JPH08272542A - Computer system - Google Patents

Abstract

PURPOSE: To reduce a processing restoring failur accompanying the occurrence of the failur in a disk controller. CONSTITUTION: This system is provided with a magnetic disk controller 14 controlling the magnetic disk devices MS00 to MS03 of a normal system, a magnetic disk controller 16 controlling the magnetic disk devices SS00 to SS03 of an alternate system, a host channel device 12 having a connection route between a host device 11 and the disk controllers 14 and 16, host channel buses 13a and 13b which connects the host channel device 12 and the magnetic disk controllers 14 and 16 corresponding to each of them and connects them to be alternate to each other, and a disk control bus connecting the magnetic disk devices MS00 to MS03 and SS00 to SS03 of both systems to the magnetic disk controllers 14 and 16. Thus, the system has a failur alternate function for controlling the disk devices of both systems by another normal system when the failur occurs in one system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system that uses a disk system having a duplicated magnetic disk device as an external storage.

[0002]

2. Description of the Related Art Some disk systems used as an external storage function in a computer system, for example, a magnetic disk system, have dual magnetic disk devices in order to improve reliability and fault tolerance. In such a magnetic disk system, the two magnetic disk control devices are magnetic disk devices corresponding to the respective magnetic disk devices (which are logically treated as one device and are physically composed of a plurality of magnetic disk devices. ) Is controlled separately.

Conventionally, in a magnetic disk system in which a magnetic disk device is duplicated, when a failure such as a power supply failure occurs in one magnetic disk controller, the magnetic disk device managed by the magnetic disk controller is changed to a magnetic disk device. Access becomes impossible, the entire system is degenerated, and only the other normal magnetic disk controller operates.

Further, at the time of failure recovery, if there is a change in data due to access to the normal system during the failure or during the failure recovery, the normal system magnetic disk device and the restored magnetic disk device are separated. Because of the need for consistency
It is essential to copy the volume (or partial if manageable) from the normal system for all (or if manageable, accessed) magnetic disk devices of the failure system.

[0005]

In the conventional dualized magnetic disk system, when a failure occurs in one of the systems, it is essential to copy the data from the normal system after the failure is recovered. It was. Therefore, when the number of magnetic disks connected under each magnetic disk control device increases, the number of magnetic disk devices that are degraded due to the occurrence of a failure in the disk control device also increases, and the processing at the time of failure recovery increases. There was a problem that it would end up.

The present invention has been made in consideration of the above circumstances, and an object of the present invention is to provide a disk controller capable of reducing the processing at the time of failure recovery due to the occurrence of a failure in the disk controller. To do.

[0007]

According to the present invention, in a computer system having a host device in which a disk device is duplicated, a first bus and a first system connected to the first bus are configured. A plurality of disk devices, a second bus, a plurality of disk devices constituting a second system connected to the second bus, and a plurality of disk devices connected to the first and second buses. A first disk device controller for controlling a plurality of disk devices constituting the first system via a bus, and the first and second buses connected to the first disk device controller via the second bus. When a failure occurs in the second disk device that controls the plurality of disk devices that make up the second system and the first system, the host device for the plurality of disk devices that makes up the first system Access request to the second disk control Characterized by comprising a substitution processing control means for processing is given to the device. Further, it is characterized by comprising control means for setting whether or not to execute the operation of the alternative processing control means when a failure occurs in one of the systems.

[0008]

According to such a configuration, when a system to which the disk device is connected is duplicated so that it can be accessed from any of the disk control devices, a failure occurs in one of the disk control devices. In addition, access control can be performed on a disk device under the disk control device in which a failure has occurred. Therefore, since the contents of access until the failure is recovered are reflected in the disk device, it is not necessary to perform processing such as data copying after the failure is recovered.

[0009]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a computer system according to this embodiment. As shown in FIG. 1, the computer system has a magnetic disk system 10 as an external storage function, and a host device 11 is connected via a host channel device 12. The host channel device 12 has two host channel buses 13
a, 13b via two magnetic disk control devices 14,
It is connected to 16.

The host channel device 12 controls data transfer and the like between the host device 10 and the magnetic disk control devices 14 and 16. Host channel bus 13a, 1
3b is a host device 11 and a magnetic disk controller 14,
16 are connected to each other, and the paths are controlled by host ports (described later) in the magnetic disk control devices 14 and 16.

In this embodiment, the magnetic disk controller 1
4, 16 are mounted on a backplane 50 (described later) and are housed in the same housing. Assuming that the magnetic disk control devices 14 and 16 are mounted on the backplane, when one of the magnetic disk control devices 14 and 16 fails, they are alternately connected via the backplane.

The magnetic disk control devices 14 and 16 are assigned to the master side and the shadow side to be duplicated. In this embodiment, the magnetic disk control device 14 is used as a master.
The magnetic disk controller 16 is used as a shadow. Under normal conditions, the magnetic disk control device 14 on the master side controls the magnetic disk devices MS00 to MS03, and the magnetic disk control device 16 on the shadow side controls the magnetic disk device SS00.
~ Control SS03. When a failure occurs in one of the magnetic disk control devices, the other magnetic disk control device causes the magnetic disk devices MS00 to MS0 of both systems.
3, the magnetic disk control device 14 for controlling SS00 to SS03 is a host channel device 12
In addition to the basic function of controlling the magnetic disk device in response to an access request from the host device 11 via the, a fault notification circuit 20, a fault detection circuit (mask circuit) 22, an alternative control circuit 24, a host port 26, Also, a power supply unit 28 is provided.

