JPH04366464A - Controlling circuit for synchronization of rotating in array disk apparatus - Google Patents

Abstract

PURPOSE:To reduce influence at the time of master drive failure by supplying a timing signal from master drive of each group, and switching a timing signal to be supplied to the drive under control according to a designation from a managing mechanism. CONSTITUTION:A plurality of magnetic disk drives 1 for constituting an array disk apparatus, is, for example, divided into two groups (groups A/B), and master drives 1a are respectively set to the groups A/B. The drives 1a supply timing signals 1 by intersecting. Slave drives 1b of the divided groups A/B are always synchronized with the signal 1 to be supplied from the normal master driver 1a. Accordingly, the drive 1a of a defect side can be removed merely by switching the signal from the drive 1a of the defect side to the signal 1 of the drive of a normal side.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to rotation synchronization in an array disk device that connects a plurality of magnetic disk drives in parallel and is equipped with a mechanism for synchronizing the rotations of those magnetic disk drives, and performs data transfer with a host device. Regarding control circuits.

One of the technologies required in the computer industry in recent years is high-speed, large-capacity data transfer. An array disk device is a subsystem that meets the above requirements by connecting multiple disk drives in parallel. Therefore, array disk devices are the subsystems that are attracting the most attention regarding data processing in future computer systems.

[0003] In such an array disk device composed of a plurality of magnetic disk drives, a failure occurs in the rotation synchronization mechanism that enables high-speed data transfer between the host device and the plurality of magnetic disk drives connected in parallel. There is a need for a rotation synchronization control method that can reduce the impact on the computer system and allow normal subsystems to continue operating without stopping them when this occurs.

0004

2. Description of the Related Art FIG. 3 is a diagram illustrating a conventional array disk device. In the conventional array disk device, in order to realize the rotation synchronization function of the magnetic disk drives 1 constituting the array, a plurality of magnetic disk drives 1 are positioned as master and slave, and one main (master) drive 1a and multiple other slave drives 1b.

[0005] As a result, a plurality of slave drives 1b simultaneously receive the timing signal ■ supplied from the master drive 1a and establish rotation synchronization by following it, and furthermore, the timing signal ■ is always sent at regular intervals. The synchronization of rotation was maintained by supplying

Specifically, for example, a microprocessor (MPU) that performs main control of the master drive 1a
10 receives the index mark signal ■ from the rotating body mechanism 11 and distributes it to the rotation control mechanism 12 of the subordinate slave drive 1b via the gate circuit (G), thereby controlling the rotating body mechanism 1 of each slave drive 1b.
1 rotation control was performed.

[0007] When a failure occurs in any one of the plurality of slave drives 1b, a microprocessor (MPU) 10 receives and decodes the failure information from each failure detection mechanism 17, and detects the failure in the magnetic disk drive. By notifying the array disk control unit 3, which is a higher-level management mechanism, through the control unit 2, the array disk control unit 3 controls which magnetic disk drive
Since it is possible to recognize whether 1 is a failure, the corresponding magnetic disk drive 1's microprocessor (MPU)
By sending a specific command to the magnetic disk drive 10, the corresponding magnetic disk drive 1 can be disconnected.

Therefore, after disconnecting the faulty magnetic disk drive 1, the other magnetic disk drives 1
It is possible to maintain rotational synchronization of the entire array disk device without affecting the system.

[0009]

[Problem to be Solved by the Invention] However, if a failure occurs in the master drive 1a, the timing signal ■ for rotation synchronization cannot be sent out.
This will also affect the rotation synchronization function of the array disk device.

[0010] Therefore, when a failure occurs in the master drive 1a (that is, when a failure notification is received from the master drive 1a), the magnetic disk drive control unit 2 that controls the plurality of magnetic disk drives 1 is collectively managed. The array disk control unit 3 monitors the index mark signal ■ from each slave drive 1b, recognizes the normality of rotational synchronization, and sends a specific command as described above to determine whether the failure has occurred. In addition to disconnecting the master drive 1a, a candidate for the next master drive 1a was selected from among the plurality of slave drives 1b, and a new master drive was set.

[0011] In this method, a new master drive 1
After setting a, connect multiple other slave drives again 1b
This is an operation that establishes and maintains synchronization with
There is a problem in that when a failure occurs in the master drive 1a, it takes a very long time and affects the computer system.

In view of the above conventional drawbacks, the present invention provides an array disk device that connects a plurality of magnetic disk drives in parallel, has a function of synchronizing the rotations of these magnetic disk drives, and transfers data with a host device. For example, if a failure occurs in the rotational synchronization mechanism of a magnetic disk drive in a master drive, it will be detected and the magnetic disk drive will be disconnected, and the other magnetic disk drives that make up the array disk will be disconnected. It is an object of the present invention to provide a rotation synchronization control circuit that can maintain a rotation synchronization function while reducing the influence on the rotation.

[0013]

[Means for Solving the Problems] FIG. 1 is a diagram showing the basic configuration of the present invention. The above problem is solved by a rotation control circuit configured as follows.

