JPH103445A - Non-stop type magnetic disk device system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a non-stop type magnetic disk device system which can improve reliability without the expense of performance through the use of a magnetic disk device which does not have a special function on reliability. SOLUTION: Information on plural access paths to one magnetic disk device are registered in the path management part 1 of a host computer 5 and the remaining access paths registered in the path management part 1 are used at the time of a channel fault. When the channel to the magnetic disk device is used at present by the other magnetic disk device, a path cache management part 3 pools an I/O request in a path cache area 2, and retransmits it when the channel becomes a usable state.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a non-stop magnetic disk drive system, and more particularly to a non-stop magnetic disk drive system that does not require a special function for a magnetic disk drive.

[0002]

2. Description of the Related Art Conventionally, this type of non-stop type magnetic disk drive system has been used in a plurality of magnetic disk drives connected to a computer or the like in order to improve their reliability.

An example of a conventional nonstop magnetic disk drive system is described in Japanese Patent Application Laid-Open No. Hei 4-315238. According to the access path closing method described in this publication, a plurality of disk control adapters for controlling a plurality of magnetic disk devices are connected under a subsystem controller, and all of the disk control adapters are connected to all the magnetic disk devices. In the event of a disk control adapter failure, the path from the subsystem controller to the failed disk control adapter is blocked so that another disk control adapter can access the target magnetic disk device.

[0004]

However, in the above-mentioned conventional non-stop type magnetic disk drive system, it is not considered that the reliability is improved only by the existing magnetic disk drive. It is necessary to have a function, and it is necessary to replace the magnetic disk device used in the existing computer.

Accordingly, the present invention has been made in view of the above-mentioned circumstances, and it is possible to improve reliability without sacrificing performance by using a magnetic disk device having no special function relating to reliability. It is an object of the present invention to provide a nonstop magnetic disk drive system.

[0006]

In order to achieve the above-mentioned object, the present invention provides a plurality of magnetic disk devices having only one channel connection portion and a point-to-point (Po
int-to-point) exchange, a plurality of channels for connecting the plurality of magnetic disk devices and the point-to-point exchange, and a plurality of channels for connecting the point-to-point exchange and a host computer. If a failure occurs in the channel when the host computer issues an I / O request to a certain magnetic disk device, the channel is used for the magnetic disk device using the channel for another magnetic disk device. I / O
Provided is a nonstop magnetic disk drive system characterized by continuing a request.

In the non-stop magnetic disk drive system according to the present invention, the host computer pools the I / O requests until a channel shared by the plurality of magnetic disk drives becomes available. It is characterized in that performance degradation during channel sharing is prevented.

[0008]

Next, embodiments of the present invention will be described with reference to the drawings.

Referring to FIG. 1, a non-stop magnetic disk drive system according to the present invention, in a first embodiment, comprises a plurality of magnetic disk drives 11 and 12, a host computer 5, and a point-to-point system. In the magnetic disk drive system connected by a plurality of paths via the exchange 8 at -point), when a channel failure occurs in a certain magnetic disk drive, a channel used by another magnetic disk drive is used.

More specifically, path switching means (path management unit 1, channel connection unit 4, channels 6, 7 and switch 8) for switching a path to a magnetic disk device when a channel to a magnetic disk device fails. Path cache means (path management unit 1, path cache area 2, and path cache management unit 3) for pooling I / O requests when a channel to the magnetic disk device is being used by another magnetic disk device. .

Referring to FIG. 3, the non-stop magnetic disk drive system according to the present invention is used in another magnetic disk drive in the second embodiment when a channel failure occurs in the magnetic disk drive. Use no channel.

More specifically, path switching means (path management unit 1, channel connection unit 4, channels 6, 7, 13 and switch 8) for switching the path to the magnetic disk device when a channel failure occurs in the magnetic disk device. And a path cache unit (path management unit 1, path cache area 2, and path cache management unit 3) that pools I / O requests when a channel to the magnetic disk device is being used by another magnetic disk device. Prepare.

