Memory Dump Handling for Rapid Server Reactivation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for storing memory dumps during system failures in high-capacity servers take a long time, leading to delayed system reactivation and potential secondary damage due to reactivation processes.
Innovation Solution
An information processing apparatus with a panic process unit, mapping process unit, reactivation process unit, and memory dump process unit that assigns a secondary memory area for the operating system, allowing for prompt reactivation and memory dump generation without overwriting critical data, using Memory Dynamic Reconfiguration to manage memory areas and avoid secondary damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If memory dumps are stored in a disk, then memory dump collection is possible, but system reactivation is delayed due to long copying time
Solution Approach 1:
The patent transitions from storing memory dumps on disk (one-dimensional storage) to using a second memory area (second-dimensional storage). This dimensional change enables much faster access and collection times while maintaining complete memory dump capability, directly resolving the contradiction between reliable dump collection and fast system reactivation.
Solution Approach 2:
The patent pre-allocates a second memory area before system failure occurs. When a panic happens, the memory dump can be collected immediately into this pre-prepared space without any allocation delay or disk I/O operations, enabling rapid system reactivation while ensuring complete and reliable dump collection.
2Loss of time
If memory dumps are stored in another memory, then system reactivation is faster, but memory dump collection takes long time for large sizes
Solution Approach 1:
The patent uses a second memory area as a dedicated dump collection space, transitioning from disk-based to memory-based storage. This provides both fast system reactivation (since memory is much faster than disk) and fast dump collection (since the entire dump process occurs in memory without I/O bottlenecks), resolving both contradictions simultaneously.
3Reliability
If the operating system collects dumps after detecting a fatal error, then dump collection is possible, but secondary damage such as hang-ups may occur
Solution Approach 1:
The patent extracts the dump collection function from the failed operating system and assigns it to a second operating system. This separation ensures that the original failed OS cannot cause secondary damage (like hang-ups) during dump collection, while the second OS safely performs the collection in an isolated environment with dedicated memory resources.
Solution Approach 2:
The second operating system acts as an intermediary between the failed first OS and the dump collection process. It safely interfaces with the first OS's memory space to collect dumps without being affected by the original failure conditions, preventing secondary damage while ensuring reliable dump collection.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An information processing apparatus that executes an operating system, the apparatus including a panic process unit configured to stop the operating system when the operating system has detected an error, a mapping process unit configured to assign, to the operating system stopped by the panic process unit, a second memory area which is other than a first memory area being used by a kernel of the operating system before stop or by a hypervisor that controls the operating system before stop of the operating system, a reactivation process unit configured to reactivate the operating system by using the second memory area as a usage area, and a memory dump process unit configured to read data in the first memory area, and to write the data to a dump file after the operating system is reactivated.