Selective Node Failover Lock Coordination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional IP failover techniques in high-availability services require all nodes in a file system to enter a grace period, leading to costly service disruptions and potential violations of service-level agreements due to paused file access and lock coordination, affecting all nodes even if they are not participating in the failover.
Innovation Solution
Implementing a method where a grace period is initiated only for the failing and secondary nodes, allowing a new lock to be written on the secondary node while maintaining old locks on other nodes, transferring the old lock to the new one, and stopping the grace period, thereby limiting application pauses and file access restrictions to only the affected nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional IP failover techniques require every node in the file system to enter a grace period, then lock coordination consistency is maintained, but service disruption increases and productivity decreases
Solution Approach 1:
The patent segments the failover process by node participation status. Nodes are divided into two groups: those participating in the failover (first and second nodes) and those not participating (other nodes). The grace period is applied selectively only to participating nodes, allowing non-participating nodes to continue servicing lock requests without interruption. This segmentation resolves the contradiction by maintaining lock coordination consistency for affected nodes while preserving service availability for unaffected nodes.
Solution Approach 2:
The patent applies local quality by making the grace period effect local rather than global. Instead of applying the grace period uniformly across all nodes in the file system, it is applied locally only to the specific nodes involved in the failover process. This allows different parts of the system to have different operational characteristics: participating nodes experience the grace period for consistent lock coordination, while non-participating nodes maintain normal operation for sustained productivity.
2Reliability
If the grace period is applied to all nodes, then lock recovery consistency is ensured, but application pause time increases
Solution Approach 1:
The patent segments the impact of the grace period by identifying and separating nodes that need lock recovery consistency (participating nodes) from those that do not (non-participating nodes). By applying the grace period only to participating nodes, the system ensures lock recovery consistency where needed while minimizing application pause time overall, as non-participating nodes continue to service requests without interruption.
Solution Approach 2:
The patent applies partial action by implementing the grace period for only a subset of nodes rather than all nodes. This partial application of the grace period is sufficient to ensure lock recovery consistency for the failover scenario while avoiding the excessive time loss that would result from applying the grace period globally to all nodes in the file system.
3Reliability
If file access is locked during the grace period, then data consistency is maintained, but service-level agreements are violated
Solution Approach 1:
The patent segments file access locking by node participation status. File access is locked only for nodes participating in the failover (first and second nodes) during the grace period, while file access remains unlocked and available for nodes not participating in the failover (other nodes). This selective locking maintains data consistency for affected files and nodes while preserving service-level agreement compliance for unaffected operations.
Solution Approach 2:
The patent applies local quality by making the file access locking effect local to specific nodes and files involved in the failover. Instead of globally locking all file access across the entire file system, the locking is applied locally only where necessary to maintain data consistency during the failover. This allows service-level agreements to be maintained for non-affected file access operations.
Data Source
AI summary
The disclosed computer-implemented method for performing node failovers may include: (1) initiating, during a failover of a first node to a second node, a grace period for the first node and the second node; (2) writing a new lock, indicating an internet protocol (IP) address, to a memory device of the second node, while maintaining: (A) an old lock of the IP address in a memory device of the first node, and (B) locks of nodes other than the first and second nodes; (3) transferring the old lock from the first node to the new lock of the second node, where requests to change the locks of nodes other than the first and second nodes are denied during transferring; and (4) stopping the grace period. Various other methods, systems, and computer-readable media are also disclosed.