The present invention includes a plurality of magnetic disk drives 1, a magnetic disk drive control section 2 that controls each of the magnetic disk drives 1, and an array disk control section 3 that manages the magnetic disk drive control sections 2. In an array disk device having a mechanism for synchronizing the rotation of the drives 1, the plurality of magnetic disk drives 1 constituting the array disk device are divided into a plurality of groups (A, B, ...),
A master drive 1a and a slave drive 1b that rotates under the master drive 1a are set for each group (A, B, ~), and each group (A, B,
The timing signal generation mechanism 15 for detecting rotation abnormality and the timing signal ■ supplied from each master drive 1a of each group (A, B, ~) to each master drive 1a of ~) are sent to the array disk control unit. 3, based on the selection instruction ■ issued based on the abnormality notification signal ■ from the timing signal generation mechanism 15 of each master drive 1a, select one of them and send the selected one to the subordinate slave drive 1b. The plurality of slave drives 1b of each group (A, B, .
Synchronization is established and maintained by the timing signal ■ supplied from the switching mechanism 16 of the master drive 1a of each group (A, B, ~), and the master drives 1a of each group (A, B, ~) mutually communicate with the timing signal ■
The switching mechanism 16 switches the timing signal ■ to be supplied to the subordinate slave drive 1b in response to an instruction signal ■ from the array disk control unit 3 indicating which timing signal ■ is to be enabled.
Configure it to do so.

[0015]

[Operation] FIG. 1 is a diagram showing the basic structure of the present invention. In the present invention, a plurality of magnetic disk drives 1 constituting an array disk device are divided into, for example, two groups (groups A/B), and a master drive 1a is set in each group A/B. .

[0016] The plurality of slave drives 1b in each group A/B always establish and maintain synchronization using the timing signal (2) supplied from the master drive 1a of its own group.

The master drives 1a set in each group A/B alternately supply the timing signal ■ to each other, and determine which timing signal is sent from the array disk control unit 3 equipped with a management mechanism. An instruction to enable the signal {This is generated based on the fault information notified from each master drive 1a of each group A/B} The signal ■ causes switching of the timing signal ■ supplied to the slave drive 1b. controlled to do.

Therefore, each group A/
Each slave drive 1b of B is always synchronized with the timing signal ■ supplied from one of the normal master drives 1a, and furthermore, since the master drives 1a are parallelized, the slave drives 1b of The timing signal from master drive 1a on the normal side from the timing signal from master drive 1a on the normal side.
Simply switch to the failed master drive 1a.
can be removed from the battle line, making it possible to re-establish synchronization in a shorter time than with conventional methods.

Therefore, when a failure occurs in the master drive 1a of a certain group A or B, the effect on the entire array disk device can be reduced compared to the conventional array disk device.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below in detail with reference to the drawings. The above-mentioned FIG. 1 is a diagram showing the principle configuration of the present invention, and FIG.
FIG. 1 is a diagram showing an embodiment of the present invention.

The present invention includes a plurality of magnetic disk drives 1, a magnetic disk drive control section 2 that controls each of the magnetic disk drives 1, and an array disk control section 3 that manages these magnetic disk drive control sections 2. In an array disk device having a mechanism (1), ~ for synchronizing the rotation of a plurality of disk drives 1, a plurality of magnetic disk drives 1 constituting the array disk device are arranged in a plurality of groups (A, B, ~).
) into each group (A, B, ~),
In addition to setting a master drive 1a and a slave drive 1b that rotates under the master drive 1a,
Each master drive of each group (A, B, ~) 1
a, the timing signal generation mechanism 15 for detecting abnormal rotation, and the timing signal ■ supplied from each master drive 1a of each group (A, B, ~),
Based on the selection instruction ■ issued from the array disk control unit 3 based on the abnormality notification signal ■ from the timing signal generation mechanism 15 of each master drive 1b, one is selected and the subordinate slave drives A switching mechanism 16 for supplying the selected timing signal 16 to 1b is a necessary means for carrying out the present invention. Note that the same reference numerals indicate the same objects throughout the figures.

The configuration and operation of the rotation synchronization control circuit in the array disk device of the present invention will be explained below with reference to FIG. 2 while referring to FIG. The two groups A and B shown in FIGS. 1 and 2 are for timing signal generation,
When detecting a failure within the master drive 1a or detecting an abnormality in each other's timing signals, it is reported to the array disk control unit 3 equipped with a management mechanism using a failure notification signal ■. The timing signal generation mechanism 15 , the switching mechanism 16 that switches the timing signal between groups A and B and supplies the output to the slave drives 1b of each group A and B, and the array disk controller 3 , signal line 1 from master drive 1a of each group A, B
It has a signal line 105 that controls the switching mechanism 16 based on the failure notification signal (2) notified by 03, 104.

[0023] Now, an array disk device is constituted by ten magnetic disk drives 1, and each group A and B
is assumed to have one master drive 1a and four slave drives 1b.

Slave drives 1 for each group A and B
b is a timing signal supplied from the master drive 1a of its own group (signal line 101, or
102) as a reference, and the master drives 1a are connected to each other's signal line 10.
1 and 102, the rotational synchronization of the ten magnetic disk drives 1 is maintained.

