关于Btrfs文件系统版本特性、空间开销及NAS部署的技术问询

Hey there! I’ve been working with Btrfs on NAS setups for a while, so let’s walk through your questions to ease those fears.

1. Does using Btrfs with Samba miss any important features? #

• Nope, you won’t lose any core Btrfs functionality by sharing it via Samba. Samba acts just as a bridge between your NAS and client computers—handling basic file read/write, permissions, and sharing. All the powerful Btrfs features (like snapshots, compression, RAID management) are still fully accessible via SSH as root on the NAS side.
• The only small caveat: Clients can’t directly interact with Btrfs-specific features (like mounting snapshots or managing subvolumes) through Samba. Those operations require root access on the NAS, but for everyday file sharing needs, Samba works perfectly with Btrfs, just like it does with ext4 or other filesystems.

2. Can I 100% recover a deleted file with Btrfs? #

• Btrfs has versioning capabilities, but this isn’t automatic—it relies entirely on snapshots that you (or an automated tool) create. If you delete a file without having a pre-existing snapshot of its parent subvolume/directory, you can’t recover it via Btrfs itself (you’d have to use generic data recovery tools, which aren’t 100% reliable).
• If you did take a snapshot before deleting the file, though? You can absolutely recover it 100% by accessing the snapshot. Think of snapshots like manual checkpoints—they don’t track every change automatically (unlike Git), but as long as you have that checkpoint, your deleted file is safe inside it.

3. Will free space run out quickly due to versioning? How to clean up non-current data? #

• This is a common misconception. Btrfs doesn’t hoard old versions unless you explicitly tell it to. If you never create snapshots or use tools that auto-generate them, your space usage will be identical to any other filesystem—only your current 5TB of data will take up space, leaving ~13TB free.
• When space does get eaten up by extra data, it’s almost always from snapshots. Thanks to Btrfs’s Copy-on-Write (CoW) system, snapshots only store the parts of files that have changed since the snapshot was taken—so they’re much more space-efficient than full backups. For example, modifying a 1GB file only adds the changed blocks to the snapshot, not the entire 1GB.
• Cleaning up is straightforward:
  • Delete individual snapshots with the command btrfs subvolume delete /path/to/snapshot
  • If you use a tool like snapper for auto-snapshots, it has built-in commands to delete old or unused snapshots in bulk
  • To fully wipe all non-current data, just delete all snapshots, then run btrfs filesystem sync followed by btrfs filesystem cleanup /mount/point to reclaim unused space. This will leave only your current, active files on the disk.

备注：内容来源于stack exchange，提问作者Chameleon