JPH0335339A - Settlement processing system for occurrence of os fault - Google Patents

Abstract

PURPOSE:To prevent the erasion of the buffing data on a main storage and the data on a cache memory of an auxiliary storage device by using a settlement processing program which is started at occurrence of an OS (operating system) fault. CONSTITUTION:A data store area 13 is secured in an auxiliary storage device 11. At the same time, a 1st store means is added to a settlement processing means 17 which is started at occurrence of a fault of an OS 16 and performs a settlement process to write the data stored in a cache memory 12 of the storage device 11 to the corresponding area of the storage 11. Then a 2nd store means writes the data stored in a buffer 15 of a main storage 14 into the area 13 of the storage device 11 or the corresponding area of the storage device 11 in accordance with a fact whether the factor of the OS fault is caused by the buffing control of the buffer 15 or not. In such a constitution, the erasion of data can be prevented even if the OS 16 has a fault in a state where a write request is through logically and not through physically to the storage device 11.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) The present invention relates to a cleanup processing method when a failure occurs in an O5 (operating system), and in particular, it relates to a method for cleaning up when a failure occurs in an O5 (operating system), and in particular, it involves buffering on main memory, The present invention relates to a method for storing data in a cache provided in an auxiliary storage device.

(Prior Art) In a data processing system equipped with an auxiliary storage device such as a magnetic disk device, input/output operations to the auxiliary storage device are extremely slow compared to input/output operations to the main memory (main memory access). In order to avoid the occurrence of physical input/output operations to the auxiliary storage device as much as possible (the apparent input/output speed of the auxiliary storage device is improved), cache memory was conventionally used as the function of the auxiliary storage device. Or, O
Normally, data for write requests to the auxiliary storage device is buffered in the main memory (in a buffer secured in the main memory) by the input/output control (input/output control program) of the S.

In such systems, when a user (application program) issues a request to update data on the auxiliary storage device, the data buffered in the main memory is not necessarily written to the auxiliary storage device immediately; Generally, actual output (writing) to the auxiliary storage device is not performed until a condition such as a storage buffer becomes full (FULL) is met. Furthermore, writing to the auxiliary storage device in response to a request from the OS is only performed to the cache memory of the auxiliary storage device, and the writing may not actually be completed. In other words, in a conventional system equipped with an auxiliary storage device, when a write request is made to the auxiliary storage device from the user or the OS, even though the above writing (logical writing) is performed immediately, the actual writing is delayed. (Physical writing) may be executed with a delay.

As described above, in a conventional system equipped with an auxiliary storage device, there is a possibility that a large time difference may occur between the apparent completion timing and the actual completion timing of a write request from a user or the like. Therefore, if the process becomes unable to continue due to an OS failure before the physical write for the above request is completed, the buffering data on the main memory and the data on the cache memory of the auxiliary storage device will be lost. Put it away. To avoid such data loss,
Alternatively, in order to reduce the damage caused by data loss, it is conceivable to periodically write the data in the buffer in the main memory and the data in the cache memory in the auxiliary storage device to the auxiliary storage device, but this would result in an increase in overhead. , it is difficult to apply because it has a negative effect on performance.

(Issue 8 to be solved by the invention) As mentioned above, in conventional systems equipped with an auxiliary storage device, physical input/output operations to the auxiliary storage device are reduced as much as possible to increase the apparent input/output speed. If you try to increase the storage capacity, there will be many situations where a write request from the user or the OS to the auxiliary storage device is completed logically but not physically completed, and in such a physically incomplete state, the OS When a failure occurs, there is a problem in that buffering data in the main memory and data in the cache memory of the auxiliary storage device are lost.

This invention was made in view of the above circumstances, and its purpose is to create a state in which a write request from a user (application program) or an OS to an auxiliary storage device is logically completed but not physically completed. To provide a cleanup processing method when an OS failure occurs, which can prevent buffering data on a main memory and data on a cache memory of an auxiliary storage device from disappearing when a failure occurs in the OS.

[Structure of the Invention] (Means for Solving the Problems) This invention provides a cache memory in the auxiliary storage device, and also provides a cache memory in the auxiliary storage device in order to minimize the occurrence of input/output operations to the slow auxiliary storage device. In a system in which a buffer for temporarily storing data to be written is secured in the main memory, and the OS controls human output to the buffer in the main memory and to the auxiliary storage device, data is saved in the auxiliary storage device. A first step for securing an area and writing data stored in the cache memory of the auxiliary storage device to a corresponding area of the auxiliary storage device in a cleanup processing means that is activated and performs cleanup processing when a failure occurs in the OS. storage means, and the data stored in the buffer on the main memory is secured on the auxiliary storage device depending on whether the cause of the OS failure is related to the control of the buffer (buffering control). The present invention is characterized by providing a second storage means for writing to a data storage area or a corresponding area on an auxiliary storage device.

