JPS6243220B2 - - Google Patents

Description

Detailed Description of the Invention (a) Technical Field of the Invention The present invention relates to a system in which a plurality of higher-level devices connect and share a single lower-level device according to their system, and the present invention relates to a system in which a plurality of higher-level devices connect and share a single lower-level device. The present invention relates to countermeasures against failures that occur in a plurality of systems that connect systems, and particularly to a shared device control method that minimizes the impact on the systems or connected systems depending on the location of the failure.

(b) Prior art and problems In a system in which multiple higher-level devices share a single lower-level device by connecting them to each other, the connection between the higher-level device and the lower-level device is used to disconnect a defective device when a failure occurs. Since the control signal that leads to the failure is single, the defective device disconnection command is uniform, and it may be necessary to unnecessarily disconnect normally operating systems to ensure system safety. .
That is, there are only two states: either the lower-level device is connected to the higher-level device, or it is not connected at all.

A disk cache subsystem, for example, will be described as a system in which a plurality of higher-level devices connect and share a single lower-level device by system. FIG. 1 shows an example of the configuration of a disk cache subsystem. The central processing unit 1 includes channels 2A and 2B, and the channel 2A includes a magnetic disk control device 3A, a disk adapter 5A, and a magnetic disk device 6A, and the channel 2B includes a magnetic disk control device 3B, a disk adapter 5B, and a magnetic disk device. The devices 6B are connected in series to form one system. The disk cache 4 is composed of a disk cache common section 4C and interface sections 4A and 4B.The interface section 4A is connected to the channel 2A system, and the interface section 4B is connected to the channel 2B system. The cache subsystem is configured. As is well known, when the central processing unit 1 accesses the tracks of the magnetic disk drive 6A or 6B, the disk cache subsystem uses the magnetic disk controller 3A or 3.
B transfers the contents of the track to the memory in the disk cache common section 4, and thereafter data is exchanged at high speed between the central processing unit 1 and the memory, thereby minimizing the influence of waiting time inherent in magnetic disk devices. It is used to

In the conventional disk cache subsystem, there is a single control signal between the disk cache 4 and the magnetic disk controllers 3A and 3B as described above, so for example, the interface section 4A of the disk cache 4 A fault has occurred in the disk cache common section 4C and interface section 4B, so the data bus 8B of the channel 2B system which can operate normally and the data bus 8A of the own system
Since the data bus 7A is also disconnected, the system of the channel 2A which can be operated via the data bus 8A is also stopped, although there is a slight decrease in performance. That is, in addition to stopping the operation of other systems that are completely unrelated to the failure, even if there is a decrease in performance, the system that can still operate is also stopped. The same applies to the case where a failure occurs in the interface section 4B, and even when the magnetic disk control device 3A has a failure, the channel 2 is completely unrelated to the failure.
Stop system B. Furthermore, the same problem occurs when the magnetic disk control device 3B fails.

(c) Purpose of the Invention The purpose of the present invention is to eliminate the above-mentioned drawbacks by providing a plurality of control signals, thereby preventing unnecessary disconnection of a system that can normally operate depending on the content of a fault that occurs within the system, and preventing the adverse effects of a faulty device. The object of the present invention is to provide a shared device control method that makes it possible to keep the amount of noise to a minimum.

(d) Structure of the Invention The structure of the present invention is such that in a system in which a plurality of higher-level devices connect and share a single lower-level device by system, a plurality of control signal lines are transmitted from the higher-level device to the lower-level device. Depending on the location of the failure in the lower-level device or the higher-level device, systems that can operate normally are left to operate as they are, and systems that are inoperable are screened and disconnected from the faulty device in the system, and the system of the faulty device is removed. And the plurality of control signal lines are used in combination so as to minimize the adverse effects on the system.

(e) Embodiment of the invention FIG. 2 is a block diagram illustrating an embodiment of the invention. Control signal lines 9A,
10A, 9B, and 10B are connected, and the following four conditions are expressed by a combination of "1"/"0" for the control signal lines 9A, 10A or 9B, 10B, respectively. For example, if the channel 2A system in Figure 1 is the own system, and the channel 2B system is the other system, 1. When the control signal line 9A is "0" and 10A is "0", both systems are connected, that is, the data buses 7A and 7B are connection.

2 When the control signal line 9A is “1” and the control signal line 10A is “0”, own system is closed and other systems are connected, that is, data bus 7
A is closed, 7B is connected.

3 When the control signal line 9A is “0” and the control signal line 10A is “1”, other systems connected to the own system are closed, that is, data bus 7
A connection 7B is blocked.

4 When the control signal line 9A is "1" and the control signal line 10A is also "1", both systems are blocked, that is, both data buses 7A and 7B are blocked. The normal operation mode is the condition 1), but whether the disk cache 4 is normal or abnormal is determined by the status signal lines 11A and 11B being connected to the magnetic disk controllers 3A and 3B, respectively, and the magnetic disk controller 3A being connected to the status signal lines 11A and 11B. The state of the disk cache 4 is detected and processed by referring to the signal line.

A: If a failure occurs in the interface unit 4A, the magnetic disk device 3A is subject to the conditions of 2) above.

B. When a failure occurs in the interface section 4B, the magnetic disk drive 3A is subject to the conditions 3) above.

C. If a failure occurs in the disk cache common section 4C, the magnetic disk device 3A
The conditions are as follows.

D When the magnetic disk control device 3A detects the occurrence of its own failure, the magnetic disk control device 3A
is the condition of 2) above.

The above description has been made regarding the operation of the magnetic disk control device 3A, but the same applies to the magnetic disk control device 3B, and generally, the one that detects the occurrence of a failure earlier in time is given priority in control.

(f) Effects of the Invention As explained above, the present invention uses two control signal lines compared to the conventional case where there is a single control signal line, and the combination of the two control signal lines makes it possible to issue a command to disconnect a defective device. It is possible to select and isolate defective devices from common ones, and avoid unnecessary blockages of normal systems and unnecessary blockages of faulty systems, making it possible to minimize their negative effects. Therefore, the processing capacity of the system is less degraded, and the hardware required for it only requires the addition of two cables, so the effect is great.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an example of the configuration of a disk cache subsystem, and FIG. 2 is a block diagram illustrating one embodiment of the present invention. 1 is a central processing unit, 2 is a channel, 3 is a magnetic disk controller, 4 is a disk cache, 5
is a disk adapter, and 6 is a magnetic disk device.

Claims (1)

[Claims] 1. In a system in which a plurality of higher-level devices connect and share a single lower-level device by system, a plurality of control signal lines are provided from the higher-level device to the lower-level device,
Depending on the location of the failure in the lower-level device or the higher-level device, the plurality of control signal lines are configured such that a system that can operate normally is allowed to continue operating, and a system that is inoperable is selected and disconnected from the faulty device in the system. A shared device control method characterized by using a combination of the following.