JPH0340152A - Fault restoring system for duplex semiconductor storage device - Google Patents

Abstract

PURPOSE:To avoid the discordance of storage contents between both semiconductor storage devices after the second application of a power supply by adding a latch relay, a copy mode F/F, and a saving function into each of semiconductor storage devices forming a duplex semiconductor storage device. CONSTITUTION:When a semiconductor storage controller 1 controls the copying operations between the semiconductor storage devices 10 and 20, the information is set to a copy mode F/F 12 of the storage device 10 set at the copy writing side to show that the storage device 10 is set at the copy writing side and kept in a copy mode. Then the copying operation is controlled to the storage device 20 for the abnormal end of a saving action by reference to the latch relays 12 and 23 of both storage devices 10 and 20 respectively when the power is supplied again after a service interruption. Thus the discordance of storage contents can be prevented between both duplex semiconductor storage devices 10 and 20.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a fault recovery method for a duplex semiconductor storage device.
In particular, the present invention relates to a failure recovery method for a duplex semiconductor memory device for preventing revision inconsistency of files in a duplex semiconductor memory device.

[Conventional technology]

Conventionally, in the failure recovery method of this type of duplex semiconductor storage device, when the revision of a file in the duplexed amphibonic conductor storage device has changed, the redundant semiconductor storage device with the new revision is replaced with the old one. By copying files to the semiconductor storage device on the revision side, inconsistency in file revisions is prevented.

On the other hand, since semiconductor storage devices are volatile storage devices, when the power supply to the semiconductor storage device stops, the stored contents (files) are stored on a save disk (magnetic disk for saving, etc.) by a bank amplifier such as a bank amplifier. It is necessary to save by power supply. Therefore, each double-handed conductor memory device in a duplex semiconductor memory device is also provided with a save disk, and in the event of a power outage (stoppage of power supply), each memory content is saved by the bank amplifier power supply.

Here, if a power outage occurs while files are being copied for failure recovery (to prevent inconsistent revisions of files in the two-handed conductor storage device), the memory contents of the two-handed conductor storage device will be The data was saved to the save disk without mismatching, and the information that it was being copied at the time of the power outage was not retained after the power was restored.

[Problem to be solved by the invention]

In the above-mentioned conventional fault recovery method for redundant semiconductor storage devices, if a power outage occurs while files are being copied between semiconductor storage devices, the storage contents of the two-handed conductive storage devices will be mismatched (
The data is saved to the save disk with inconsistent file revisions), and the information that it was being copied at the time of the power outage is not retained even after the power is resupplied. This has the disadvantage that the stored contents of the storage device remain inconsistent and normal failure recovery cannot be achieved.

In view of the above-mentioned points, an object of the present invention is to provide a fault recovery method for a duplex semiconductor memory device that can prevent the storage contents of the duplex two-handed conductor memory device from becoming inconsistent even in the above-mentioned cases. The goal is to provide a method.

[Means to solve the problem]

The fault recovery method for the duplex semiconductor storage device of the present invention is as follows:
A latch relay in each semiconductor storage device that retains information indicating abnormal termination of a save operation even after power supply is stopped, and a semiconductor storage device on the writing side of the copy in a copy operation between semiconductor devices to indicate this fact. The copying F/F in the semiconductor storage device that retains information and the copying F/F in the semiconductor storage device during the save operation in each semiconductor storage device by the bank amplifier power supply after the original power supply is stopped. If F holds information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy, sends information indicating abnormal termination of the save operation to the latch relay in the semiconductor storage device and saves. A save function in the semiconductor storage device that abnormally terminates the operation, and the copying F/F in the semiconductor storage device on the writing side of the copy when controlling the copy operation between the semiconductor storage devices.
information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy is sent, and the save operation is abnormally terminated based on reference to the latch relay in each semiconductor storage device when power is supplied again after a power outage. and a semiconductor storage control device that controls a copy operation to the semiconductor storage device according to the invention.

[Effect]

In the fault recovery method for a redundant semiconductor storage device of the present invention, a latch relay in each semiconductor storage device retains information indicating abnormal termination of a save operation even after the power supply is stopped, and a copy operation between semiconductor devices The copying F/F in the semiconductor storage device on the writing side of the copy holds information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy, and the save function in the semiconductor storage device is switched off from the original power source. Information indicating that the copying F/F in the semiconductor storage device is in the process of copying and is on the writing side of the copy during a save operation in the semiconductor storage device using the bank amplifier power supply after the supply of the bank amplifier is stopped. , the semiconductor storage controller sends information indicating abnormal termination of the save operation to the latch relay in the semiconductor storage device to abnormally terminate the save operation, and the semiconductor storage control device controls the copy operation between the semiconductor storage devices. Information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy is sent to the copying F/F in the semiconductor storage device on the writing side of the copy, and when the power is resupplied after a power outage, each semiconductor storage Based on reference to the latch relay in the device, the copy operation to one device is controlled based on the abnormal termination of the save operation.

〔Example) Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of a fault recovery method for a duplex semiconductor memory device according to the present invention. The fault recovery system for a duplexed semiconductor storage device according to this embodiment includes a semiconductor storage control device 1, semiconductor storage devices 10 and 20, and save disks 14 and 24.

The semiconductor storage device 10 or 20 has a save function 11 or 21 (assumed to be realized as a function of a microprogram in the semiconductor storage device 10 or 20) and a copying F/F (FI i p/F). 1 op
.

flip-flop) 12 or 22 and latch relay 1
3 or 23.

Referring to FIG. 2, the processing of the save function 11 or 21 includes a latch relay step 101, a copying F/F cent determination step 102, a save operation step 103, a save operation normality determination step 101,
It consists of a latch relay recent step 105.

