The WAL-G team is pleased to announce the release of version 3.0.8 of WAL-G.
WAL-G is a tool for archival database restoration for PostgreSQL and several other databases. This release includes several improvements relevant to PostgreSQL users: * Direct-IO Reader for PostgreSQL ([#2003](https://github.com/wal-g/wal-g/pull/2003)) - This can improve performance of database during backup operations by bypassing the operating system's page cache. * Support for xid64 patch page layout ([#2040](https://github.com/wal-g/wal-g/pull/2040)) - WAL-G now works correctly with PostgreSQL instances using the xid64 patch for extended transaction IDs. * Fixes for WAL page size, block size, and checksums ([#1968](https://github.com/wal-g/wal-g/pull/1968), [#2135](https://github.com/wal-g/wal-g/pull/2135)) - These improvements enhance the reliability of backup and restore operations. * Reduced memory usage ([#2056](https://github.com/wal-g/wal-g/pull/2056)) - PostgreSQL and Greenplum operations now use less memory, making WAL-G more efficient for large-scale deployments. * Improved error handling ([#2021](https://github.com/wal-g/wal-g/pull/2021), [#2061](https://github.com/wal-g/wal-g/pull/2061)) - Better handling of corrupted WAL metadata and connection issues makes the tool more robust in production environments. * Retries for restore point creation ([#2000](https://github.com/wal-g/wal-g/pull/2000)) - This improves reliability during point-in-time recovery operations. Additionally, this release includes new commands for continuous recovery in Cloudberry and GreenplumDB, enhanced storage backend support (including Alibaba Cloud OSS with multipart upload), and various bug fixes and dependency updates. WAL-G v3.0.8 is available for download on our [GitHub releases page](https://github.com/wal-g/wal-g/releases/tag/v3.0.8). Have a nice day!
