Borg FUSE mount maintains a persistent repo lock

[marekkrk]

When an entire repository is mounted with "borg mount", the corresponding FUSE process maintains a constant lock on the repository. This prevents "borg create" from creating new archives in the locked repository. New archives can only be created once the repository is unmounted.

Is it possible for the FUSE code to only acquire locks when read operations are performed? That way it should be possible to add new archives while the repository is mounted as a directory.

My use case for that is using a borg repository as a deduplicating version control system for large binary files with ability to mount a repo as a read-only file system.

[ThomasWaldmann]

Yes, that is known and this likely will be solved with borg 2 (which works very differently already, but is still in beta and not to be used in production).

Due to the way how it works, borg 1.x does exclusive locking for most operations.