Live Kernel Dump Analysis for Near-Real-Time Cloud Issue Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing environments face issues with physical computing devices that impact virtual machine performance, leading to unexpected reboots, downtime, and service outages, with conventional live kernel dump processing being offline and inefficient, resulting in prolonged issue resolution times.
Innovation Solution
Implementing an agent that flags live kernel dumps for prioritized processing, compresses and encodes them, and queues them for expedited transmission to an analysis server for near-real-time issue resolution, using predefined identifiers to facilitate rapid issue detection and least disruptive remedial actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If live kernel dump files are processed offline with conventional methods, then complete analysis can be performed, but issue resolution time increases to 26 minutes causing virtual machine downtime
Solution Approach 1:
The agent performs preliminary actions by flagging the kernel dump file for prioritized processing, compressing it, and queuing it for expedited transmission before the issue fully impacts virtual machines. This preliminary preparation enables rapid analysis and resolution, reducing downtime from 26 minutes to approximately 3 minutes.
Solution Approach 2:
The system dynamically adjusts processing priorities by creating a prioritized queue for flagged kernel dump files separate from the normal processing queue. This dynamic prioritization ensures that critical issues are processed faster than routine diagnostics, enabling near-real-time issue resolution while maintaining normal operational flow.
2Measurement precision
If physical computing devices are rebooted to resolve issues offline, then complete diagnostic analysis can be performed, but virtual machine unavailability occurs during the reboot process
Solution Approach 1:
The agent extracts the kernel dump file from the running physical computing device and transmits it to the analysis server for processing. By extracting the diagnostic data externally rather than requiring system shutdown, the system maintains virtual machine availability while still obtaining complete issue information for accurate diagnosis.
Solution Approach 2:
The analysis server acts as an intermediary that receives, processes, and analyzes the kernel dump file remotely. This intermediary processing eliminates the need for direct physical access or system reboot to perform diagnostics, enabling accurate issue detection while keeping virtual machines running.
3Ease of manufacture
If virtual machines are moved to another physical computing device for offline diagnosis, then issue analysis can be performed, but service continuity is disrupted
Solution Approach 1:
Instead of moving virtual machines between physical devices, the system extracts diagnostic information (kernel dump files) from the running physical device and transmits it to the analysis server. This extraction approach provides diagnostic accessibility while maintaining service continuity, eliminating the need to disrupt virtual machine operations for diagnosis.
Data Source
AI summary
A method, computer program product, and computing system for detecting, using an agent, an issue associated with a target computing device. A kernel dump file is obtained by initiating a kernel dump concerning the issue from the target computing device to the agent. A flagged kernel dump file is generated by flagging the kernel dump file for prioritized processing on an agent processing queue. The flagged kernel dump file is queued in a prioritized kernel dump queue. The flagged kernel dump file is uploaded from the prioritized kernel dump queue to an analysis server for online resolution.