(Function) According to the above configuration, when a failure occurs in the OS, during the cleanup process, the stored data (unwritten data) in the cache memory of the auxiliary storage device is transferred to the correct area of the auxiliary storage device (the failure occurs in the OS). will be written to the area that was being attempted to be written before this occurred).

In addition, the data stored in the buffer on the main memory is handled differently depending on the cause of the OS failure, and if the cause of the OS failure is other than buffering control, it is written to the correct area of the auxiliary storage device, and the data is written to the correct area of the auxiliary storage device, and data related to buffering IIJm is handled differently depending on the cause of the OS failure. If so, it is written to a data storage area secured on the auxiliary storage device. Therefore, if the cause of the OS failure is other than buffering control, all data (unwritten data) stored in the cache memory of the auxiliary storage device and data stored in the buffer on the main memory are guaranteed. Furthermore, even if there is an OS failure factor within the buffering system, data can be recovered by reading data from a specific area (data storage area) on the auxiliary storage device.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of a data processing system to which the present invention is applied. In the same figure, 11
12 is an auxiliary storage device such as a magnetic disk device; 12 is a cache memory such as a disk cache provided in the auxiliary storage device 11 to prevent input/output operations from occurring to the auxiliary storage device 11 and improve input/output speed; 13 is a data storage area (for storing buffering data in the buffer 15 described later) secured in a specific area of (the auxiliary storage medium of) the auxiliary storage device 11. 14 is main memory, 15
is a buffer secured on the main memory 14 for storing write data (data of a write request from the user) to the auxiliary storage device 11. The main memory 14 includes an O3 (operating system) 1B, which is a system program for managing various system controls including human output control for the buffer 15 and the auxiliary storage device 11, and an operating system (O3) 1B that is used in the event of a failure in the OS 16. A cleanup processing program 17 and the like for performing cleanup processing are placed. 18 is O
This is a CPU for executing S1B, a cleanup processing program 17, and various application programs.

Next, the operation of one embodiment of the present invention will be explained.

First, a buffer 15 and an auxiliary storage device 1 secured on the main memory 14 are similar to those described in the [Prior Art] section.
Access to the cache memory 12 provided in 1 will be explained. When a write request to the auxiliary storage device 11 occurs due to the execution of an application program (that is, a write request from the user to the auxiliary storage device 11), it is difficult to check whether the writing to the auxiliary storage device 11 is actually performed each time. The input/output operation of the auxiliary storage device 11 is the main memory 1
This is very slow compared to accessing 4, resulting in poor efficiency.

Therefore, in the system shown in FIG. 1, a buffer 15 is secured on the main memory 14 under the management of 0816, and when a write request from a user to the auxiliary storage device 11 occurs, the requested data is stored in accordance with OS 1B. The CPU 18 temporarily stores the data in the buffer 15. At this time, the OS lli specifies that no output request is issued to the auxiliary storage device 11 until the write size requested by the user reaches the specified size. OS 1B
(CPU 1g that executes), when a write request from the user occurs, once the requested data is written to the buffer 15, the CPU 1g that executes the write request is sent to the user without issuing an output request to the auxiliary storage device 11. Notify completion. This notification indicates the completion of logical writing in response to the write request from the user (application program). It is not possible to determine whether it has been written or not. If this is the case, in which the data requested by the user has been written to the buffer 15 but not to the auxiliary storage device fill, the OS 1B may fail (OSIB cannot run). If a failure occurs and the system stops, conventionally the buffer 15
The stored data will be lost, and when the OS 16 is started next time, a contradiction will occur in the user data.

Now, in the system shown in Figure 1, the C code that executes 0516 (
Even if an output request is issued to the auxiliary storage device 11 from the auxiliary storage device 11 (auxiliary storage medium of the auxiliary storage device 1g)
Actual writing to is not necessarily performed. That is, the first
In the system shown in the figure, user data retrieved from the buffer 15 in response to an output request to the auxiliary storage device 11 is
A situation occurs in which data is stored (cached) in the cache memory 12 provided in the auxiliary storage device 11 and is not written to (the auxiliary storage medium of) the auxiliary storage device 11. Therefore, conventionally, if the OS 1B fails and stops in such a state, the data in the cache memory 12 will be lost.

