JPH05233466A - Fault recovery system of doubled auxiliary storage device - Google Patents

Abstract

PURPOSE:To make a recovery from double fault trouble by preventing input/ output performance from deteriorating owing to the trouble recovery of the doubled auxiliary storage device. CONSTITUTION:If fault occurs when information 15 is read out of the auxiliary storage device 17, this information 15 is read out of the auxiliary storage device 18 and stored in a main storage device 16, and this information is held by an information control part 19 until the fault is removed. A fault acceptance part 11 registers fault information in a fault control table 13. A fault recovery part 12 inspects an input/output table 14 and performs a fault recovering process when there is no input/output request to the auxiliary storage device 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure recovery system for a redundant auxiliary storage device in an information processing system.

[0002]

2. Description of the Related Art In an information processing system using a duplicated auxiliary storage device, the same information is stored in each of a pair of duplicated auxiliary storage devices.

Conventionally, the recovery of a failure that has occurred in one of such dual auxiliary storage devices has been performed by the procedure shown in FIG. First, the recovery processing of the medium of the auxiliary storage device in which the failure has occurred, that is, the reformation of the failed portion and the allocation of the alternate area are performed (step 32). Then, the information to be stored at the failure location is read from the other auxiliary storage device and written into the failure storage device to recover the information at the failure location (step 33).

As a publicly known example of this type of failure recovery system, there is, for example, the one described in Japanese Patent Publication No. 1-244610.

[0005]

In the above conventional method, the information stored in the one auxiliary storage device lost due to the occurrence of the failure is stored only in the auxiliary storage device which was normally operating until the completion of the recovery. If the existing information is lost due to a double failure, it becomes impossible to continue processing and recover the information. Therefore,
In order to avoid this, the failure recovery process must be executed immediately when the failure occurs.

However, in the method of executing the failure recovery processing immediately when a failure occurs, since the input / output operation to the auxiliary storage device for the failure recovery processing, which does not normally occur when the failure occurs, occurs. There is a problem that the input / output performance temporarily deteriorates when the input / output frequency to the volume is high. In particular, as the number of information processing systems that do not allow temporary deterioration of I / O performance such as entrusting human life or commerce, the temporary deterioration of I / O performance due to failure recovery as described above cannot be overlooked. It

Therefore, an object of the present invention is to provide a failure recovery system for a redundant auxiliary storage device which can solve the above problems.

[0008]

In order to achieve the above object, according to the present invention, in an information processing system using a duplicated auxiliary storage device, a failure occurs in one of the duplicated auxiliary storage devices. In this case, the information to be stored at the location of occurrence of the failure is stored in another storage device until the failure recovery processing is completed, and the failure recovery processing is performed by an input / output request to the failure occurrence auxiliary storage device. It is to be executed when there is no.

[0009]

As described above, since the information to be stored at the fault location is kept in the other storage device until the end of the fault recovery processing, the information duplication state is maintained. Even if the information stored in the normal auxiliary storage device is lost due to the double failure, the failure recovery is not disabled. As a result, it becomes unnecessary to immediately execute the failure recovery processing when the failure occurs. Since the processing of the input / output request to the auxiliary storage device in which the failure has occurred has priority over the failure recovery processing, the deterioration of the input / output performance due to the failure recovery processing can be substantially avoided.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the configuration diagram of FIG. 1 and the flowchart of FIG.

In the information processing system according to this embodiment, as shown in FIG. 1, a dual auxiliary storage device 1 is used.
7 and 18 are used, the central processing unit 10 is provided with a failure reception unit 11, a failure recovery unit 12, a failure management table 13, and an information management unit 19 in connection with the failure recovery. The input / output table 14 and the main storage device 16 are not added for the purpose of failure recovery, but are shown because they are used for failure recovery.

The outline of the failure recovery operation is as follows. For example, when a failure occurs in one of the auxiliary storage devices 17, the failure reception unit 11 registers the failure information in the failure management table 13. The failure recovery unit 12 monitors the input / output table 14 of the device corresponding to the auxiliary storage device 17 requiring the failure recovery, and maintains the input / output performance if the failure recovery process is executed due to the input / output request. When it is difficult to do so, the execution of the failure recovery process is postponed, the failure recovery process is executed when there is no input / output request, and when this is completed, the failure information is deleted from the failure management table 13. Also,
The information management unit 19 holds the information 15 (information to be stored at the failure location) that has been input / output until the failure recovery is completed in the storage device 16, so that the information 15 that is not in the duplicated state can be replaced by a new failure. Try not to lose.

The structure and operation will be described in detail below. The fault management table 13 includes a fault device name 13a and a fault position 13
b and the main memory 1 of the information 15 to be stored in the faulty part
Position 13c on 6 can be stored. The information 15 is held in the page 16a, which is a unit of management of the main storage device 16, and corresponds to the information 17a, 18a held in the auxiliary storage devices 17, 18. A page in the main storage device 16 and a record, which is a unit for holding information in the auxiliary storage devices 17 and 18, have a one-to-one correspondence in FIG. May be different. The input / output table 14 is the device name 14a.
And a queue 14b for input / output requests to each device.

