JPH03105434A - Error logging system - Google Patents

Abstract

PURPOSE:To attain error logging as against the occurrence of plural faults in a computer system by providing plural error logging storage areas and a storage area control area. CONSTITUTION:Error logging information are recorded and held in the storage areas 41-4n. CPU 1 decides to which area error logging information is to be recorded by referring to the control area 3 controlling the use state of the storage areas, and new error logging information is written so that important information is not impaired. Error logging information outputted from the error logging storage areas by an operator is taken as disuse for storage in the control area 3, and subsequent error logging information when the fault occurs is treated so that it can be overwritten on the information. Error logging is performed as against the occurrence of plural faults according to the system.

Description

[Detailed Description of the Invention] [Summary] Regarding an error logging system when a failure occurs in a computer system, the purpose is to record error logging information when a system failure occurs, with the aim of enabling error logging even when multiple failures occur. The error logging system has a plurality of error logging storage areas and a storage area management means for managing the usage status of the plurality of error logging storage areas, and when a failure occurs, the system manages the storage area management area. For reference, error logging information is stored in a writable area among the plurality of error logging storage areas.

[Industrial application field]

The present invention relates to an error logging system when a failure occurs in a computer system.

If a failure occurs due to a hardware failure or program error while the computer system is in operation, automatic IPL (initial program loading) processing is performed as a recovery process in systems that operate continuously for two hours.

In such a case, in order to be able to analyze the cause of the failure later, the memory state immediately before the failure occurred (more precisely, just before automatic IPL), various internal registers'r, and hardware error information, etc. Error 1] Stored in auxiliary storage as logging. The system operator then automatically
After the completion of PE, errors in the auxiliary storage device were output and the information analyzed to identify faulty locations.

[Prior Art and Problems to be Solved by the Invention] Conventionally, in the above-mentioned error logging system, there is only one area for recording error logging information, as shown in FIG. For single failures, it was possible to analyze the cause of the error. However, if multiple failures occur and automatic IPL is performed continuously, it becomes difficult to analyze the cause of the failure, resulting in decreased system reliability and MTTR (rr+ean time to
The present invention has been made in view of the above problems, and provides an error logging system that enables error logging even when multiple failures occur. The purpose is to

[Means to solve the problem]

FIG. 1 is a diagram explaining the principle of the present invention.

In FIG. i, 1 is the CPU of the system. 2 is a system memory; 41 to 4n are storage areas for multiple error logging provided outside the system;
3 is a storage area management area (control area) that manages the usage status of the plurality of error logging storage areas.

[Effect]

When recording and retaining error logging information along with automatic IPL, CPU By referring to the control area 3, multiple error logging information storage areas 4l to 4n can be stored.
It is determined in which area error logging should be recorded, and new error logging information is written so as not to overwrite important information. For example, the error logging information output from the error logging storage area by the operator is treated as unnecessary to be stored in the control area 3, and the error logging information at the time of the next failure can be overwritten thereon.

〔Example〕

FIG. 2 is a diagram showing the structure of an error logging storage area and a control area in an embodiment of the error logging system of the present invention.

In FIG. 2, 30 is the aforementioned control area, 51 to 5n
are multiple error logging storage areas.

As described later, the error logging area is area 1 (51) that stores error logging at the time of the first failure.
and areas 2 to n (52 to 5n) for storing error logging in the event of a subsequent failure. The control area 30 has a start storage address, a maximum number of registrations, a current number of registrations (storage), a retrieval pointer, and a registration pointer.

The start storage address indicates the start address of area 1 (51) above, the maximum number of registrations indicates the number of storage areas that can store error logging, and the current number of registrations indicates the number of error loggings currently stored. and the fetch pointer is
It indicates the address of the storage area when error logging is read by the operator, and the registration pointer indicates the address of the storage area when error logging is stored.

FIG. 3 is a diagram showing the procedure for storing error logging in the embodiment of the present invention.

In step 6l of FIG. 3, it is determined whether the first storage address of the control area 30 is equal to the initial value xoooo. If the first storage address is the initial value xoooo, this indicates that the failure that has occurred is the first failure, so the process advances to step 62 and this error logging is specially stored in area 1 (51). do. in general,
Error logging when the first failure occurs provides particularly important information for failure cause analysis, so error logging when the first failure occurs will not be overwritten no matter how many failures occur thereafter. This is to ensure that.

If in step 61, the first storage address is the initial value
If not OOO, the process proceeds to step 63, where it is determined whether the current number of registrations is equal to the maximum number of registrations (n)-1 in areas 2 to n.

If in step 63 the current number of registrations is the maximum number of registrations (
n)-1, the process proceeds to step 64 and stores the error logging in one of the areas 2 to n indicated by the registration pointer. Then, in step 65, the current registration number is incremented, and in step 66, the registration pointer is updated within the range of areas 2 to n.

In step 63, the current number of registrations is set to the maximum number of registrations (n) - 1
If it is equal to
The new error logging is stored in the area where the oldest error logging is stored, that is, the storage area indicated by the registration pointer. Then, at step 68,
Both the retrieval pointer and the registration pointer are updated.

FIG. 4 is a diagram showing a procedure when reading error logging in an embodiment of the present invention.

In step 71 of FIG. 4, if the first storage address is xo o o o, this indicates that no failure has occurred, or at least that the storage of error logging has not been exhausted, so the process is continued. finish.

If the first storage address is not XOO00 in step 71, this indicates that a failure has occurred and error logging is disabled, so proceed to step 72 and deal with the first failure stored in area 1. Read the error logging and set the first storage address to XO O 0 0
Set to .

Then, in step 73, it is determined whether the current number of registrations held in the control area 30 is zero. If it is 0, there is no more error logging retained, so the process ends.

In step 73, if the current number of registrations held in the control area 30 is 0, the process advances to step 74, and the error logging stored in areas 2 to n is read out until the retrieval pointer catches up with the registration pointer. . Then, in step 75, the retrieval pointer is updated and the process ends.

〔Effect of the invention〕

According to the error logging system of the present invention, error logging can be performed even when multiple failures occur.

[Brief explanation of drawings]

FIG. 1 is a diagram explaining the principle of the present invention. FIG. 2 is a diagram showing the structure of the error logging storage area and control area in an embodiment of the error logging system of the present invention. FIG. 4 is a diagram showing the procedure for reading error logging in an embodiment of the present invention, and FIG. 5 is an explanatory diagram of a conventional error logging system. [Explanation of symbols] 1... System CPU, 2... System memory, 41 to 4n... Storage area for error logging, 3
...Storage area management area (control area).

Claims (1)

[Scope of Claims] 1. In an error logging system that records and retains error logging information when a system failure occurs, a plurality of error logging storage areas (41 to 4n) and the usage status of the plurality of error logging storage areas are managed. The system has a storage area management area (3), and when a failure occurs, the system refers to the storage area management area (3) and
An error logging system characterized in that error logging information is stored in a writable area among the plurality of error logging storage areas (41 to 4n).