Multipath Driver Cognitive Analysis for SAN Intermittent Faults
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage area network (SAN) systems face challenges in isolating and resolving recurring intermittent failures, which can lead to performance degradation and application failures due to the elusive nature of these faults, causing repeated recovery loops and false triggers.
Innovation Solution
A multipath driver cognitive analysis method that detects recurring intermittent errors by analyzing historical fault data and path management algorithms, implementing additional path states to handle transient recovery operations, and using timeout policies to manage path health checks and selection algorithms, thereby preventing cycling between on-line and off-line states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If path recovery actions are automatically initiated to remedy intermittent failures, then system availability is improved, but recurring intermittent faults cause repeated recovery loops that degrade performance
Solution Approach 1:
The system dynamically adjusts path states between on-line, off-line, and degraded sub-state based on real-time error detection and cognitive analysis. Paths transition through different states depending on their health status, allowing the system to adapt to changing conditions and avoid rigid recovery loops while maintaining availability.
Solution Approach 2:
The multipath driver implements cognitive analysis that continuously monitors path errors and provides feedback to adjust path selection and recovery actions. This feedback mechanism prevents repeated recovery loops by learning from historical error patterns and making intelligent decisions about when to retry paths versus when to leave them degraded, thereby maintaining both availability and performance.
2Reliability
If paths are taken offline to circumvent solid failures, then fault tolerance is achieved, but intermittent faults cause unnecessary path takeovers that degrade performance
Solution Approach 1:
The system changes the state parameter of paths based on error analysis, introducing a degraded sub-state that allows paths to remain partially available rather than being completely taken offline. This parameter change enables the system to maintain fault tolerance while avoiding unnecessary path takeovers that would degrade application performance.
Solution Approach 2:
Path states are dynamically adjusted based on real-time error detection and cognitive analysis. Rather than statically taking paths offline, the system dynamically transitions paths between on-line, off-line, and degraded states, allowing for more nuanced fault tolerance that preserves performance when possible.
3Reliability
If multiple path states are implemented to handle transient recovery operations, then recovery loop cycling is prevented, but device complexity increases
Solution Approach 1:
The path state machine is segmented into distinct states (on-line, off-line, degraded sub-state) with clear transition rules for each. This segmentation provides structured control over recovery operations, preventing cycling by ensuring paths progress through defined states rather than oscillating between on-line and off-line, while keeping complexity manageable through clear state boundaries.
4Reliability
If paths are monitored with health checks and timeout policies, then recurring intermittent errors are detected, but measurement precision requirements increase
Solution Approach 1:
The multipath driver acts as an intermediary that monitors path health through controlled interactions (health checks, timeout policies) without requiring direct observation of underlying hardware faults. This intermediary approach detects recurring intermittent errors through observable behavior patterns rather than requiring precise measurement of the root cause, reducing measurement precision requirements while maintaining detection accuracy.
Data Source
AI summary
An aspect includes detecting a recurring intermittent error in a path of a network in a system that includes at least one data transmission port configured for connection to at least one shared data storage device via a plurality of paths of the network. It is determined whether a cause of the recurring intermittent error is a previous path recovery action. In response to determining that the cause of the recurring intermittent error is not a previous path recovery action, the data transmission port is prevented from accessing the path for a specified time period by moving the path into a degraded sub-state, and subsequent to the specified time period the data transmission port is provided access to the path. In response to determining that the cause of the recurring intermittent error is a previous path recovery action, the data transmission port is provided access to the path.