[0025] Here, one of the slave drives 1
If a failure occurs in disk drive b, the nine disk drives can maintain rotational synchronization without affecting the entire array disk, as in the conventional method.

Furthermore, in the rotation synchronization control circuit of the present invention, when a failure occurs in either one of the master drives 1a, the signal from the array disk control unit 3 that has received the failure notification via the signal line 103 or 104 is The timing signal is supplied to the slave drive 1b under the failed master drive 1a according to the instruction from the control signal {instruction signal for which timing signal ■ of groups A/B to be enabled} on the signal line 105. ■{by signal line 101 or 102),
The switching mechanism 16 switches to the timing signal ■ from the other group, group A or B.
to each signal line 106, or signal line 10
7.

Therefore, the slave drive 1b of each group A/B is always the normal master drive 1a.
Synchronization can be established based on the timing signal ■. As described above, the present invention provides a rotation synchronization control circuit for an array disk device that connects a plurality of magnetic disk drives in parallel, has a mechanism for synchronizing the rotations of these magnetic disk drives, and performs data transfer with a host device. In this method, a plurality of magnetic disk drives constituting the array disk device are divided into a plurality of groups, a master drive is set in each group, and the plurality of slave drives in each group are always assigned to their own group. Synchronization is established and maintained by the timing signal (■) supplied from the master drive, and the master drives of each group supply the above-mentioned timing signal (■) to each other from the management mechanism section (array disk control section) of the array disk device. The feature is that the timing signal (2) to be supplied to the subordinate slave drives is switched in accordance with the instruction signal (2) indicating which group's timing signal (2) is to be enabled.

[0028]

As described above in detail, the rotation synchronization control circuit in the array disk device of the present invention divides the plurality of magnetic disk drives 1 constituting the array disk device into, for example, two groups ( Group A
/B), set the master drive 1a in each group A/B, respectively, and set the master drive 1a in each group A/B,
The plurality of slave drives 1b always establish and maintain synchronization using the timing signal ■ supplied from the master drive 1a of their own group, and each master drive 1 set in each group A/B.
a is from an array disk control unit 3 which supplies timing signals (■) to each other in a crossed manner and is equipped with a management mechanism;
Instructions for which timing signal to enable {this is
The timing signal ■ generated based on the fault information notified from each master drive 1a of each group A/B is configured to switch the timing signal ■ supplied to the slave drive 1b, so it is divided into two. Each slave drive 1b of each divided group A/B is always one of the normal master drives.
It is synchronized with the timing signal ■ supplied from the master drive 1a, and since the master drive 1b is parallelized, the timing signal from the faulty master drive 1a is simply synchronized with the timing signal ■ from the normal master drive 1a. By simply switching to the signal ■, the master drive 1a on the faulty side can be removed from the battle line, and synchronization can be re-established in a shorter time than in the conventional method.

Therefore, when a failure occurs in the master drive 1a of a certain group A or B, the effect on the entire array disk device can be reduced compared to the conventional array disk device.

[Brief explanation of the drawing]

[Figure 1] Principle configuration diagram of the present invention

[Fig. 2] A diagram showing an embodiment of the present invention

[Figure 3] Diagram explaining a conventional array disk device

[Explanation of symbols]

1 Magnetic disk drive 1
a Master drive 1b Slave drive 10 Microprocessor (MPU) 11
Rotating body mechanism 1
2 Rotation control section 15 Timing signal generation mechanism 16
Switching mechanism 17 Failure detection mechanism 2 Magnetic disk drive control unit 3
Array disk control unit 101 to 107 Signal line ■ Timing signal ■ Switching instruction signal or selection signal ■
Failure notification signal

Claims (1)

[Claims] Claim 1: A plurality of magnetic disk drives (1);
It has a magnetic disk drive control unit (2) that controls each magnetic disk drive (1), and an array disk control unit (3) that manages these magnetic disk drive control units (2). )
In an array disk device having a mechanism (■, ~) for synchronizing the rotation of the disks, a plurality of magnetic disk drives (1) constituting the array disk device are divided into a plurality of groups (A, B, ~), each with a Each group (A, B, ~) has a master drive (1
a) and a slave drive that rotates under it (
1b), and each master drive (1a) of each group (A, B, ~) is provided with a timing signal generation mechanism (15) for detecting abnormal rotation;
The timing signal (■) supplied from each master drive (1a) of ~) is sent to the array disk control unit (3) as an abnormality notification from the timing signal generation mechanism (15) of each master drive (1a). Signal (■
) Based on the selection instruction (■) issued based on
b) is provided with a switching mechanism (16) that supplies the selected timing signal (■), and each group (A, B
, ~) are always synchronized by the timing signal (■) supplied from the switching mechanism (16) of the master drive (1a) of its own group (A, or B, ~). Establish and maintain
Master drive (1a) of each group (A, B, ~)
) supply the above-mentioned timing signal (■) to each other, and an instruction signal from the array disk control unit (3) indicating which timing signal (■) is to be enabled.
A rotation synchronization control circuit in an array disk device, characterized in that the timing signal (■) supplied to the subordinate slave drive (1b) is switched by the switching mechanism (16) according to (■).