Array disk subsystem

Details Array disk subsystem Array disk subsystem

1994-12-09

2001-06-05

Tu, Christine T. (Department: 2133)

Error detection/correction and fault detection/recovery

Data processing system error or fault handling

Reliability and availability

Reexamination Certificate

active

06243824

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates to a secondary storage subsystem including a secondary storage controller and a secondary storage in which the secondary storage subsystem is connected to a host such as a central processing unit (CPU) to store therein data necessary for the host and to restore the data when required. More specifically, the present invention relates to an array disk subsystem having a plurality of secondary Storage controllers and a plurality of secondary storages wherein the subsystem is capable of achieving ordinary input/output Processing even while a backup processing is being executed. The backup processing here is accomplished to store information of the secondary storage in another store means for a backup use, which is effective for an acquisition of data when an access to the data stored in the subsystem is prevented due to a failure such as an erroneous operation conducted by an operator, a software error, or a hardware error.
In computer systems at the present stage of art, data necessary for the host is loaded in a secondary storage such that the data is obtained from the secondary storage when a data request is requested from the host. As the secondary storage, a nonvolatile storage is employed in general; typically, a direct access storage device such as a magnetic disk or an optical disk is adopted. Although these secondary storage devices have a considerably high reliability, there may still exist a chance of a failure which prevents data stored therein from being accessed.
In order to cope with the difficulty above, there has already been known a backup processing method in which, as shown in
FIG. 6
, during a period of time when a computer system is not utilized from an external device, data stored in a magnetic disk
11
of a secondary storage is transferred via a disk controller (DKC)
14
to a CPU
1
, which then records the data on a magnetic tape (MT)
12
via a magnetic tape controller (MTC)
15
or on an optical disk (OD)
13
via an optical disk controller (ODC)
16
.
Furthermore, in accordance with the present invention, there is accomplished a parallel processing in which a plurality of disk drives each being associated with a relatively small capacity are prepared so that data transferred from the CPU
1
is split into subdata items to be stored in the respective disk drives. A parallel processing system of this type has already been put to the market, for example, from Storage Concepts Corp. A method of improving the reliability in such a system has been described in “Providing Fault Tolerance In Parallel Secondary Storage ystems” written by Arvin Park and K. Balasubramanian of Department Computer Science, Princeton University (Nov. 7, 1986). The configuration of the system has also been discussed in the report above.
SUMMARY OF THE INVENTION
In the secondary storage of the computer system of the prior art, the data handling on the secondary storage is interrupted while the backup processing is taking place. Due to the development of computer applications in various fields and the growing volume of information, the amount of data to be stored in the secondary storage can be considered to continue increasing. Resultantly, according to the backup processing method of the existing secondary storage, the backup processing takes quite a long period of time, which leads a considerable problem with respect to the utilization efficiency of the computer system. Moreover, in a case of a 24-hour online system continuously running for daily jobs in a bank or a stock brokerage firm, when the backup processing prevents the secondary storage from being used for such a long period of time, the throughput of the system will be remarkably lowered. In this situation, there has been needed a backup processing in which the CPU can conduct an ordinary input/output processing without taking the backup processing into consideration and even while the backup processing is in process.
In accordance with the present invention, there is provided an array disk subsystem including a unit for achieving input and output operations of data items with a host, a unit for subdividing a data item from the host into subdata items, a unit for generating error correction codes (ECCs) from the subdata items, a group of disk units for respectively storing therein the subdivided data items and the ECCs, a backup unit for sequentially achieving a backup operation on the data items in the group of disk units, a unit operative in response to a read command required from the host to the group of disk units including a disk unit under a backup operation for reading a data item and an ECC from one of the disk units not under the backup operation, and a unit for restoring, based on the data item and the ECC thus read from the disk unit, data stored in the disk unit under the backup operation.
Moreover, according to an aspect of the present invention, there is provided a data processing apparatus including a CPU, a unit for subdividing a data item from the CPU into subdata items, a unit for generating ECCs from the subdata items, a group of disk units for respectively storing therein the subdivided data items and the ECCs, a backup unit for sequentially achieving a backup operation on the data items in the group of disk units, a unit operative in response to a read command required from the CPU to the group of disk units including a disk unit under a backup operation for reading a data item and an ECC from one of the disk units not under the backup operation, and a unit for restoring, from the data item and the ECC thus read from the disk unit, data stored in the disk unit under the backup operation.
In addition, in accordance with another aspect of the present invention, there is provided an array disk subsystem including a unit for achieving input and output operations of data items with a host, a unit for subdividing a data item from the host into subdata items, a unit for generating ECCs from the subdata items, a group of disk units for respectively storing therein the subdivided data items and the ECCs, a backup unit for sequentially achieving a backup operation on the data items in the group of disk units, a store unit operative in response to a write command issued from the host to the group of disk units including a disk unit under a backup operation for storing therein a data item to be written in the group of disk units, and a unit for transferring, after the backup operation is completed, the data from the store unit for recording the data item in the group of disk units.
Moreover, in accordance with still another aspect of the present invention, there is provided an array disk subsystem comprising a unit for achieving input and output operations of data items with a host, a nit for subdividing a data item from the host into subdata items, a unit for generating ECCs from the subdata items, a group of disk units for respectively storing therein the subdivided data items and ECCs, a backup unit for sequentially achieving a backup operation on the data items associated with the group of disk units, a unit for monitoring input/output commands from the host, and a unit operative when a frequency of the input/output commands is less than a fixed value for transferring data from the group of disk units to the backup unit to accomplish a backup operation of the data.
A data item sent from the CPU is subdivided into subdata items to produce ECCs for the subdata items and then the ECCs and the associated subdata items are respectively stored in disk drive units. An ECC which is a code generated from a subdata item has information related to the subdata item. When an error occurs in a subdata item, the ECCs and the remaining subdata items thus stored can be used to correct the subdata item. More specifically, a data item transferred from the CPU is split by the data control unit or a data controller into subdata items. An ECC is created for each subdata item such that the ECC and the subdata item are stored in different disk drive units. With this provisi

Affiliated with

Also associated with

Rating

Array disk subsystem does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Array disk subsystem, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Array disk subsystem will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2513965