A FUSE passthrough filesystem with asynchronous replication and self-healing. helmetfs sits transparently between your applications and disk, adding BLAKE3 checksums, background replication to a second directory, and periodic scrubbing to detect and repair silent data corruption.
helmetfs mount <source-dir> <mountpoint> --replica <replica-dir> [options]helmetfs unmount <mountpoint>| Flag | Description | Default |
|---|---|---|
--replica <path> |
Replica directory (required) | -- |
--replication-workers <n> |
Number of background worker threads | 4 |
--scrub-time HH:MM |
Daily scrub schedule (24-hour format) | 01:00 |
--metrics-addr :PORT |
Enable Prometheus metrics endpoint | off |
mkdir -p /data/backing /data/replica /mnt/protected
helmetfs mount /data/backing /mnt/protected --replica /data/replica --scrub-time 03:00 --metrics-addr :9090Use /mnt/protected as a normal filesystem. Data lives in /data/backing and is asynchronously replicated to /data/replica.
Internal files (.sum sidecars and the .helmetfs/ directory) are hidden from the mount.