JPH09237163A - Duplex disk storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent data transfer from being interrupted so as not to prolong a read processing time by a reason such as a case where a disk storage device selected at the time of reading allows reading to move to an exchange area within an instructed range and a case it becomes a rotation waiting state owing to a retry in a duplex disk storage device where the same data are written in the two disk storage devices. SOLUTION: When a reading instruction is received from a host device 1, the first SCSI(small computer system interface) disk storage device 9A and the second SCSI disk storage device 9B respectively store independently read data in a first data buffer circuit 7A and the second data buffer circuit 7B, and a means is provided for counting the number of blocks stored in both circuits respectively the number of the blocks which are stored is reported to a DMA control circuit 3 at the point of time when storage is completed for each block. The DMA control circuit 3 permits data of the data buffer circuit stored with data earlier to be transferred red to the host device 11 through a high-order SCSI bus 12.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control system for a dual disk storage device, and more particularly to a control system for reading data from a dual disk storage device.

[0002]

2. Description of the Related Art A disk storage device is a device for storing a large amount of data, and erasing the stored data due to a failure is a fatal problem for a computer system. In order to prevent this problem, by storing the same data in two disk storage devices, even if one disk storage device fails, another normal disk storage device continues normal operation. Duplex disk storage devices are known.

Conventionally, when reading data from a dual disk storage device, a method has been used in which two disk storage devices are divided into primary disks and secondary disks in advance, and a read command is issued only to the primary disk to read data. It was taken.

Further, Japanese Laid-Open Patent Publication No. 3-259321 discloses a technique for shortening the read time by issuing a read command to both of the two disk storage devices to read the data of the disk which is first capable of data transfer. ing.

Further, Japanese Patent Laid-Open No. 3-253933 discloses a memory in a dual disk control device, which is provided with a management table for storing various states of respective dual media, and at the time of reading, a read processing disk is determined by referring to the management table. , A technique is disclosed in which, while performing access according to the state, the next read access disk is changed to the other disk so as to alternately read from both duplex media.

[0006]

In the process of reading data from the conventional dual disk storage device, the data may be read only from a predetermined primary disk,
In the case of Japanese Unexamined Patent Publication No. 3-259321, since the data in the disk storage device that starts data transfer earlier among the two devices is transferred, the data transferred to the host device is 2
Data read from only one of the two disk storage devices.

For this reason, in the selected disk storage device, when there is a block for which replacement allocation processing has been performed in a block in the range in which reading is instructed, or when moving to the replacement area, or when reading fails and rotation is waited for retry. There is a problem in that a read processing time is greatly extended due to a pause time during which data transfer is interrupted due to such a case.

Further, in the case of Japanese Patent Laid-Open No. 3-253933, since it is premised that two disks are alternately used, the effect only appears when the lead is continuously read many times, and the independent single lead has no effect. There is a problem. Furthermore, when moving to the alternate area during reading, or when the reading fails and the rotation waits due to a retry, a pause time occurs in which data transfer is interrupted, and the read processing time is greatly extended. The point that there is is JP-A-3-2
This is similar to the case of 59321.

[0009]

SUMMARY OF THE INVENTION It is an object of the present invention to reduce the processing time for reading data from a dual disk storage device.

[0010]

A dual disk storage device according to the present invention has a first disk storage device and a second disk storage device, and the first disk storage device responds to a data write command of a host device. In the duplicated disk storage device for writing the same data to the second disk storage device, data is independently written from the first disk storage device and the second disk storage device in response to a read command from the host device. It is characterized in that the data in the disk storage device, which has been read out and read out earlier for each block, is transferred to the higher-level device.

In the dual disk storage device of the present invention, when the data transfer of all the blocks of the disk storage device, which has completed the reading of all the blocks in response to the read command, is completed, the read processing is terminated. It is characterized by reporting to.