Next, the operation of the fault recovery method of the duplex semiconductor memory device of this embodiment configured as described above will be explained. In addition,
Here, a duplex semiconductor memory device (semiconductor memory device IO
Regarding the operation when copying from the semiconductor storage device 10 to the semiconductor storage device 20 is performed under the control of the semiconductor storage control device 1 in order to recover from a failure in the semiconductor storage device (semiconductor storage device duplicated by the semiconductor storage device 10 and 20) explain.

In this copying, the memory contents for one block are read from the semiconductor memory device 10, the memory contents for the one block are written to the semiconductor memory device 20, and these reading and writing operations are repeated for the number of files.

During the above-described copying, the semiconductor storage device 1 sets the copying F/F 22 in the semiconductor storage device 20 on the copy writing side (copied side; old revision side).

A save operation when a power outage occurs during such copying and a save request is generated will be described with reference to FIG. 2.

Power for this save operation is supplied by a backup power source such as a power source, which is driven after the original power supply is stopped.

Further, this save operation is controlled by save functions 11 and 21, which are functions of microprograms in semiconductor memory devices 10 and 20. That is, the microprogram that receives the save request confirms that the request is a save request, and operates the save functions 11 and 21. As a result, each of the save functions 11 and 21 performs the following operations.

First, the operation of the save function 11 in the semiconductor storage device 10 on the copy reading side (the new revision side that is the copy source) will be described.

First, the latch relay 13 is set (as described later, this state can indicate "abnormal end of save operation") (step 101).

Here, the latch relay 13 is a relay that maintains its state even if the supply of power (original power and backup power) is stopped.

Next, it is determined whether or not the copying F/F 12 is being written (whether or not the semiconductor storage device 10 was in the process of copying and was on the writing side of the copy at the time of the original power outage) (step 102).

As a result of this determination, the copying F/F 12 has not been read, so a save operation is performed to save the memory contents of the semiconductor storage device 10 to the save disk 14 (step 1).
03), it is determined whether this save operation is normal (step 104).

If it is determined that the save operation is normal, the latch relay 13 is reset (step 105), and the save operation is normally completed.

Note that if the save operation is not normal as determined in step 104 (assumed to be normal in this embodiment), the latch relay 13 is left as is (remains in the seso I state) and the save operation is abnormally terminated.

Next, the operation of the save function 21 in the semiconductor storage device 20 on the copy writing side will be described.

First, the latch relay 23 is set (step 101).
).

Next, it is determined whether or not the copying F/F 22 is being sent (step 102).

In this judgment, since the F/F 22 during copying has been sent, the save operation is abnormally terminated while leaving the latch relay 23 as is (the semiconductor storage device 20 is transferred to the save disk 24).
(The contents of the memory will not be saved.)

When the save operation by the save functions 11 and 21 as described above is completed, the supply of backup power is stopped (even after the supply of backup power is stopped, the latch relay 23
(The current state of the data is maintained, and the information indicating the abnormal termination of the save operation in the semiconductor memory device 20 is retained.)

When the original power supply is resumed, an attempt is made to start up the semiconductor storage devices 10 and 20 from the control number of the semiconductor storage control device 1. (See latching relays 13 and 23 in storage devices 10 and 20).

Here, since the latch relay 13 of the semiconductor storage device 10 has been reset, normal startup is performed.
The storage contents are stored from the save disk 14 to the semiconductor storage device 10.

On the other hand, regarding the semiconductor storage device 20, the latch relay 2
3 remains in the cent state, startup is not performed (memory contents are not stored from the save disk 24 to the semiconductor storage device 20, and the semiconductor storage device 20 becomes unusable).

Based on recognition of the unusable state of the semiconductor storage device 20, the semiconductor storage control device 1 resets the latch relay 23 and starts controlling the copy operation from the semiconductor storage device IO to the semiconductor storage device 20.

In this embodiment, we have described the case in which the save functions 11 and 2 are realized as functions of a microprogram, but it goes without saying that the same functions can be realized by hardware means. .

〔Effect of the invention〕

As explained above, the present invention provides a latch relay, a copying F/F, and a save function in each semiconductor storage device constituting a duplex semiconductor storage device, thereby preventing copying due to a power outage during copying. If the storage contents of each semiconductor storage device are saved due to interruption of the operation, the storage contents of the bidirectional conductor storage devices will become inconsistent after power is supplied again (inconsistency in file revisions will occur). It has the effect of being able to prevent it from happening.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a flow chart showing the processing of the save function in FIG. In the figure, 1... Semiconductor storage control device, 10, 20... Semiconductor storage device, Z 21, 22, 23, 24... - Save function, - F/F during copying, - Latch relay save disk It is.

Claims (1)

[Scope of Claims] A latch relay in each semiconductor storage device that retains information indicating abnormal termination of a save operation even after power supply is stopped, and a semiconductor memory on the writing side of a copy in a copy operation between semiconductor devices. The F/F in the semiconductor storage device during copying that retains information indicating that fact in the device, and the F/F in the semiconductor storage device during the save operation using the backup power supply after the original power supply is stopped. If the copying F/F holds information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy, the latch relay in the semiconductor storage device indicates an abnormal end of the save operation. A save function in the semiconductor storage device that sets information and abnormally terminates the save operation, and a save function in the copying F/F in the semiconductor storage device on the writing side of the copy when controlling the copy operation between semiconductor storage devices. Information indicating that the semiconductor storage device is in the process of copying and is on the writing side of the copy is set, and when the power is resupplied after a power outage, the save operation is abnormally terminated based on reference to the latch relay in each semiconductor storage device. 1. A fault recovery method for a duplex semiconductor storage device, comprising: a semiconductor storage control device that controls a copy operation to such a semiconductor storage device.