Worked on stabilizing VIP-based service update synchronization in the kubeovn/kube-ovn repository, focusing on backend development and networking with Go and Kubernetes. Addressed a race condition that caused inconsistent updates when multiple services shared the same VIP, which previously led to endpoint slice churn and loss of SwitchLBRules. The solution involved re-queuing affected endpoint slices and ensuring SwitchLBRules were preserved during service updates, directly resolving issues #5597 and #5646. This targeted bug fix improved the reliability of service updates in environments with shared VIPs, demonstrating a methodical approach to maintaining consistency across distributed networking resources.