Failure Analysis Model Training for Storage Log Packages
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Technical support engineers face challenges in efficiently analyzing failure causes in large-scale and complex storage systems due to the limitations of existing methods, which are primarily designed for individual log files rather than log packages, leading to increased time and expertise requirements for identifying root causes.
Innovation Solution
A failure analysis model is trained using preprocessed log files that include potential failure causes and associated scores, allowing for prediction and updating of model parameters based on the probability of the predicted failure cause, enabling efficient analysis of failure causes within log packages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing methods designed for individual log files are used to analyze failure causes in storage systems, then the analysis process can be performed with simpler tools, but the time required and expertise needed increase significantly when dealing with log packages from large-scale complex systems
Solution Approach 1:
The patent segments the log package analysis into multiple independent log file analyses. Each log file is processed separately to generate potential failure causes and scores, which are then aggregated. This segmentation allows parallel processing and reduces the overall time required for analyzing large log packages while maintaining comprehensive coverage of all potential failure causes.
Solution Approach 2:
The patent performs preliminary processing on each log file before aggregation, including extracting potential failure causes, calculating scores, and filtering relevant information. This preliminary action prepares the data in advance for efficient aggregation and analysis, reducing the time required for the final synthesis step and improving overall failure analysis productivity.
2Device complexity
If existing methods are used for analyzing log packages, then fewer resources are required, but the complexity of handling multiple log files and aggregating results increases the difficulty of the analysis process
Solution Approach 1:
The patent implements automated score calculation and aggregation mechanisms that perform the complex analysis operations without requiring manual intervention. The system automatically processes each log file, calculates failure cause scores, aggregates results, and generates the final failure cause determination. This self-service approach reduces operational complexity while maintaining ease of use for the end user.
3Measurement precision
If manual analysis by technical support engineers is performed on log packages, then expert judgment can be applied, but the cost in terms of engineering experience requirements and time consumption increases
Solution Approach 1:
The patent implements a feedback mechanism where the scores and potential failure causes from individual log files are aggregated and used to determine the final failure cause. The system continuously refines the failure cause determination by incorporating feedback from multiple log file analyses, improving accuracy while reducing the time required compared to manual expert analysis. The feedback loop enables automated systems to achieve expert-level precision.
Data Source
AI summary
Embodiments of the present disclosure include a method, an electronic device, and a computer program product for training a failure analysis model. In a method for training a failure analysis model in an illustrative embodiment, at least one set of log files including multiple preprocessed log files is obtained, the at least one set of log files including a marked failure cause of a storage system, and preprocessed log files in the multiple preprocessed log files including one or more potential failure causes of the storage system and scores associated with the potential failure causes; a failure cause of the storage system is predicted according to a failure analysis model and based on the potential failure causes and the scores in the multiple preprocessed log files; and parameters of the failure analysis model are updated based on a probability that the predicted failure cause is the marked failure cause.