January 2025 monthly summary for longhorn-manager: Stabilized backup system operations and tightened resource management to reduce downtime and improve reliability. Key work focused on eliminating a resource leak in the backup controller and ensuring safe, conflict-free application of data engine settings during detachment. The changes enhance backup workflow predictability and overall system stability in production.

December 2024 – Longhorn Manager: Delivered a targeted bug fix to stabilize RWX fast-failover when a node becomes not ready. Implemented clearDelinquentLeasesIfNodeNotReady to clear delinquent RWX leases for volumes owned by not-ready nodes, preventing incorrect fast-failover status. The patch is implemented in commit f37529b99dd9004096543725a0361b754f01d696 with message 'fix: clear fast-failover status when node goes not ready'.

November 2024 was focused on reliability, cross-version compatibility, and observability across Longhorn components. Key features were delivered to centralize storage operations and unify logging, while several cross-version logging fixes improved operational stability. The work reduces maintenance overhead, enables safer upgrades, and enhances debugging capabilities for operators and developers. Key outcomes: - Centralized and delegated RWX filesystem resize logic to the share-manager, ensuring consistent behavior across storage types. - RPC-based logging configuration for the instance-manager with cross-version compatibility, improving reliability of log level application across data engine versions. - Unified logging system across V1 and V2 in the longhorn-instance-manager, including explicit separation of internal vs SPDK logging configurations, addressing V1 log level handling and preventing internal log levels from impacting SPDK settings.