In the above system, it is assumed that a failure occurs when reading the information 17a from the auxiliary storage device 17. In this case, the same information 18a as the information 17a is read from the other normal auxiliary storage device 18 and the main storage device 16
It is stored as information 15 in the upper page 16a (step 22).

When the failure reception unit 11 detects the occurrence of the failure, it also manages failure information, that is, the auxiliary storage device name 13a in which the failure has occurred, the failure position 13b, and the position 13c of the information 15 on the main storage device 16. The data is registered in the table 13 and the input / output processing is terminated (step 23). However, if the same failure information is already registered in the failure management table 13, registration is not performed to avoid duplication.

On the other hand, the fault recovery unit 12 regularly inspects the fault management table 13. Then, when the fault is registered, the corresponding device name 14a in the input / output table 14 is based on the faulty auxiliary storage device name 13a.
Check queue 14b for I / O requests for (step 24). That is, in this example, it is checked whether or not there is an input / output request to the auxiliary storage device 17. Queue 1
If there is an I / O request to 4b, the failure recovery is postponed.

If the queue 14b is empty, the failure recovery unit 12 starts the following failure recovery processing. First, the recovery processing of the medium of the auxiliary storage device 17 in which the failure has occurred, that is, reformatting of the failed portion and allocation of the replacement area are performed (step 25). Next, the information 15 relating to the failure stored in the main storage device 15 is stored in the auxiliary storage device 17
And the information is recovered (step 26). Finally, the auxiliary storage device name 13a whose failure recovery has been completed and its failure position 13b are deleted from the failure management table 13 (step 27).

However, if an output request for the information 15 is registered in the queue 14b after starting such a failure recovery process, the failure recovery process is stopped and the output request is prioritized. In this case, the failure recovery process is restarted when the queue 14b becomes empty next time. When the information 15 exists, the input request from the auxiliary storage device 17 does not occur.

In the information processing system, the page of the main storage device 16 is used by a plurality of processes. When there are a large number of processing execution requests and it is not possible to secure enough pages to execute the processing, some of the used pages are output to the auxiliary storage device in order to be used by other processing. Is done. Therefore, the page 16a having the information 15 is allocated to another process before the failure is recovered, and the information is stored in the auxiliary storage device 1.
If it is output to 8, the double holding state of the information at the failure point is lost, and the recovery from the double failure cannot be guaranteed.

In order to avoid this, the information management section 19
By referring to the information position 13c registered in the failure management table 13, the page 16a holding the information 15 is prevented from being used for execution of other processing until failure recovery is completed.

However, the information management unit 19 inspects the failure management table 13 and finds the information position 1 without the failure device name 13a.
When there is 3c, it is judged that the failure recovery is completed, the information 1c is deleted from the failure management table 13, and the information 1
5 protection is terminated (step 28).

As an application example of the present invention, a method of arranging information in a cache memory or an arbitrary auxiliary storage device in which no failure has occurred can be considered between the occurrence of failure and recovery from failure. In this case, it is not necessary to protect the information on the main storage device, and the burden on the information management unit can be reduced. However,
In order not to affect other input / output to / from the auxiliary storage device for saving information, it is necessary to secure the auxiliary storage device having a low input / output frequency.

[0023]

As can be understood from the above description, according to the present invention, even if a failure occurs in one of the duplicated auxiliary storage devices under the condition of high input / output frequency, other input / output requests are eliminated. Since failure recovery processing is not performed up to this point, deterioration of I / O performance can be prevented, and a system with high load, high speed, and uniform performance can be realized. In addition, even if the information on the auxiliary storage device that survived from the occurrence of the failure to the failure recovery is lost due to the failure (double failure), the information to be stored at the failure location is saved in another storage device. Therefore, it is not irrecoverable, and high reliability due to the duplication of the auxiliary storage device can be maintained.

[Brief description of drawings]

FIG. 1 shows a configuration related to a redundant auxiliary storage device and its failure recovery processing.

FIG. 2 shows a procedure for failure recovery.

FIG. 3 shows a procedure for failure recovery of a conventional duplicated auxiliary storage device.

[Explanation of symbols]

 11 ... Failure reception section 12 ... Failure recovery section 13 ... Failure management table 14 ... Input / output table 15 ... Information 16 ... Storage device 17, 18 ... Duplicated auxiliary storage device 19 ... Information management section

Claims (1)

[Claims] 1. In an information processing system using a duplicated auxiliary storage device, when a failure occurs in one of the duplicated auxiliary storage devices, information to be stored at the location of the failure is provided. A redundant auxiliary storage device, characterized in that it is saved in another storage device until the failure recovery process is completed, and that the failure recovery process is executed when there is no I / O request to the failure occurrence auxiliary storage device. Disaster recovery method.