The fault notifying circuit 20 notifies the magnetic disc control device 16 of the other system and the host device 10 of the fault occurrence when a fault (power supply abnormality) occurs in the magnetic disc control device of its own system. Circuit 20a, output circuit 20b
have. The monitoring circuit 20a monitors a signal (alarm signal) output from the power supply unit 28 (power supply control unit 28a) when an abnormality occurs, and outputs a signal indicating the occurrence of the abnormality to the output circuit 20b when the abnormality occurs. The output circuit 20b outputs the output signal from the monitoring circuit 20a for a certain period of time to a disk control bus connected to the magnetic disk controller 16 (described later).
To (reset signal). Also, the output circuit 20b
Notifies the status indicating that a failure has occurred to the host device 11 from the host port 26 via the failure detection circuit 22 and the alternative control circuit 24, and the host device 11
Do not accept access from.

The fault detection circuit 22 is a mask circuit which can be set by firmware or software.
This is for controlling whether or not the failure replacement function by the other magnetic disk device is executed when a failure occurs. The fault detection circuit 22 is configured to execute the fault substitution function, and the fault notification circuit 20 (output circuit 20b).
Output to the alternative control circuit 24. Further, the failure detection circuit 22 detects a notification (reset signal) from the magnetic disk control device 16 indicating that a failure has occurred, and notifies the replacement control circuit 24 if it is set to execute the failure replacement function. . The fault detection circuit 22 can be set to execute the fault substitution function only for an arbitrary period.

The alternative control circuit 24 is a failure notification circuit 20.
When the output from the (output circuit 20b) (indicating the occurrence of a failure in the other magnetic disk controller) is received via the failure detection circuit 22, the alternative control of the failure alternative function is performed.
The alternative control circuit 24 performs alternative control of the port number of the host port 26 and the data path in the magnetic disk controller.

The host port 26 is an access port from the host channel device 12 to the magnetic disk controller 14, and is controlled by a port number (ID). In this embodiment, two access ports (port numbers 01, 0
9) and the port number 09 is connected to the host channel bus 13b of the shadow side magnetic disk control device 16. Under normal conditions, the magnetic disk controller 14 is accessed using the port with port number 01. On the other hand, when a failure occurs in the magnetic disk control device 16, the magnetic disk control device 14 is switched to another port by switching the port with the port number 01 to the port number 09.

The power supply unit 28 supplies electric power to each unit in the magnetic disk control device and is controlled by the power supply control unit 28a. The power supply control unit 28a sends an alarm signal to the monitoring circuit 2 of the failure notification circuit 20 when an abnormality occurs in the power supply unit 28.
0a, and the power supply output is cut off.

The magnetic disk controller 16 has a basic function of controlling the magnetic disk device in response to an access request from the host device 10, a failure notification circuit 30, a failure detection circuit (mask circuit) 32, and an alternative control. A circuit 34, a host port 36, and a power supply unit 38 are provided. The function of each unit is the same as that of the magnetic disk control device 14, and the description is omitted.

The magnetic disk controller 14 is connected with two systems of disk control buses 40a and 40b, and one disk control bus 41a is connected to the magnetic disk device MS00.
~ MS03 is connected. Magnetic disk controller 1
6, two systems of disk control buses 42a and 42b are connected, and one disk control bus 42b is connected to the magnetic disk devices SS00 to SS03.

The magnetic disk devices MS00 to MS0
3, SS00 to SS03 each have a power supply device that is not affected even if a failure such as power failure occurs in the corresponding magnetic disk control device.

By mounting the magnetic disk controllers 14 and 16 on the backplane 50, the disk control bus 40a of the magnetic disk controller 14 is connected to the disk control bus 42a of the magnetic disk controller 16 via the backplane 50. Similarly, the disk control bus 40b
And the disk control bus 42b are connected. The two-system disk control buses are for ensuring an alternative path for controlling the magnetic disk device from the normal magnetic disk control device when an abnormality occurs in the magnetic disk control device. In addition, the magnetic disk control device 14, 16
By mounting on the backplane 50, the host channel buses 13a and 13b are mutually connected to the host port (port number 09) on the other side via the backplane 50, and substitution is performed when an abnormality occurs in the magnetic disk controller. Secure a route.

The magnetic disk control devices 14 and 16 are mounted on a back plane 50, for example, as shown in FIG. In FIG. 2, modules realizing the magnetic disk control devices 14 and 16 are mounted on the front and back of the backplane 50 and housed in the same housing.

The backplane 50 provides a common path for mounted modules. The modules of the respective magnetic disk control devices 14 and 16 are connected using the predetermined pins common to the backplane 50 so that the host channel buses 13a and 13b and the disk control buses 40a, 40b, 42a and 42b are connected to each other as shown in FIG. They are connected in the relationship shown inside.

