JPH09305320A - Data storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the depression of a specific disk drive due to fixed bus priority and to improve bus use efficiency when disk drives are present on one bus. SOLUTION: The bus 8 has a release state, an arbitration state, and an occupation state and a disk controller 10 and the respective magnetic disk drives 9-0-9-3 are given priority levels of bus use right acquisition respectively and obtain the right to use the bus through the release state and arbitration state to occupy the bus 8. Each magnetic disk drive functions to temporarily release the bus 8 in the occupation state of the bus 8 and then pass the right to use the bus to the disk controller 10 or another magnetic disk drive on the bus 8. The respective magnetic disk drives 9-0-9-3 confirm that the bus 8 is not occupied and join in arbitration a certain wait time after the bus 8 is released, and obtain the right to use the bus, thereby transferring data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data storage device having a disk controller and a plurality of disk devices on one bus.

[0002]

2. Description of the Related Art SCSI (Small Computer System Inte
FIG. 9 shows a configuration of a conventional data storage device using a magnetic disk device conforming to the rface) standard or the standard. SC
In SI, each of the disk control device and the magnetic disk device has an ID (identification) number (0 to 7).
The higher the number, the higher the priority in obtaining the right to use the bus. Normally, the disk controller has an ID number 7 (highest priority), and the magnetic disk device has a lower ID number, for example, 0 to 2 when connecting three magnetic disk devices.

A bus use sequence in the data storage device of FIG. 9 will be described with reference to FIG. In FIG. 9, a disk controller 51 and three magnetic disk devices 50-0
Each of .about.50-2 can use the bus (occupied state) when the bus use right is acquired through the bus free phase (released state) of the bus and the arbitration phase (arbitration state). Arbitration (bus arbitration) is performed as follows. When the bus control request is issued from the disk control device 51 or the disk devices 50-0 to 50-2 in the bus released state, the bus enters the arbitration phase, and a bus use request is issued from another device for a certain period of time T1. Wait for it to be served. When a certain period of time T2 (arbitration period: T1 <T2) elapses after entering the arbitration phase, the priorities of the devices participating in the arbitration are compared, and the device with the highest priority acquires the right to use the bus. .

The magnetic disk device can temporarily release the bus (disconnect) in the occupied state of the bus (during command execution) and pass the usage right to another.

Asynchronous I / O control has heretofore been performed in order to quickly process each access request in a plurality of magnetic disk devices connected on one bus. In this control, when the magnetic disk device disconnects during command execution, the disk controller 23 issues a bus use request to acquire the bus use right, and issues the command to the magnetic disk device that is not currently executing the command. It is a thing.
The magnetic disk device that has been disconnected and handed the usage right continues to transfer data between the disk and the internal buffer of the magnetic disk device even during disconnection, and when the data in the internal buffer reaches a certain amount, issues a bus usage request and releases the bus usage right. , And reconnect to restart data transfer on the bus.

In the data storage device shown in FIG. 9, the disk control device 51 includes disk devices 50-0 to 50-2.
When reading data from, the following operation is performed.
The disk controller 51 receives the access request from the external device 52, acquires the bus use right, and acquires the disk device 5
Issue a command to 0-0. The external device 51 receives the reply signal, generates the next access request, and passes it to the disk control device 51. The disk controller 51 further acquires the bus use right and issues a command to the disk device 50-1.
By repeating this, the requests are sequentially processed. Thus, the disk control device 51 sends a reply signal to the disk device after issuing a command without waiting for the end of the command, receives the next access request, and issues the command. This allows
Multiple magnetic disk units can alternately use the bus to simultaneously process commands, and seek time /
Since it is possible to hide the time during which no data transfer is performed, such as the rotation waiting time, efficient data transfer can be performed.