In the path management section 1 of the host computer 5, information on a plurality of access paths is registered for one magnetic disk device, and the remaining information registered in the path management section 1 when a channel failure occurs. Use an access path.

When the channel to the magnetic disk device is being used by another magnetic disk device, the path cache management unit 3 of the host computer 5 pools the I / O request in the path cache area 2 and the channel is Resend when available.

[0015]

Next, in order to more specifically describe the above-described embodiment of the present invention, an embodiment of the present invention will be described in detail with reference to the drawings.

[0016]

FIG. 1 is a block diagram showing the configuration of a first embodiment of a nonstop magnetic disk drive system according to the present invention.

Referring to FIG. 1, a first embodiment of the present invention comprises magnetic disk devices 11, 12, an exchange 8, a channel 9 connecting the exchange 8 and the magnetic disk device 11, an exchange 8 and a magnetic disk. Channel 10 connecting device 12
, A host computer 5, and channels 6 and 7 connecting the channel connection unit 4 of the host computer 5 and the exchange 8.
And

The exchange 8 is a point-to-point (Poin)
It is a t-to-point switch, which allows access to the target connecting device by specifying the inlet and outlet channels.

The host computer 5 includes a channel connection unit 4, a channel from the channel connection unit 4 to the exchange 8, and a channel from the exchange 8 to the magnetic disk units 11, 12, that is, from the channel connection unit 4 to the magnetic disk unit 11, 12, a path cache area 2 for preventing performance degradation during path degeneration operation of the magnetic disk drive system, and a path cache management for managing the path cache area 2. Part 3
And

The path management unit 1 and the path cache management unit 3 are realized by a function on an operating system (OS) operating on the host computer 5, and the path cache area 2 is provided in the main storage device included in the host computer 5. Implemented using a part.

The path management unit 1 manages the state of the path, that is, whether the path is in use or not in use, in addition to the function of managing the path. In the path management unit 1, a normally used path is called a "primary path", and a path used when a failure occurs in the primary path is called a "sub path".

When a program running on the OS of the host computer 5 instructs access to the magnetic disk device, an inquiry about the path to the target magnetic disk device is made to the path management unit 1. If the target path is not in use, data is issued to the channel connection unit 4, and if the target path is in use, data is issued to the path cache management unit 3.
The path cache management unit 3 pools the received data in the path cache area 2.

Next, the operation of the first embodiment of the present invention will be described with reference to FIG.

The channel 6 is used as the primary path of the magnetic disk drive 11, the channel 7 is used as the primary path of the magnetic disk drive 12, and the channel 7 is used as the secondary path of the magnetic disk drive 11. Device 12
The channel 6 is used as a sub-path of the above, and such information is managed by the path management unit 1.

In the magnetic disk drive system of this embodiment, when no failure occurs, the channel 6 is used to issue / receive data to / from the magnetic disk drive 11 and to the magnetic disk drive 12 Channel 7 is used for issuing / receiving data.

When a failure occurs in the channel 6 during transmission and reception of data between the host computer 5 and the magnetic disk device 11, the path management unit 1 detects the failure and
The I / O request to the magnetic disk device 11 which triggered the failure is suspended, and the I / O request to the magnetic disk device 11 from the OS on the host computer 5 is suppressed.

Next, a sub system path for the magnetic disk device 11 is checked. As a result, it is understood that the channel 7 is used for the sub system path of the magnetic disk device 11. Next, the channel 7 is registered in the primary path of the magnetic disk device 11,
The information of the channel 7 registered in the secondary path of the magnetic disk device 11 is deleted. As a result, the information of the path management unit 1 is as shown in FIG. That is, as shown in FIG.
Is that the main path is from channel 6 to channel 7
And the sub system path is not set.