Next, the operation of this embodiment will be described.
First, the normal operation will be described. Host device 11
Accesses the magnetic disk control devices 14 and 16 via the host channel device 12. The magnetic disk controller 14 is accessed via the port number 01 of the host port 26 using the host channel bus 13a. The magnetic disk control device 16 is accessed via the port number 01 of the host port 36 using the host channel bus 13b. The magnetic disk control devices 14 and 16 perform duplication by controlling the magnetic disk devices MS00 to MS03 and SS00 to SS03 under their control.

Next, the operation when a failure occurs in one magnetic disk controller will be described. Here, it is assumed that a failure has occurred in the power supply unit 38 of the shadow side magnetic disk control device 16. In addition, the fault detection circuits 22 and 3
It is assumed that 2 is set (no mask) to execute the failure replacement function.

When a voltage drop such as a power failure occurs in one magnetic disk controller 16, the power supply controller 38a of the power supply unit 30 outputs an alarm signal. The monitoring circuit 30a of the failure notification circuit 30 detects the occurrence of a power supply failure and notifies the output circuit 30b of the occurrence.

The output circuit 30b has a disk control bus 42.
The failure is notified to the normal magnetic disk controller 14 via b (may be the disk control bus 42a) (reset signal). This notification is continued for a certain time or longer. Further, the output circuit 30b notifies the alternative control circuit 34 via the failure detection circuit 32. The alternative control circuit 34 controls the host port 36, returns a wait status indicating that some failure has occurred in response to an access from the host device 11, and does not accept the access. On the host side, it will be in a wait (or retry) state.

On the other hand, a normal magnetic disk controller 14
Detects a signal (reset signal or the like), which is continuously notified for a certain period of time or more, via the disk control bus 40b, and receives a notification of occurrence of a failure.

Since the fault detection circuit 22 is set to execute the fault substitution function, it notifies the substitution control circuit 24 of the received fault occurrence notification. The alternative control circuit 24 changes the port number 09 of the host port 26 to the port number 0.
The value is changed to 1 and the alternative control of the data path in the magnetic disk controller 14 is performed.

As a result, the access from the host channel device 12 to the shadow side magnetic disk control device 16 via the host channel bus 13b is made to the master side host port 26 (port changed to port number 01).
Is input to Then, the command for the shadow side that has been waited is input to execute the alternative control. That is, the magnetic disk controller 14 controls the magnetic disk device M.
In addition to the control of S00 to MS03, the magnetic disk devices SS00 to SS00 are connected via the magnetic disk control buses 40b and 42b.
The control for SS03 is also executed.

In this way, when a failure occurs in the magnetic disk controller of one system, an alternative route to the other system is realized, and a normal magnetic disk controller continues until the failure is recovered. Both-system magnetic disk unit MS00
~ MS03, SS00 to SS03 can be controlled, so that processing such as data copy is not necessary after failure recovery, and processing at the time of failure recovery accompanying failure occurrence is reduced.

In the above embodiment, the magnetic disk control devices 14 and 16 of both systems are mounted on the backplane 50 to connect the bus, but the connection form may be other means such as a cable. It is also possible to use. Further, the magnetic disk system has been described as an example, but the system may be another storage medium, for example, an optical disk or a magneto-optical disk.

[0033]

As described in detail above, according to the present invention, 2
If a failure occurs in one of the redundant disk control units, the other system that is operating normally will take over the alternative control until the system is restored. It is possible to reduce it.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a computer system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a state in which a magnetic disk control device is mounted on a backplane in the present embodiment.

[Explanation of symbols]

10 ... Magnetic disk system, 11 ... Host device, 12
... host channel device, 13a, 13b ... host channel bus, 14, 16 ... magnetic disk controller, 20,
30 ... Fault notification circuit, 20a, 30a ... Monitoring circuit, 20
b, 30b ... Output circuit, 22, 32 ... Fault detection circuit (mask circuit), 24, 34 ... Alternative control circuit, 26, 36 ...
Host port, 28, 38 ... Power supply unit, 28a, 38a ...
Power control unit, 40a, 40b, 42a, 42b ... Disk control bus, 50 ... Backplane, MS00-MS0
3, SS00 to SS03 ... Magnetic disk device.

Claims (2)

[Claims] 1. A computer system having a host device in which disk devices are duplicated, a first bus, and a plurality of disk devices constituting a first system connected to the first bus; A second bus; a plurality of disk devices constituting a second system connected to the second bus; and a first bus connected to the first and second buses and the first bus via the first bus. A first disk device control device for controlling a plurality of disk devices configuring the system, and a plurality of devices configuring the second system via the second bus and the first and second buses. Second disk device for controlling the second disk device and the first system, when a failure occurs in the first system, an access request from the host device to the plurality of disk devices constituting the first system is issued to the second disk device. Given to the controller and processed Computer system characterized by comprising a substitution processing control means that. 2. When a failure occurs in one of the systems,
The computer system according to claim 1, further comprising control means for setting whether or not the operation of the alternative processing control means is executed.