[0007]

However, when the number of disk devices connected to one bus increases, the disk device with a low priority cannot acquire the bus use right in the bus arbitration, and the disk with a high priority cannot be acquired. A so-called "subtraction" occurs, in which the execution of all the commands of the device has to be completed. In such a state, after the disk device with the higher priority executes all the commands, only the disk device with the lower priority will execute some commands in succession.

FIG. 11 shows a timing chart of such a state. In the figure, four disk devices are connected, and each disk device executes this read command by occupying the bus twice by issuing one data read command from the disk control device. As shown in the figure, since the bus is completely occupied by the command issuance from the disk control device and the data read command from the three disk devices, the disk device 2-0.
Causes subduction. For this reason, the response time to this disk device becomes extremely poor, and since the time during which no data transfer is performed, such as the seek time / rotation wait time of this disk device, is not concealed, the data transfer efficiency drops significantly.

[0009]

In order to solve the above problems, a data storage device according to claim 1 comprises a plurality of disk devices and a disk control device connected to one bus, and obtains a bus use right. Each of a plurality of disk devices connected to a bus having a release state, an arbitration state, and an occupied state is temporarily released the bus in the occupied state of the disk controller. Or, the function to pass the bus usage right to other disk devices, and the function to re-examine the bus status after a certain period of time after the bus is released when the bus is used and participate in arbitration to obtain the bus usage right. Equipped with. By setting the time interval from when the bus release status is detected when the bus is used to when it participates in bus arbitration, the fixed priority uniquely set for each disk device can be temporarily changed. Therefore, the data transfer from the disk device can be preferentially performed.

A data storage device according to a second aspect of the present invention is
The data storage device according to claim 1, further comprising a function capable of individually setting, by a disk control device, a length of time that each of the disk devices waits from detection of a bus release state to re-inspection of the bus state. By setting the time interval from the detection of the released state of the bus to the participation in bus arbitration when the bus is used by the disk controller, the fixed priority uniquely set for each disk device is temporarily set. The data transfer from any disk device can be preferentially performed.

According to a third aspect of the present invention, there is provided the data storage device according to the first aspect, wherein each of the disk devices has a length of time for waiting from the detection of the release state of the bus to the reinspection of the bus state. They are given in the reverse order of their own priority. As a result, the degree of fixed priority uniquely set for each disk device can be relaxed, or the priority can be set in a desired order.

A data storage device according to a fourth aspect is the data storage device according to the first, second or third aspect, wherein the disk controller waits for a predetermined time after the bus is released and then becomes the bus state. It is equipped with the function of reinspecting and confirming that the bus is not occupied and participating in arbitration to acquire the bus use right. As a result, the frequency of command issuance from the disk control device to the disk device having a high priority is reduced, so that the sinking of the disk device having a low priority can be avoided.

According to a fifth aspect of the present invention, there is provided the data storage device according to the fourth aspect, wherein the disk controller waits from the detection of the bus release state to the re-inspection of the bus state for a length of time of each disk. It is set to be longer than any or all of the waiting times set in the device. As a result, the relationship between the priority of command issuance from the disk control device and the priority of each disk device can be set arbitrarily.

[0014]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 shows a first embodiment of the present invention.
FIG. 3 is a block diagram of a file server including the data storage device in the embodiment of the present invention. In the figure, 1 is a data storage device, 2 is a buffer memory, 3 is a request issuing unit, 4 is a data management unit, 5 is a user management unit, and 6 is a LAN.
Device, which is a user terminal 7-0 to 7-3 via a LAN.
Are connected and deliver the data stored in the data storage device 1 in response to read requests from these user terminals.

The request for data supply from the user terminal is L
User management unit 5 and data management unit 4 via the AN device 6
Sent to The data management unit 4 calculates and manages the position of the supplied data in the data storage device 1 and the usage allocation of the data storage device 1, and sends this information to the request issuing unit 3. Based on this management information, the request issuing unit 3 sends a data read command to the data storage device 1 in response to a request from each user. The data read from the data storage device 1 is temporarily stored in the buffer memory 2 and sent to the user terminal via the LAN device 6.

