing process is performed by the parity generation unit 6 under the control of the control unit 3.

Generally, the write processing can be roughly classified into two types, update write and format write. The update write is a write process that rewrites only the data field D in the record and does not affect the subsequent record at all, but the format write wants the software to change the format on the track to record the data. This is a write process sometimes used, and records after the valid data are erased.

For example, when the format write is performed on the record R1, the record R2 and the subsequent records are erased. In this way, the CKD format data 5 is erased from the last record written by the format write to the end point index IDX2.

Now, when the write instruction from the computer main body 1 is a format write, the data 5 on the cache 4 has a significant valid data 5a portion and a meaningless invalid data 5b portion as shown in FIG. You can At this time, in order to divide the data 5 for one logical track and to generate the parity of each divided data, the data 5 including the valid data 5a and the invalid data 5b is placed in the cache 4, and the division and Parity generation is performed.

[0013]

As described above, the conventional set type magnetic disk device, in the case of erasing the last record including the valid data 5a (the record including the invalid data 5b) at the time of executing the format write, In order to generate the parity by dividing the data 5, not only the valid data 5a but also the erased invalid data 5b must be stored in the cache 4, and a meaningless area is required in the cache 4, resulting in utilization efficiency. However, there was a problem that
Further, since the extra work of writing the invalid data 5b in the cache 4 occurs, there is a problem that the format write process takes a lot of time.

The present invention has been made to solve the above problems, and an object of the present invention is to obtain an aggregate type magnetic disk device which improves the utilization efficiency of the cache and shortens the format / write processing time. And

[0015]

SUMMARY OF THE INVENTION A collective magnetic disk drive according to the present invention is provided with an invalid data generation unit for generating invalid data which is meaningless at the time of performing a format write among data included in one logical track. The control unit stores only valid data that is valid at the time of executing the format write in the cache and causes the invalid data generation unit to generate the invalid data.

[0016]

According to the present invention, invalid data is not stored in the cache at the time of executing the format write, but is transmitted from the invalid data generating unit outside the cache.

[0017]

【Example】

Example 1. Embodiment 1 of the present invention will be described below with reference to the drawings. First Embodiment FIG. 1 is a block diagram showing a first embodiment of the present invention, 2A and 3A correspond to a disk subsystem 2 and a control unit 3, respectively, and 1, 4, 5a and 6 to 8 are the same as those described above. Is. The format structure of the data 5 for one logical track is as shown in FIG.

In this case, only valid data 5a which is valid at the time of executing the format write is stored in the cache 4. Reference numeral 9 denotes an invalid data generation unit that generates invalid data 5b, which is meaningless at the time of executing the format write, of the data 5.

Next, the operation of the first embodiment of the present invention shown in FIG. 1 will be described with reference to FIG. First, similarly to the above, when a write instruction is issued from the computer main body 1 to the disk subsystem 2A, the control unit 3A
Takes in data from the computer main unit 1, temporarily stores the data in the cache 4, divides the data 5 for one logical track at a certain timing, and causes the parity generation unit 6 to generate the parity. Are written to the data disk 7 and the parity disk 8, respectively.

However, when the write processing is format write, the control unit 3A caches only valid data 5a that has meaning at the time of format write, excluding meaningless invalid data 5b at the time of format write. Store on top. And valid data at a certain timing
Only 5a is divided and the parity is generated via the parity generation unit 6.

The control unit 3A also causes the invalid data generation unit 9 outside the cache 4 to generate invalid data 5b, and causes the parity generation unit 6 to generate the parity of each invalid data. As a result, data division and parity generation are performed on the data 5 for one logical track including the valid data 5a and the invalid data 5b, and the obtained data and parity are stored in the data disk 7 and the parity disk 8, respectively. .

As described above, since meaningless invalid data 5b is not stored in the cache 4, a larger amount of valid data 5a can be stored in the cache 4 and the utilization efficiency of the cache 4 is improved. Further, since it is unnecessary to write the invalid data 5b to the cache 4, the processing time at the time of executing the format write is shortened.

Example 2. In the first embodiment, the parity generation unit 6 also generates the parity for the invalid data 5b. However, if the data content of the invalid data 5b is erased (all defined by 0 pattern), the invalid data 5b
It can be seen that the parity of b need not be changed at all.

FIG. 2 is a block diagram showing a second embodiment of the present invention, and 2B and 3B correspond to the disk subsystem 2A and the control unit 3A, respectively. In this case, cache 4
Only the valid data 5a above is sent to the parity generation unit 6 and contributes to parity generation, and the invalid data 5b from the invalid data generation unit 9 is sent to the data disk 7 and the parity disk 8 as it is.

In FIG. 2, when the write instruction from the computer main body 1 is format write, only the valid data 5a is stored in the cache 4 as in the first embodiment, and the valid data 5a is divided at a certain timing. And is sent to the parity generation unit 6. The invalid data 5b is generated by the invalid data generating unit 9.

At this time, the invalid data 5b is directly sent to the data disk 7 and the parity disk 8 without being sent to the parity generator 6, and the data and parity are written as they are to the data disk 7 and the parity disk 8. The parity generation unit 6 also generates a parity for only the valid data 5a, and writes the obtained data and parity in the data disk 7 and the parity disk 8, respectively.

Thus, also in the case of the second embodiment, the first embodiment
Similarly, since the invalid data 5b is not stored in the cache 4, the utilization efficiency of the cache 4 is improved and the processing time at the time of executing the format write is shortened.

[0028]

As described above, according to the present invention, the invalid data generating section for generating invalid data which becomes meaningless at the time of performing the format write among the data included in one logical track is provided, and the control section Only valid data that is valid at the time of format write is stored in the cache, and invalid data is generated from the invalid data generator. Therefore, the cache usage efficiency is improved and the format write processing time is shortened. There is an effect that a collective type magnetic disk device can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

FIG. 3 is a block diagram showing a conventional collective type magnetic disk device.

FIG. 4 is an explanatory diagram showing a format of general CKD format data.

[Explanation of symbols]

 1 Computer Main Body 2A, 2B Disk Subsystem 3A, 3B Control Unit 4 Cache 5 1 Logical Track Data 5a Valid Data 5b Invalid Data 6 Parity Generator 7 Data Disk 8 Parity Disk 9 Invalid Data Generator

Claims (1)

[Claims] 1. A CKD input from a computer body
A control unit that temporarily stores the data of the format on the cache and also divides the data on the cache individually, and stores the divided data individually on a plurality of data disks, and from the divided data. In a set type magnetic disk device including a parity generation unit for generating parity and storing it in a parity disk, invalid data that is meaningless at the time of format write is executed among data included in one logical track. And an invalid data generating unit that stores only valid data that is valid when the format / write is executed in the cache, and causes the invalid data generating unit to generate the invalid data. Collective type magnetic disk device.