JP2562834B2 - Redundant control method - Google Patents

Description

The present invention relates to a duplication control method for controlling and processing only a primary volume when performing read access to a duplication volume, and further simplifies the processing of the disk control unit and accurately determines the volume detachment. The central processing unit notifies the disk controller of the redundant logical volume address, primary and secondary volume addresses, and information indicating the operating status of the redundant control, and then switches to the redundant volume. In the duplication control method in which only the primary volume is read-accessed during read access, the disk control unit notifies the central processing unit of the abnormality occurrence when the primary volume abnormality occurs, and the central processing unit causes the abnormality. By issuing an occurrence notification, a setting command is issued to the disk control unit and After the switching process between the volume and the secondary volume is performed, a read command is reissued to the disk control unit to perform the read process of the secondary volume switched to the primary volume.

[Industrial applications]

The present invention relates to a duplication control method, and more particularly to a duplication control method that controls and processes only a primary volume during read access to a duplication volume.

[Conventional technology]

A conventional redundant control method is a central processing unit (CPU), a disk control unit, and two disks (# 1 disk and # 2 disk).
In a system composed of disks, etc., at the time of read access, only the read processing of the primary disk is performed based on the read command.

This conventional method will be described with reference to FIG.
Issues a read command (by logical address specification) to the disk controller as indicated by a ₁ , the disk controller analyzes this command (b _{1 in} FIG. 4) and refers to the duplication setting information to read the normal system. The disk (in this case, the # 1 disk) is accessed (b _{2 in} FIG. 4), and the # 1 disk read processing (the same, c ₁ ) is performed.

The disk control unit normally performs the read process, and when the read process is completed, a status notification of normal completion (in FIG. 4,
b ₃ ) to the CPU.

Next, the CPU issues the read command again (in Fig. 4, a
_2), which are disk control unit analyzes (same, b _4), to access the main system disk # 1 (same, b _5). As a result, the read process of the # 1 disc (same as c ₂ ) is started, but if an error occurs in the read process at this time, necessary recovery (for example, retry by command reissue) is performed.

However, when the disk control unit determines that recovery is impossible due to a hardware error or the like, the primary disk is disconnected and the secondary disk (# 2 disk in this case) is accessed as shown in FIG. As a result, the # 2 disk read process is performed (d _{1 in} FIG. 4), and when the process is normally performed and the process ends, the disk control section returns a normal end status.
Notifies the CPU (the same, b _7). If disconnection occurs, the fact that disconnection has been performed is also notified.

[Problems to be Solved by the Invention]

According to the above-mentioned conventional duplex control method, since the disk controller performs the recovery process and the judgment of disconnection, the judgment of disconnection may not be normally performed.

For example, when the same disk is used by multiple host computers, access to the disk may be busy forever, and in such a case, the disk control unit may determine that the hardware is abnormal. There wasn't.

The present invention has been made in view of the above points, and it is an object of the present invention to provide a duplexing control method that simplifies the processing of the disk control unit and, moreover, accurately determines the volume disconnection and switches the volume.

[Means for solving the problem]

According to the duplex control method of the present invention, the disk control unit notifies the central processing unit of the occurrence of an abnormality in the primary volume, whereby the central processing unit issues a setting command to the disk control unit to send the primary volume and the secondary system. The switching process with the volume is performed.

Thereafter, the central processing unit reissues the read command to the disk control unit to cause the secondary volume switched to the primary volume to perform the read processing.

[Action]

Upon receiving the abnormality notification from the disk control unit, the central processing unit determines the details of the abnormality and performs necessary recovery processing.
If an error occurs even in this recovery process, or if recovery is not possible due to a hardware error, a setting command is issued to switch the primary and secondary disks.

Therefore, the central processing unit only needs to change the settings of the primary / secondary disks, and the volumes can be switched by the same processing before the occurrence of the abnormality.

Further, the disk control unit need only have the read function of the primary volume, and does not need the read function of the secondary volume.

Further, when there are a plurality of central processing units, and these access the same primary volume, each central processing unit has a communication means and the system configuration is known, so that access to the same primary volume is not possible. Even if it stays busy forever, it is possible to accurately determine whether it is due to a hardware abnormality.

〔Example〕

FIG. 1 is an explanatory view of a control procedure of an embodiment of the method of the present invention, and FIG. 2 is a system configuration diagram of an embodiment of the present invention. Second
In the figure, 1 is a central processing unit (CPU), 2 is a disk controller, and 3 and 4 are disks # 1 and # 2, respectively. The disk controller 2 is provided corresponding to the CPU 1. A memory (MEM) 5 stores the operation program of the CPU 1 and the like. Further, the # 0 disk in the disk control unit 2 is a disk that the disk control unit 2 logically looks like, and the disks 3 and 4 of # 1 and # 2 are duplicated so that the CPU 1 has the # 0 disk. Indicates that it is shown as a disc (the # 0 disc does not actually exist).