The data storage device 1 includes four magnetic disk devices 9-0 to 9- connected to one SCSI bus 8.
3 and the disk controller 10. The disk control device 10 and the magnetic disk devices 9-0 to 9-3 each have an ID number, and the disk control device 10 has the highest priority ID number 7, and three magnetic disk devices 9-0.
9-3 have ID numbers 0 to 3, respectively. The disk control device 10 and the magnetic disk devices 9-0 to 9-3 acquire the bus use right through the bus free phase (released state) and the arbitration phase (arbitration state) and occupy the bus (occupied state). The magnetic disk device that is executing the command can temporarily release the bus (disconnect) and allow the disk controller 10 or another magnetic disk device to use the bus. The magnetic disk device continues to transfer data from the disk to the buffer inside the disk device even during disconnection in the reading process. When the buffer inside the magnetic disk device is filled with data, the bus usage right is acquired (reconnect), and the data transfer to the bus 8 is restarted. When two or more magnetic disk devices and disk control devices 10 try to use the bus 8 at the same time and compete with each other, the bus with the highest priority gets the right to use the bus due to arbitration. As a result of the arbitration, the magnetic disk device for which the usage right has not been obtained waits until the next bus-free operation and joins the arbitration again.

Data to be accessed by the user terminals 7-0 to 7-3 are stored in the magnetic disk devices 2-0 to 2-3, and a substantially constant amount of data is supplied by one access. I shall.

FIG. 2 is a diagram showing the configuration of the disk controller 10. The disk control device 10 includes a bus monitoring unit 11 that monitors the status of the SCSI bus 8, a bus controller 12 that controls the bus by this output, a data buffer 13 that temporarily stores data when inputting / outputting data, and a request issue. Command issuing unit 1 for decoding a command from the unit 3 and sending the command to a desired magnetic disk device
4

FIG. 3 is a diagram showing the configuration of the magnetic disk device 9. The magnetic disk device 9 includes a bus monitoring unit 15 that monitors the state of the SCSI bus 8, a bus controller 16 that controls the bus by this output, a data buffer 17 that temporarily stores data when data is input / output, and a data buffer 17. It is composed of a magnetic disk 18 for storage. The bus monitoring unit 15 has a bus control signal register indicating the state of the control signal of the bus 8. The bus controller 16 confirms that the bus 8 is in the bus free phase after a certain time has elapsed after the bus 8 was disconnected, and then enters the arbitration. Then, the bus use right is acquired and the data is sent to the disk controller 10. This waiting time Tw is set by the disk controller 10 by a command.

In this data storage device, the request issuing unit 3 manages a read request and issues a request for each magnetic disk device based on a data request from a user terminal. First, an access request assigned to the magnetic disk device 9-0 is generated and passed to the disk control device 10. After receiving the reply message from the disk control device 10, the same process is performed for the request assigned to the magnetic disk device 9-1. If there is no request, move to the next magnetic disk unit. After the process has been completed for all the terminals, the second process is performed for each magnetic disk device. In this way, the disk controller 1 cyclically generates access requests for all the magnetic disks.
Pass to 0. FIG. 4 shows a flowchart of the request issuing unit 3.

The disk controller 10 receives the access request from the request issuing unit 3 and receives the magnetic disk device 9
-Sends commands to 0 to 3 and transfers data. When the bus is occupied by the disk device when the access request from the request issuing unit 3 is received, the bus is in a waiting state. After the bus is released, a delay for the arbitration phase (2, 2 microseconds) is set, and then the state of the bus control signal is checked by the bus control signal register to determine whether it is in the bus free phase. If it is in the bus free phase, it immediately enters the arbitration, acquires the right to use the bus, issues a command to the disk device, and sends a reply message to the request issuing unit 3. Then, the next access request is received from the request issuing unit 3 and processed in the same manner. If it is not in the bus free phase, it waits until the bus is disconnected again, and repeats until a command can be issued. FIG. 5 shows a flowchart of the disk control device 10. When the disk device occupies the bus and performs data output processing, the disk controller 10 transfers data to the buffer memory.

