On Thu, Jan 16, 2020 at 4:34 PM Patrick Donnelly <pdonnell@redhat.com> wrote:
On Fri, Jan 10, 2020 at 4:22 AM Neal Gompa <ngompa13@gmail.com> wrote:
The current zfs2ceph implementation handles zvol sends and transforms them into rbd v1 import streams. I don't recall exactly the reason why we don't use v2 anymore, but I think there was some gaps that made it so it wasn't usable for our case back then (we were using Ceph Luminous). I'm unsure if this is improved now, though it wouldn't surprise me if it has. However, zvols aren't enough for us. Most of our ZFS datasets are in the ZFS filesystem form, not the ZVol block device form. Unfortunately, there is no import equivalent for CephFS, which blocked an implemented of this capability[3]. I had filed a request about it on the issue tracker, but it was rejected on the basis of something was being worked on[4]. However, I haven't seen something exactly like what I need land in CephFS yet.
The reason I closed the ticket is that I thought the request was for a mechanism to geo-replicate CephFS. There is on-going work to make that feasible lead by Jan and Venky.
If what you actually want is a way to import ZFS file systems, I would tell you that rsync is the answer. If you have a binary stream of a ZFS file system from `zfs export`, then you will need to import it into a new zfs file system and rsync that.
That's fairly painful if I want to preserve all the history of ZFS snapshots while not retaining ZFS at the end of it. Yes, the problem is a lot simpler if I just kept ZFS even on the other end. But I want to be able to map a ZFS send stream to a CephFS import stream. That should be technically possible, albeit complicated, if CephFS supported import/export like RBD does. The idea, after all, is to be pure Ceph at the end. -- 真実はいつも一つ！/ Always, there's only one truth!