The dual disk storage device of the present invention has a first disk storage device and a second disk storage device, and the first disk storage device and the second disk storage device respond to a data write command of a host device. In a duplicated disk storage device for writing the same data to a storage device, in response to a read instruction from the host device, a first disk storage device that stores data read independently from the first disk storage device and the second disk storage device The data storage means, the second data storage means, and the data transfer in the duplicated disk storage device are controlled to judge the storage status of the first data storage means and the second data storage means for each block. Then, the connection switching means is instructed to connect the data storage means that has completed the storage earlier and the host interface, and the faster storage The number of blocks that can be stored in the data transfer control means for transferring the received data to the host device and the first data storage means is counted, and the storage is completed when the storage is completed for each block in the first data storage means. The first storage block counting means for reporting the number of blocks to the data transfer control means and the number of blocks stored in the second data storage means are counted, and at the time when the storage is completed for each block in the second data storage means. Second storage block counting means for reporting the number of stored blocks to the data transfer control means, and the first data storage means and the second data storage means in response to an instruction from the data transfer control means It is characterized by comprising the connection switching means for switching a connection with a host interface for connecting to the host device.

By the data transfer control means, the first data storage means and the second data storage means store the data read from the first and second disk storage devices connected to each other. The first and second storage block number counting means report to the data transfer control means the number of blocks that have been stored when the storage in the corresponding first and second data storage means is completed. The data transfer control means determines the storage status of the first and second storage means for each block, and instructs the connection switching means to connect the storage means that has finished storing earlier and the upper interface. , The data that can be stored quickly is transferred to the host device.

Therefore, of the data transferred from the two disk storage devices, the data that can be read earlier is transferred to the higher-level device, and the reading of all blocks of the disk storage device that has completed the reading is completed. At this point, it is possible to report the end of read processing to the host device.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention. In FIG. 1, a duplicated disk storage device 1
Is an upper SCSI (Small Computer S
It is connected to the higher-level device 11 via a systems interfaces) bus 12. The dual disk storage device 1 has an upper SCSI control circuit 2 which is connected to the upper SCSI bus 12 and operates as a target. The first SCSI disk storage device 9A and the second SCSI disk storage device 9B that hold the duplicated data are respectively a first lower SCSI bus 10A and a second lower SCSI bus 1
It is connected to the first lower SCSI control circuit 8A and the second lower SCSI control circuit 8B via 0B. These first subordinate S
The CSI control circuit 8A and the second lower SCSI control circuit 8B operate as an initiator and control the first SCSI disk storage device 9A and the second SCSI disk storage device 9B that operate as targets.

The first data buffer circuit 7A and the second data buffer circuit 7B are connected to the first lower SCSI control circuit 8A and the second lower SCSI control circuit 8B by a data bus, and further connected to the data buffer switching circuit 5. It The data buffer switching circuit 5 is further connected to the upper SCSI control circuit 2 by a data bus and switches the connection of the data bus to the first data buffer circuit or the second data buffer circuit. In FIG. 1, the data bus in the duplicated disk storage device 1 is represented by a double line.

The DMA control circuit 3 comprises an upper SCSI control circuit 2, a first lower SCSI control circuit 8A and a second lower SCS.
It is connected to the I control circuit 8B, the first data buffer circuit 7A, the second data buffer circuit 7B, and the data buffer switching circuit 5 by a local bus, and controls data transfer in the duplicated disk storage device 1. The first data buffer storage block number counting circuit 6A and the second data buffer storage block number counting circuit 6B respectively count the number of blocks of data stored in the first data buffer circuit 7A and the second data buffer circuit 7B, DMA that information
Report to the control circuit 3.

The microprocessor unit 4 is an upper S
CSI control circuit 2, first lower SCSI control circuit 8A, second lower SCSI control circuit 8B, DMA control circuit 3, first
It is connected to the data buffer storage block number counting circuit 6A and the second data buffer storage block number counting circuit 6B by a local bus and controls the operation of the entire duplicated disk storage device 1.

Next, the data read operation in the dual disk storage device of the present invention will be described in detail with reference to FIG.

