File Systems

The following file systems features and enhancements are introduced in UEK 8:

Btrfs

The following notable Btrfs file system changes are introduced in UEK 8:

  • Compressed data can be sent or received without transformation, and data chunks larger than 64K are now handled for writes.

  • Quota accounting is simplified. Simple quotas can be used instead of quota groups for straightforward tracking of space usage by linking extents to their subvolumes. This approach can improve performance, but simple quotas are unable to track shared data, so are best suited to environments where extents are immutable and persist longer than any copies.

  • The introduction of a temporary FSID makes it possible to mount cloned devices. The file system gets a randomly generated UUID on mount.

  • Improved NOCOW write checks improve throughput by 9%.

  • A new mount option discard=async is enabled by default for devices that support trim/discard, applying asynchronous discard for the whole file system.

  • The mount option ignoremetacsums ignores invalid metadata checksums, and the ignoresuperflags mount option can be set to ignore superblock flags tracking conversion progress.

  • Send and relocation tasks, such as balance, device removal, shrink, and block group reclaim, run in parallel.

  • Devices can be added during a paused balance.

XFS

The following notable XFS file system changes are introduced in UEK 8:

  • You can now mount a file system with the blocksize larger than the pagesize.

  • Large extent counts are available for big virtual disk images.

  • Atomic file content commits are now available.

  • Fully autonomous online fsck and repair are available as a technical preview.

  • An update to the mkfs.xfs command sets a minimum XFS file system size to 300 MB to prevent the creation of small file systems that caused performance and redundancy problems. This change differs from the command included in the earlier xfsprogs package available in the ol9_baseos_latest repository on Oracle Linux 9 systems.

NFS

The following notable NFS file system changes are introduced in UEK 8:

  • NFSv4.2 READ_PLUS feature is enabled by default within the kernel to improve handling of sparse files by including a description of holes, or data blocks that are uninitialized.

  • Various older protocol features for NFS are removed in UEK 8. See Deprecated and Removed Features.