Errors or typos? Topics missing? Hard to read? Let us know or open an issue on GitHub.

See also: mount, How to share data with an instance

In Multipass, a mount is a directory mapping from the host to an instance, making its contents, and changes therein, available on both ends.

Because mounts are performed as privileged users (root on Linux and macOS, SYSTEM on Windows), they allow write access to the whole host operating system.

On Linux and macOS, only privileged users (members of sudo, wheel, admin groups) can use Multipass, so this isn’t a concern.

On Windows, mounts are disabled by default, as anyone with TCP access to localhost ( can use Multipass, and by extension, gets access to the entire filesystem.

In Multipass, there are two types of mounts: classic (default) and native.

Classic mounts

Classic mounts use SSHFS (SSH File System) to achieve file/directory sharing. This option is available across all our backends.

SSHFS is based on SSH, which pays a performance penalty to achieve secure communication.

Native mounts

Native mounts use driver-dependent technologies to achieve the high performance. They are only available in the following cases:

  • On Hyper-V, where they are implemented with SMB/CIFS.
  • On QEMU, where they are implemented with 9P.
  • On LXD, using that backend’s own mounts, which also rely on 9P.

See also: Driver (backend) - Feature disparities

Last updated 11 days ago. Help improve this document in the forum.