On Wed, Jun 05, 2019 at 02:16:43PM -0700, Patrick Donnelly wrote:
Hello Jan,
On Mon, Jun 3, 2019 at 2:26 AM Jan Fajerski <jfajerski@xxxxxxxx> wrote:
I've asked about this in IRC already, but due to timezone foo ceph-devel might
be more effective.
I was wondering if there was a plan or expectation of creating cephfs subvolumes
using a luminous ceph_volume_client on a nautilus cluster (or any other sensible
version combination)?
I do not think so. We do plan to have Nautilus clusters continue to
function with the old ceph_volume_client.py clients.
Currently this does not work, due to the volume client using the now removed
'ceph mds dump' command. The fix is straight forward, but depending on if that
should work this could be more complex (essentially making ceph_volume_client
aware of the version of the ceph cluster).
... so this is a bug. Is there a tracker ticket open for this yet?
Not yet, will open one.
I'm aware of the current refactor of the volume client as a mgr module. Will we
backport this to luminous?
No.
Or is there an expectation that the volume client and
the ceph cluster have to run the same version?
That's what we'd like yes. I think the tricky part is dealing with
applications (like Manila) using an older ceph_volume_client.py. We
could backport a switch in the library so that it uses the new `ceph
fs volume` commands if the cluster is Nautilus+. I'm not sure that is
really needed though.
We see a case like that (luminous manila client against a nautilus cluster), so
I'd argue something is needed. I have a patch for this (with a switch), though
it might be enough to simply use 'ceph fs dump' instead of 'ceph mds dump' (i.e.
without a switch). luminous has the fs command already after all. I'm just not
sure if that would break something else.
--
Patrick Donnelly, Ph.D.
He / Him / His
Senior Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
--
Jan Fajerski
Engineer Enterprise Storage
SUSE Linux GmbH, GF: Felix Imendörffer, Mary Higgins, Sri Rasiah
HRB 21284 (AG Nürnberg)