Thereafter, the suppression of the I / O request from the OS on the host computer 5 to the magnetic disk device 11 is stopped, and the suspended I / O request to the magnetic disk device 11 is resumed. At this time, the pending I /
The O request is issued to the magnetic disk device 11 using the channel 7 according to the information of the path management unit 1 shown in FIG. That is, even if the channel 6 becomes unusable due to a failure, access to the magnetic disk device 11 is not interrupted.

The magnetic disk drive 11 and the magnetic disk drive 12 based on the information of the path management section 1 shown in FIG.
The magnetic disk drive system in a state where it is accessed
It is a kind of degenerate state. That is, the magnetic disk device 1
The transmission and reception of data to and from the magnetic disk drive 12 share the channel 7 because both the primary path of the magnetic disk drive 11 and the primary path of the magnetic disk drive 12 are channels 7.

In such a state, if an I / O request to the magnetic disk device 11 and the magnetic disk device 12 is made simultaneously from the OS operating on the host computer 5, any of the I / O requests Until the end of
Although the target magnetic disk device is in a usable state, the other I / O request cannot be made.

In the present embodiment, the path cache management unit 3 uses the path cache area 2 in order to prevent such performance degradation.

That is, in the state of the path management unit 1 shown in FIG. 2, an I / O request is issued from the user program running on the OS of the host computer 5 to the magnetic disk device 11 and the magnetic disk device 12. In this case, the path management unit 1 checks the state of the channel 7 and confirms that the channel 7 is not being used, and then issues an I / O request to the magnetic disk device 11.

Next, before processing an I / O request to the magnetic disk device 12, the state of the channel 7 is checked. At this time, the state of the channel 7
Since it is being used for the / O request, the I / O request to the magnetic disk device 12 is passed to the path cache management unit 3. The path cache management unit 3 pools the received I / O request to the magnetic disk device 12 in the path cache area 2. After that, the path management unit 1 is notified that the I / O requests to the magnetic disk device 11 have been pooled.

After that, the I
It is assumed that the / O request has been completed. After recognizing this, the path management unit 1 notifies the path cache management unit 3 that the I / O request using the channel 7 has been completed. The path cache management unit 3 that has received the notification checks whether the I / O requests for the channel 7 are pooled in the path cache area 2. At this time, since I / O requests for the magnetic disk device 12 are pooled in the path cache area 2, the request is taken out of the path cache area 2 and issued to the magnetic disk device 12.

With the above processing, it is possible to prevent performance degradation when a plurality of magnetic disk devices share a channel.

[0036]

Embodiment 2 Next, a second embodiment of the present invention will be described. FIG. 3 is a block diagram showing the configuration of a second embodiment of the nonstop magnetic disk drive system of the present invention.

Referring to FIG. 3, the second embodiment of the present invention has basically the same configuration as that of the first embodiment described above, except that the channel 1 connecting the channel connection unit 4 and the exchange 8
3 that the path management unit 1 has the magnetic disk device 11,
12 in that it has two sub-system paths.

Next, the operation of the second embodiment of the present invention will be described with reference to FIG.

The channel 6 is used as a primary path of the magnetic disk drive 11, and the channel 7 is used as a primary path of the magnetic disk drive 12.
The channel 13 is used as the sub-path 1 of the magnetic disk 11, the channel 7 is used as the sub-path 2 of the magnetic disk 11, and the channel 6 is used as the sub-path 2 of the magnetic disk drive 12. It is managed by the management unit 1.

In the magnetic disk drive system of this embodiment, when no failure occurs, the channel 6 is used for issuing / receiving data to / from the magnetic disk drive 11, and the data is transmitted to / from the magnetic disk drive 12. Channel 7 is used for issuing / receiving data.

When a failure occurs in the channel 6 during data transmission / reception between the host computer 5 and the magnetic disk device 11, the path management unit 1 detects the failure and
The I / O request to the magnetic disk device 11 which triggered the failure is suspended, and the I / O request to the magnetic disk device 11 from the OS on the host computer 5 is suppressed.