When the upper SCSI control circuit 2 receives a read command from the upper device 11, the first lower SCSI
The control circuit 8A and the second lower SCSI control circuit 8B are connected to the first SCSI disk storage device 9 respectively.
A, read to the second SCSI disk storage device 9B
Issue a command to request reading of data. Second
Shows a specific example when a read of 10 blocks is requested. First SCSI disk storage device 9A, second SCS
The data read from the I disk storage device 9B is the first lower SCSI control circuit 8A and the second lower SCS, respectively.
After passing through the I control circuit 8B, the first data buffer circuit 7A,
It is temporarily stored in the second data buffer circuit 7B.

This data is received by the first series (first
SCSI disk storage device 9A, first lower SCSI control circuit 8A, first data buffer circuit 7A and second series (second SCSI disk storage device 9B, second lower SCS)
The I control circuit 8B and the second data buffer circuit 7B) are independently operated without affecting each other. Even if the reception of the data of one series is delayed, the other series precedes the reception of the data regardless of the delay. I do.

FIG. 2 shows the case where the first SCSI disk storage device 9A starts data transfer earlier. The first data buffer storage block number counting circuit 6A and the second data buffer storage block number counting circuit 6B count up each time the first data buffer circuit 7A and the second data buffer circuit 7B store one block of data. Then, the number of blocks stored in the first data buffer circuit 7A and the second data buffer circuit 7B is counted. This block count is microprocessor unit 4
Is cleared by issuing a new transfer command.

The DMA control circuit 3 has a first block for each block.
Data buffer storage block number counting circuit 6A and second
The information of the number of storage blocks held by the data buffer storage block number counting circuit 6B is read, and the data of the first data buffer circuit 7A and the second data buffer circuit 7B, which has finished storing the corresponding block data earlier, is read. Transfer to upper SCSI control circuit 2. At this time, the DMA control circuit 3 issues an instruction to the data buffer switching circuit 5 to switch the data bus to switch the upper SCSI control circuit 2 and the second data buffer circuit 7A or the upper S.
The CSI control circuit 2 and the second data buffer circuit 7B are connected.

In FIG. 2, since the first data buffer circuit 7A stores data earlier from the first block to the fifth block, the data of the first data buffer circuit 7A is sent to the upper SCSI control circuit 2. ing.

FIG. 2 shows a case where the transfer is suspended before the data of the sixth block is transferred due to the reason that the first SCSI disk storage device 9A has moved to the spare area or has been retried. In the meantime, the second SCSI disk storage device 9B, in which the start of the data transfer of the first block is delayed, stores the data of the sixth block and thereafter in the second data buffer circuit 7B earlier, and the higher order by the instruction of the DMA control circuit 3. The case of sending to the SCSI control circuit 2 is given as an example. The upper SCSI control circuit 2 is the first data buffer circuit 7A or the second data buffer circuit 7A.
The data received from B is transferred to the higher-level device 11. When either the first SCSI disk storage device 9A or the second SCSI disk storage device 9B, whichever comes first, completes the data transfer of all blocks, the DMA control circuit causes the first data buffer storage block number counting circuit 6A and the second data buffer storage The DMA control circuit 3 determines from the contents of the block number counting circuit 6B that the DMA processing circuit 3 has finished, and reports the completion of the read processing to the upper level device 11.

In FIG. 2, data read from the first SCSI disk storage device 9A is read from the first block to the fifth block, and data read from the second SCSI disk storage device 9B is read from the sixth block to the tenth block. When the read processing of the second SCSI disk storage device 9B is completed, the completion report is sent to the higher-level device 11. You may continue to temporarily store the slower series that has not completed data transfer,
You may issue a message and cancel read processing. FIG. 2 shows a case where all 10 blocks of data of the slower first SCSI disk storage device 9A are stored in the first data buffer circuit 7A.