In the system shown in FIG. 2, when the CPU 1 (or software) performs redundant volume control, first,
The disk controller 2 is notified of the information shown in FIG. 3 by the duplex setting command. As shown in FIG. 3, this notification information is composed of a redundant logical volume address, a primary volume address, a secondary volume address, and information indicating the operating status.

The duplicated logical volume address is the disk controller 2
Indicates the address of the duplicated virtual volume shown to the CPU 1, which is # 0 in the embodiment of FIG.

The primary volume address indicates the address of the primary volume in the duplicated volume, and is # 1 in the embodiment of FIG. The secondary volume address indicates the address of the secondary volume in the duplicated volume, and is # 2 in the embodiment of FIG.

Furthermore, the operating status indicates the operating status of the duplex control, and when disconnecting the primary volume, the internal disconnection flag is turned on to instruct the disconnection.

After the above-mentioned notification of the duplication setting command to the disk control unit 2, the CPU 1 designates the duplication logical volume address and the duplication control is performed to issue the read / write command. However, in the read control, when instructed by the dual logical volume address designation, only the primary volume is accessed and the read processing is performed.

Next, in such a system, the operation in the case where an abnormality occurs during the read processing during duplex control will be described with reference to FIG. First, the CPU 1 issues a read command by designating a duplicated logical volume address to the disk control unit 2 (7a in FIG. 1).

As a result, the disk control unit 2 analyzes this read command as shown by 8a, refers to the duplicated configuration table (created by the setting command), and reads the primary volume (that is, # 1 disk 3). Perform 9.
However, at the time of this read processing 9, the disk control unit 2 determines that it is abnormal if read data is not obtained from the # 1 disk 3, and notifies the CPU 1 of abnormal termination.

The CPU 1 determines the content of the abnormality based on this abnormal end notification and performs necessary recovery processing. If recovery is abnormal or cannot be recovered due to hardware error, CP
U1 issues a setting command to the disk control unit 2 (first
Figure 7b). As a result, the disk control unit 2 disconnects the # 1 disk 3, and as shown by 8b, the # 1 disk 3 that was the primary volume in the duplicated configuration table until then becomes the secondary volume and the secondary volume until then. Then, the # 2 disk 4 which has been described above is switched to the primary volume (that is, the duplex configuration is reset).

The CPU 1 waits for the end notification of the duplicated configuration resetting from the disk control unit 2 and reissues the read command by the logical address designation as shown by 7c in FIG. Then, the disk control unit 2 analyzes the read command as shown by 8c in FIG. 1 (refer to the redundant configuration table) and determines that it is the primary volume at that time. Access the # 2 disk 4 which was
Perform When this read processing 10 is completed, the disk control unit 2 notifies the CPU 1 of that fact.

When an abnormality occurs, not only is the volume of the primary system and the volume of the secondary system switched, but the disconnection flag in the operating status information is turned on to disconnect the primary volume that has the error.

〔The invention's effect〕

As described above, according to the present invention, the disk control unit does not need the function of determining the abnormality content, and the read function of the secondary volume is also unnecessary. Since the central processing unit (software) whose system configuration is known determines the details of the abnormality,
It is possible to accurately determine whether the cause of the abnormality is due to a hardware abnormality, a recovery processing abnormality, or another cause, and it is possible to perform more reliable duplex control.

[Brief description of drawings]

 FIG. 1 is an explanatory diagram of a control procedure of an embodiment of the method of the present invention, FIG. 2 is a system configuration diagram of an embodiment of the present invention, FIG. FIG. 7 is an explanatory diagram of a control procedure of an example of a conventional method. In the figure, 1 is a central processing unit (CPU), 2 is a disk controller, and 3 and 4 are disks.

Claims (1)

(57) [Claims] 1. A duplicated volume after a central processing unit (1) notifies a disk control unit (2) of a duplicated logical volume address, primary and secondary volume addresses, and information indicating an operation state of duplicated control. In the duplication control method of performing read processing only on the primary volume (3) during read access to the disk, the disk control unit (2) includes the primary volume (3).
When the abnormality occurs, the central processing unit (1) is notified of the abnormality occurrence, and the central processing unit (1) issues a setting command to the disk control unit (2) in response to the abnormality occurrence notification to issue the primary volume (3). ) And the secondary volume (4) are switched, and then a read command is reissued to the disk control unit (2) to read the secondary volume (4) switched to the primary volume. A duplex control method characterized by performing processing.