Re:Asynchronous filesystem replication with FUSE and rsync
by icebraining on Thursday December 11, 2014 @11:49AM
(#8735716)
Sounds clear to me:
Because of no locks, the last write wins according to rsync update semantics based on file timestamp + checksum. In case two files have the same mtime, rsync compares the checksum to decide which one wins.
Because of no locks, the last write wins according to rsync update semantics based on file timestamp + checksum. In case two files have the same mtime, rsync compares the checksum to decide which one wins.
https://github.com/immobiliare/sfs#consistency-model