JPH03109633A - Load sharing system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a load sharing system, and more particularly to a load sharing system using cross-call mirrored disks.

(Prior Art) A load sharing system is a computer system intended to combine a plurality of processors and distribute functions and loads to the individual processors.

Mirrored disks are sometimes used in such load sharing systems. Mirrored disks write the same content on two disks so that even if one disk fails, the system will not go down.
This is a highly reliable system. In addition, cross calls are used for failure countermeasures, performance improvements, data sharing, and other purposes.

FIG. 3 is a block diagram showing the schematic structure of a load sharing system using the Roscall mirror F disk as described above.

In the conventional device configured in this way, CPU3, CPU
Both U4 and DISKI are used as data disks, and D
ISK2 was used as a standby side.

(Problem to be solved by the invention) However, in a system with such a configuration, the DI S
When the KI is busy, when the other CPU tries to use the DISKI, it has to wait, resulting in wasted time.

The present invention has been made in view of these problems, and an object of the present invention is to provide a load sharing system that minimizes wasted time when accessing mirrored disks.

(Means for Solving the Problems) The present invention for solving the above-mentioned problems provides a first disk and a second disk constituting a mirrored disk, a setting section in which one of the disks to be used is set, and a first disk. a port to which the second disk is connected, a port to which the second disk is connected, and a port selection section that selects one of the ports when reading based on the state of the setting section
A setting section where the PU and one disk to be used are set.

A second CPU includes a port to which the first disk is connected, a port to which the second disk is connected, and a port selection unit that selects one of the ports when reading based on the state of the setting unit. , is characterized in that the disk to be used is set differently in the setting section of each CPU.

(Function) In the load sharing system of the present invention, the first CPU
Different disks are set by the respective setting sections for the and second CPU, and different disks are used when each CPU performs reading simultaneously.

(Example) Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In the figure, parts corresponding to those in FIG. 3 are designated by the same reference numerals.

In the figure, 1 is a first disk, and 2 is a second disk. These two disks are part of a mirrored disk system. 10 is one CPU, and 20 is the other CPU. In addition, since the internal configurations of CPU10 and CPU20 are the same, in order to simplify the explanation,
The inside of CPU10 will be shown and explained in detail.

11 is disk 1. C when both disks are normal
This is a setting section where a disk that can be dedicated to PUl0 is set. 12 is a failure detection unit that detects a failure of the disk 1;
13 is a failure detection unit that detects a failure of the disk 2; 14 is a setting unit 11. This is a port selection unit that selects a port for exchanging data with either the disk 1 or the disk 2 based on instructions from the failure detection unit 12.13.

This port selection unit 14 receives a control signal D 1 wrtte for writing data to the disk 1 from the setting unit 6
Logic gate 14a1 that receives the detection result from the failure detection section 13; Logic gate 1 that receives the output of this logic gate 14a and the inverted signal of the detection result from the failure detection section 12.
Control signal D for writing 4b1 data to disk 2
2write, from the setting unit 11] Logic gate 14c receiving the detection result from the failure detection unit 12
1 consists of a logic gate 14d that receives the output of this logic gate 14c and an inverted signal of the detection result from the failure detection section 13. Reference numeral 15 indicates a driver/receiver section that constitutes a port for data transfer with the disk 1, 16 a driver/receiver section that constitutes a port for data transfer with the disk 2, and 17 all remaining parts of the CPU 10. In this embodiment, it is called a control section.

Moreover, FIG. 2 is a flowchart showing the general operation of CPU10.

When a write request occurs, data is written to both disk 1 and disk 2 (FIG. 2 (S2)). The failure detection unit 12.13 also detects the disk 1. A failure of the disk 2 is detected, and the detection result is supplied to a logic gate in the port detection section 14. That is, the control section 17
from write signal D 1 write along with data.
e and D2write occur and the logic gates 14a, 14
c. On the other hand, logical groups) 14b and 14d are
It receives the outputs of these logic gates 14a and 14c, as well as inverts the failure detection signal, and if there is no failure, outputs PIEN and P2EN to put the driver/receiver sections 1.5 and 1.6 into a passing state. . As a result, data from the control unit 17 is written to both disk 1 and disk 2.

Next, a case where a read request occurs will be explained.

If both disks 1 and 2 are normal, reading is performed from the disk on the side set by the setting unit 11 (the side to which the sketch in the setting unit 11 is not connected) ((S7), (S8))
. In the example shown in FIG. 1, since the switch of the setting section 11 is set to the disk 1 side, the input of the logic gate 14c becomes H level, and P2EN for putting the driver/receiver section 16 into the passing state is activated. Occur. Therefore, reading from the disk 2 is performed. Although not shown,
It is assumed that the CPU 20 is set by the setting unit 21 to read data from the disk 1. For this reason, each CPU reads from a separate disk, so
No unnecessary waiting time occurs.

Now, consider a case where one of the disks is faulty. For example, when the disk 2 fails, the failure detection section 13 outputs an H-level failure detection signal. This H-level failure detection signal is supplied to the logic gate 1.4a, and a PIEN is generated that causes the normal driver/receiver unit 15 on the disk 1 side to pass. Further, since the failure detection signal is supplied to the inverting input terminal of the logic gate 14d, this logic gate 14d is closed and P2EN does not occur regardless of the setting of the setting section 11. The same applies when the disk 1 fails. Note that when both disks are in failure, both logic gates 14b and 14d are closed by the failure detection signal, so PIEN and P2EN are not generated and no reading is performed.

As explained above in detail in this embodiment, when both disks are normal, each CPU sets the disk to be accessed for reading separately, so each CPU can access it independently, resulting in unnecessary waiting. Time does not occur.

(Effects of the Invention) As explained in detail above, according to the present invention, each CPU accesses a different disk when the disk is normal, so that the wasted time during mirrored disk access is reduced. This makes it possible to realize a load sharing system that minimizes the

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a flowchart for explaining the operation, and FIG. 3 is a block diagram showing the schematic structure of a conventional device. 1.2...Disk 10.20...CPU1
1.21... Setting section 12.13, 22.23... Failure detection section 14.24.
...Port selection sections 14a to 14d...Logic gates 15.16, 25.26...Driver/receiver section 1
7.27...Control unit

Claims (1)

[Claims] A first disk (1) and a second disk (2) constituting a mirrored disk, and a setting section (11) in which one of the disks to be used is set.
, a port (15) to which the first disk (1) is connected,
A first CPU ( 10) and a setting section (21) where one of the disks to be used is set.
, a port (25) to which the first disk (1) is connected,
A second CPU having a port (26) to which the second disk (2) is connected, and a port selection section (24) that selects one of the ports when reading depending on the state of the setting section (21); A load sure system comprising: a setting section (11) and a setting section (21) each having different settings for the disks to be used.