In the system of FIG. 1, the magnetic disk device 9
The standby time of −0 is Tw0, and the magnetic disk devices 9-1 to 9-3
The waiting time is Twn, and the difference (Twn−Tw
0) is set to be longer than the bus arbitration time. Therefore, when a plurality of magnetic disk devices including the magnetic disk device 9-0 have a bus use request when the bus is released from the occupied state, the bus request is output earlier than other magnetic disk devices. Magnetic disk unit 9-
0 gets the right to use the next bus. This is shown in FIG. Unless the bus use request of the magnetic disk device 9-0 is reserved by another magnetic disk device or the disk control device 10, the normal operation as shown in the conventional example is performed.

As described above, according to the embodiment,
The bus controller of each magnetic disk unit waits a certain time after the bus is disconnected, confirms that it is in the bus free phase, and enters arbitration. The bus usage (reconnect) from this disk unit is prioritized over the bus usage by the. Therefore, by setting a shorter standby time for a disk device having a lower priority, it is possible to prevent the sinking of these magnetic disk devices.

(Second Embodiment) A file server equipped with a data storage device according to a second embodiment of the present invention will be described. The block diagram of the file server in this embodiment is the same as that in the first embodiment. Further, the disk control device 10 and the magnetic disk device 9-0
The operation of each unit other than to 3 is the same as the operation in the first embodiment.

In this system, the magnetic disk device 9-
The waiting time of 0 to 3 is Tw0 to Tw3 (Tw0 <Tw1 <
Tw2 <Tw3), and these values are set such that two adjacent waiting times among them are included during the bus arbitration period. Therefore, when the magnetic disk devices 9-0 and 9-1 issue a bus request when the bus is released from the occupied state, the magnetic disk device 9
-1, and when the magnetic disk device 9-0 issues a bus request and 9-1 does not issue, the magnetic disk device 9-0 acquires the next bus use right. This is shown in FIG. As described above, according to the embodiment of the present invention, the bus controller of each magnetic disk device confirms that the bus is in the bus free phase after a certain time has elapsed after the bus is disconnected, and enters the arbitration. By manipulating the priority of each device on a fixed bus, or increasing the probability that a lower priority device will win the bus (or lowering the probability that a higher priority device will win the bus). )be able to. For this reason,
By setting a shorter standby time for a disk device having a lower priority, it is possible to reduce the probability of sinking of these magnetic disk devices. Furthermore, by setting a longer standby time for a disk device having a higher priority, the bus acquisition probability among the magnetic disk devices can be equalized.

(Third Embodiment) A file server having a data storage device according to a third embodiment of the present invention will be described. The block diagram of the file server in this embodiment is the same as that in the first embodiment. Further, the disk control device 10 and the magnetic disk device 9-0
The operation of each unit other than to 3 is the same as the operation in the first embodiment.

Magnetic disk unit 9-0 in this system
The standby times Tw0 to Tw3 of .about.3 are the same as those of the second embodiment, but the standby time Twh of the disk controller 10 is set to be longer than Tw0 to Tw3. Therefore, when the disk controller 10 issues a bus request when the bus is released from the occupied state, in the second embodiment, the disk controller 10 according to the priority order.
Of the magnetic disk device having the bus request, in the present embodiment, the next bus use right is acquired. In this way, the reconnection request from the magnetic disk device can be prioritized with respect to the command issuance from the disk control device 10.
This is shown in FIG.