By the way, the system shown in FIG. 1 is provided with a clean-up processing program 17 that performs clean-up processing of the OS 1B, such as collecting the cause of the failure and information for factor analysis when a failure occurs in the OSTE. There is.

This type of cleanup processing program is normally prepared in data processing systems, but in the system shown in FIG. 1, a new processing routine is added to the cleanup processing program 17 to process the data We are trying to prevent it from disappearing. Cleanup processing when an O3 failure occurs using the new cleanup processing program 17 will be described below with reference to the flowchart in FIG.

First, when a failure occurs in the OS 16, the cleanup processing program 17 is started (step S2).

As a result, the CPU 1B starts the cleanup process when an O3 failure occurs according to the cleanup process program 17, and first, the storage contents of the cache memory 12 of the auxiliary storage device 11 (of which data has not been written to the auxiliary storage device 11) ) to the correct location of the auxiliary storage device 11 (the auxiliary storage medium attached to it) (the location where the OS 1B was trying to write before the failure occurred).
Step S2).

As a result, an operation is performed to write the data remaining in the cache memory 12 (unwritten data) to the auxiliary storage device 11. Next, the CPU 8 checks whether the cause of the failure in the OS 1B is related to the control of the buffer 15 on the main memory 14, that is, the buffering control (step S3), and if it is not related to the buffering control. If so, the contents of buffer 15 on main memory i4 are saved to OS 1.
Auxiliary storage device 11 that 6 was trying to write to before the failure occurred
(step S4). On the other hand, if the cause of the failure in 0816 is related to buffering control, the contents of the buffer 15 on the main memory 14 may not be guaranteed, so the CPU 18
Instead, a process is performed in which the contents of the buffer 15 are written into the data storage area 13 secured in advance in the auxiliary storage device 11 (step S5). After completing step S4 or step S5, the CP 018 performs normal cleanup processing when an OS failure occurs (step S6). When this step S6 is completed, the OS is stopped (system stopped) (step S7).

As described above, according to this embodiment, the cleanup processing program 17 is activated when a failure occurs in the OSTE.
A new data storage function is realized using the , and user data stored in the cache memory 12 of the auxiliary storage device 11 and the buffer 15 on the main memory 14, which would have been lost in the past, is saved in the auxiliary storage device It. As a result, the user data is guaranteed the next time the OS is started. However, in the event of an O5 failure related to buffering control, it is necessary to read the data stored in the data storage area 13 of the auxiliary storage device 11 and perform data recovery.

Note that the present invention includes a power-off detection means for detecting the occurrence of a power-off, notifying the 031B, and calling the clean-up processing program 17, and a backup battery having a capacity necessary for the clean-up processing program 17 to complete the processing. With this, it is possible to save the contents of the cache memory 12 and buffer 15 in the auxiliary storage device 11 even when the power is turned off.

[Effects of the Invention] As detailed above, according to the present invention, if a failure occurs in the OS while a write request to the auxiliary storage device is logically completed but not physically completed. In addition, buffering data on the main memory and data remaining in the cache memory of the auxiliary storage device, which were lost in the past, can be saved in the auxiliary storage device, thereby preventing data loss.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a flowchart for explaining the operation.

Claims (1)

[Claims] An auxiliary storage device equipped with a cache memory, a main memory in which a buffer for temporarily storing data to be written to the auxiliary storage device is secured, and the above-mentioned buffer on this main memory and the above-mentioned auxiliary storage device. It is equipped with an operating system (OS) for managing various system controls including input/output control for the storage device, and instead of physical input/output processing for the auxiliary storage device, a buffer on the main memory or the auxiliary storage device is provided. In a data processing system that performs logical input/output processing to the auxiliary storage device by accessing the cache memory of the device, a failure occurs in the data storage area secured on the auxiliary storage device and the operating system. A cleanup processing means that is activated and performs cleanup processing when the
a first storage means for writing data stored in the cache memory of the auxiliary storage device to a corresponding area of the auxiliary storage device; and a first storage means for writing data stored in the cache memory of the auxiliary storage device to a corresponding area of the auxiliary storage device; cleanup processing means having a second storage means for writing to the corresponding area or the data storage area secured on the auxiliary storage device;
If the cause of the failure of the operating system is not related to the control of the buffer, write the data stored in the buffer to the corresponding area of the auxiliary storage device,
A cleanup processing method when an OS failure occurs, characterized in that when the cause of the failure is related to control of the buffer, data stored in the buffer is written to the data storage area.