Cluster Management Computer for Rapid Server Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cluster systems face challenges in quickly switching to a backup server due to differences in disk images and service states, leading to incomplete failover and obsolete backup images.
Innovation Solution
A management computer system that generates restoration pairs of active and auxiliary servers with common software configurations, allowing for efficient backup image distribution and quick switching by identifying and restoring common images and differential data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If a disk image is distributed to the auxiliary server in advance, then the failover time is reduced, but the disk image becomes obsolete when the service configuration changes
Solution Approach 1:
The patent applies preliminary action by distributing a common disk image to the auxiliary server in advance, before failure occurs. This pre-positioning of the disk image enables rapid failover when the active server fails, as the image is already available on the auxiliary server rather than needing to be transferred during failover.
Solution Approach 2:
The patent implements feedback by monitoring configuration changes on the active server and automatically updating the common disk image on both active and auxiliary servers when changes are detected. This feedback mechanism ensures the disk image remains valid and current, preventing obsolescence while maintaining the benefits of pre-distribution.
2Measurement precision
If the disk image is updated to reflect current service state, then the backup accuracy is improved, but the complexity of image management increases
Solution Approach 1:
The patent merges the disk images of the active server and auxiliary server into a single common disk image that is identical on both systems. This consolidation simplifies image management by eliminating the need to maintain separate images, while still achieving accurate backups through the feedback mechanism that updates the common image when changes occur.
Solution Approach 2:
The feedback mechanism automatically detects configuration changes and triggers updates to the common disk image, eliminating the need for manual image management interventions. This automated feedback loop maintains backup accuracy without increasing operational complexity, as the system self-manages the updates.
3Speed
If a common disk image is used for both active and auxiliary servers, then the failover speed is improved, but the ability to support different service configurations is reduced
Solution Approach 1:
The patent creates a universal common disk image that serves multiple functions: it acts as the boot image for both active and auxiliary servers, and dynamically adapts to reflect different service configurations. The feedback mechanism ensures the common image contains the correct configuration data regardless of which server is active, providing both speed and adaptability.
Solution Approach 2:
The common disk image is made dynamic rather than static - it automatically updates its contents based on configuration changes detected through the feedback mechanism. This dynamic nature allows the same common image to adapt to different service configurations while maintaining the failover speed benefits of pre-positioning.
Data Source
AI summary
A computer system, comprising: a plurality of computers; and a management computer for managing a cluster which includes the plurality of computers, the cluster include a plurality of first computers each of which executes a service and a plurality of second computers for taking over the service, wherein the management computer includes: a backup obtaining module; and a restoration pair management module, wherein the restoration pair management module is configured to: sort the plurality of first computers into a plurality of groups of first computers by which a software configuration for providing the service is common; generate at least one restoration pair for each of the plurality of groups by associating at least one first computer that is included in one group with at least one second computer; determine a common image for each of the plurality of restoration pairs.