As described above, according to the embodiment of the present invention, it is confirmed that the bus controllers of the magnetic disk devices and the disk control device are in the bus free phase after a certain time has elapsed since the bus was disconnected. By entering the arbitration by controlling the bus arbitration, it is possible to control the priority of bus use between the disk controller and the magnetic disk unit when a plurality of devices have bus requests. Furthermore, by setting a long standby time in the disk controller having a high priority, the bus acquisition probability between the magnetic disk and the disk controller can be leveled.

[0030]

According to the data storage device of the present invention, when the disk controller issues a new command or when the magnetic disk device reconnects the bus for data transfer, the data storage device is kept constant after the bus is disconnected. By setting a delay of at least the time (at least for the arbitration phase), command issuance by the disk controller and reconnection from a disk device with a lower priority will take precedence over a disk device with a higher priority, so a fixed priority is given. Substantially changeable, lower priority disk drives do not sink.

Further, by setting the waiting time when the disk control device and each magnetic disk device acquire the right to use the bus, the probability that each device acquires the bus can be adjusted, and the priority is low. It is possible to increase the access frequency of the device, or decrease the access frequency of the high priority device.

[Brief description of drawings]

FIG. 1 is a block diagram of a file server using a data storage device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a disk control device 10.

FIG. 3 is a block diagram showing the configuration of a magnetic disk device 9.

FIG. 4 is an operation flowchart of the disk control device 10.

FIG. 5 is an operation flowchart of the magnetic disk device 9.

FIG. 6 is a disk control device 1 according to the first embodiment.
0 and timing chart of operation of magnetic disk device 9

FIG. 7 is a disk control device 1 according to the second embodiment.
0 and timing chart of operation of magnetic disk device 9

FIG. 8 is a disk control device 1 according to a third embodiment.
0 and timing chart of operation of magnetic disk device 9

FIG. 9 is a diagram showing a configuration of a conventional data storage device.

FIG. 10 is a diagram showing a sequence of bus usage in a conventional data storage device.

FIG. 11 “Subduction” in a conventional data storage device
Timing chart of the state where

[Explanation of symbols]

 1 Data Transfer Device 2 Buffer Memory 3 Request Issuing Unit 4 Data Management Unit 5 User Management Unit 6 LAN Device 7-0 to 7-3 User Terminal 8 Bus 9-0 to 9-3 Magnetic Disk Device 10 Disk Control Device 11, 15 Bus monitoring unit 12,16 Bus controller 13,17 Data buffer 14 Command issuing unit 18 Magnetic disk

Claims (5)

[Claims] 1. A plurality of disk devices and a disk controller connected to one bus, wherein each of the disk controller and the plurality of disk devices is given a fixed priority for obtaining the bus use right. The bus has a released state, an arbitration state, and an occupied state, and each of the disk control device and the plurality of disk devices acquires the bus use right from the released state of the bus through the arbitration state to occupy the bus, The disk device has a function of temporarily releasing the bus in the occupied state of the bus and passing the bus use right to the disk control device or another disk device. Further, when the bus is used, the bus is released. After a certain period of time after that, the bus status is re-inspected, it is confirmed that the bus is not occupied and it participates in arbitration to acquire the bus use right. That data storage device. 2. The disk controller according to claim 1, wherein the disk controller sets a length of time that each of the plurality of disk devices waits from detection of a bus release state to re-inspection of the bus state. Storage device. 3. The fixed priority given to each of the plurality of disk devices is the length of time that each of the plurality of disk devices waits from the detection of the released state of the bus to the recheck of the bus status. 2. The data storage device according to claim 1, wherein the data storage devices are provided in the reverse order. 4. When using the bus, the disk controller rechecks the bus status after a predetermined time has elapsed since the bus was released, and confirms that the bus is not occupied and participates in arbitration. The data storage device according to claim 1, 2 or 3, wherein the bus use right is acquired. 5. The length of time that the disk controller waits from detection of a bus release state to re-inspection of the bus status depends on any or all of the waiting times set for each of the plurality of disk devices. 5. The length is also long.
A data storage device as described.