I am using ceph (master branch) + nfs-ganesha(master branch), and mount a bucket with sync option.
But even using sync option, the write offset sent by nfs client can also may overlap, and rgw file checks the offset is non-contiguous write position and return IO ERROR.
- Backport set to luminous, jewel, kraken
- Status changed from New to Need Review
- Assignee set to Matt Benjamin
- Status changed from Need Review to Pending Backport
- Backport changed from luminous, jewel, kraken to luminous, jewel
Matt, kraken is EoL - removing on the assumption it was added by mistake. Feel free to reinstate it if there is a particular reason why it is needed.
Sorry, thanks for correcting.
Also available in: Atom