Next, a sub path for the magnetic disk device 11 is checked. As a result, it is understood that the channel 13 is used for the sub system path of the magnetic disk device 11. next,
The channel 13 is registered in the primary path of the magnetic disk device 11 and the information of the channel 13 registered in the secondary path 1 of the magnetic disk device 11 is deleted. Next, the information of the channel 7 registered in the sub-path 2 of the magnetic disk device 11 is deleted, and the information of the channel 7 is registered in the sub-path 1 of the magnetic disk device 11. As a result, the information of the path management unit 1 is as shown in FIG. That is, as shown in FIG.
Is that the main path is from channel 6 to channel 1.
3, the sub system path 1 is changed from the channel 13 to the channel 7, and the sub system path 2 is not set.

Thereafter, the suppression of the I / O request from the OS on the host computer 5 to the magnetic disk device 11 is stopped, and the suspended I / O request to the magnetic disk device 11 is resumed. At this time, the pending I /
The O request is issued to the magnetic disk device 11 using the channel 13 according to the information of the path management unit 1 shown in FIG. That is, even if the channel 6 becomes unusable due to a failure, access to the magnetic disk device 11 is not interrupted. Further, since the channel 13 is used only for I / O requests to the magnetic disk device 11, the performance of the magnetic disk device system is not reduced.

[0044]

As described above, according to the present invention,
By connecting the magnetic disk device and the host computer via a point-to-point exchange instead of directly connecting the magnetic disk device and the host computer, it is possible to have an extra channel, so that it was not possible with the conventional magnetic disk device system. Reliability can be achieved.

According to the present invention, during the path degeneration operation, the path cache management unit temporarily pools the I / O requests in the path cache area, and the channel shared by a plurality of magnetic disk devices can be used. Since the I / O request is issued when the state is changed, it is possible to prevent the performance of the magnetic disk drive system from deteriorating even during the path degeneration operation.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of a nonstop magnetic disk drive system according to the present invention.

FIG. 2 is a diagram showing another state (state at the time of a channel failure) of the path management unit 1 shown in FIG.

FIG. 3 is a block diagram showing the configuration of a second embodiment of the non-stop magnetic disk drive system of the present invention.

4 is a diagram showing another state (state at the time of a channel failure) of the path management unit 1 shown in FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Path management part 2 Path cache area 3 Path cache management part 4 Channel connection part 5 Host computer 6, 7 channels 8 Exchange 9, 10 channels 11, 12 Magnetic disk unit 13 channels

Claims (5)

[Claims] A plurality of magnetic disk devices having only one channel connection portion and a point-to-point (Point-to-point)
a plurality of channels for connecting the plurality of magnetic disk devices and the point-to-point switch, and a plurality of channels for connecting the point-to-point switch and a host computer. When a failure occurs in a channel when the host computer issues an I / O request to a certain magnetic disk device, the I / O to the magnetic disk device is performed using the channel for another magnetic disk device. A non-stop magnetic disk drive system characterized by continuing an / O request. 2. The method according to claim 1, wherein the host computer pools the I / O requests until a channel shared by the plurality of magnetic disk devices becomes available, thereby preventing performance degradation during channel sharing. The non-stop magnetic disk drive system according to claim 1. 3. In a magnetic disk drive system in which a plurality of magnetic disk drives and a host computer are connected by a plurality of paths via a point-to-point switch, when a channel failure to a certain magnetic disk drive occurs, the magnetic disk drive is connected to the host A non-stop type magnetic disk drive system comprising a path switching means for switching a path. 4. The non-stop system according to claim 3, further comprising a path cache means for pooling I / O requests when a channel to a certain magnetic disk device is being used by another magnetic disk device. Type magnetic disk drive system. 5. The apparatus according to claim 1, wherein said path switching means switches a path to the magnetic disk device so as to use a channel not used by another magnetic disk device when a channel failure occurs in the magnetic disk device. Item 4. The non-stop magnetic disk drive system according to item 3.