Automatic allocation method and device for dumps

Abstract

Embodiments of the invention relate to an automatic allocation method and device for dumps. The method comprises the steps of calculating a stack similarity score between every two dumps in unallocated dumps and allocated dumps, and determining all the allocated dumps with the stack similarity scores higher than a stack similarity score threshold as related allocated dumps; in response to operation of determining the related allocated dumps, calculating scores of other features of each dump in the related allocated dumps, wherein other features include at least one of recency of the dump associated with the dump, a state of the dump, a release number of the dump, a version number of the dump and a duplicate value of the dump; and calculating a total similarity score according to the stack similarity scores and the scores of other features, and performing automatic allocation on the unallocated dumps based on the total similarity score. According to the method and the device provided by the embodiments of the invention, the dumps can be automatically allocated quickly and accurately by calculating the stack similarity of the dumps and the scores of other features.

Description

technical field [0001] Embodiments of the present disclosure generally relate to the field of computers, and more specifically relate to a method and device for automatically allocating dumps. Background technique [0002] A crash dump is a snapshot of the state of a computer system or process when it crashed. The crash dump saves the fault information or environment information when the system crashes or is abnormal, so that relevant personnel can use it for troubleshooting and analysis. In large systems with multiple functional component areas, many crash dumps can be produced during development, testing, and deployment. After a dump is obtained, it is often necessary to determine which specific functional area in a large system should be responsible for dump analysis. [0003] Typically, crash dumps can be categorized or assigned to appropriate functional areas based on their cause, rationale, or signature. Signatures can be stack traces or other internal state in the ...

AI Technical Summary

Problems solved by technology

However, in the prior art, the manual allocation of crash dumps not only consumes a lot of time and energy, but also is limited by the knowledge of analysts, resulting in low accuracy of dump allocation.

Also, different functional areas can cause stack dumps to have similar or even the same, so classification based on information in dumped stacks alone is not accurate enough

Images