Although not shown in FIG. 2, switching between the first series and the second series can be performed any number of times. For example, a transfer interruption first occurs in the first SCSI disk storage device 9A,
When the data of the block of the data buffer circuit 7B catches up with that of the first data buffer circuit 7A, and then is switched to the data of the second data buffer circuit 7B, and then the transfer of the data of the second SCSI disk storage device 9B is interrupted. Immediately after the data transfer of the first SCSI disk storage device 9A is recovered and the data is transferred to the first data buffer circuit 7A up to the block where the transfer interruption of the second SCSI disk storage device 9B has occurred. Switch to the data of the 1 SCSI disk storage device 9A.

Either when data transfer has not been completed to the first data buffer circuit 7A up to the block where the transfer interruption has occurred, or when data interruption has occurred simultaneously in the first SCSI disk storage device 9A and the second SCSI disk storage device 9B. After returning to the previous one, the data transfer catches up with the interrupted block, and the data in the data buffer circuit on the side that can be continuously transferred enters the upper SCSI control circuit 2 and is transferred to the upper device 11.

[0030]

As described above, in the dual disk storage device according to the present invention, the data which can be read earlier among the data transferred from the two disk storage devices is transferred to the host device, so that the data can be transferred quickly. The end of the reading process can be reported to the higher-level device at the end of the reading of the disk storage device that has finished the reading, and it is possible to prevent an increase in the processing time due to the movement to the spare area and the retry.

Further, not only when reading is continuously performed many times, but also when independent reading is performed once, it is possible to prevent an increase in processing time.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a duplicated disk storage device of the present invention.

FIG. 2 is a sequence diagram showing an embodiment of a duplicated disk storage device of the present invention.

[Explanation of symbols]

 1 Duplex Disk Storage Device 2 Upper SCSI Control Circuit 3 DMA Control Circuit 4 Microprocessor Unit 5 Data Buffer Switching Circuit 6A First Data Buffer Storage Block Number Counting Circuit 6B Second Data Buffer Storage Block Number Counting Circuit 7A First Data Buffer Circuit 7B Second data buffer circuit 8A First lower SCSI control circuit 8B Second lower SCSI control circuit 9A First SCSI disk device 9B Second SCSI disk device 10A First lower SCSI bus 10B Second lower SCSI bus 11 Upper device 12 Upper SCSI bus

Claims (4)

[Claims] 1. A first disk storage device and a second disk storage device, wherein the same data is stored in the first disk storage device and the second disk storage device in response to a data write command of a host device. In a dual disk storage device for writing data, in response to a read command from the host device, data is read independently from the first disk storage device and the second disk storage device A dual disk storage device, characterized in that data in the disk storage device is transferred to the higher-level device. 2. The end of read processing is reported to the higher-level device when the data transfer of all the blocks of the disk storage device, which has completed the read of all the blocks receiving the read command earlier, is completed. The dual disk storage device according to claim 1. 3. A first disk storage device and a second disk storage device, wherein the same data is stored in the first disk storage device and the second disk storage device in response to a data write command of a host device. A dual disk storage device for writing data, the first data storage means for storing data read independently from the first disk storage device and the second disk storage device in response to a read command from the host device; Data storage means and the data transfer in the duplicated disk storage device are controlled, the storage status of the first data storage means and the second data storage means is determined for each block, and the storage is performed quickly. The connection switching means is instructed to connect the completed data storage means and the host interface, and the data that can be stored earlier is transferred to the host device. The number of blocks which the data transfer control means and the first data storage means can store are counted, and the number of blocks which has been stored at the time when the storage is completed for each block in the first data storage means is the data transfer. The number of blocks that can be stored in the first storage block counting means and the second data storage means that is reported to the control means is counted, and the storage is completed when the storage is completed for each block in the second data storage means. Second storage block counting means for reporting the number of blocks to the data transfer control means; and the first storage block counting means for receiving the instruction from the data transfer control means.
And a connection switching means for switching the connection between the data storage means and the second data storage means and the upper interface for connecting to the upper apparatus. 4. The end of the read process is reported to the higher-level device when the data transfer of all the blocks of the disk storage device, which has completed the read of all the blocks that have received the read command, is completed earlier. The dual disk storage device according to claim 3, wherein the dual disk storage device is a storage device.