[openstack-dev] [tripleo] FFE nfs_ganesha integration

Tom Barron tpb at dyncloud.net
Sun Feb 4 21:51:12 UTC 2018


Just to follow up, CI is passing for the three patches outstanding
and the last one has a release note for the overall feature.  The trick to
getting CI to pass was to introduce a new variant Controller
role for when we actually deploy with CephNFS and the VIP for the
server on the StorageNFS network.  Using the variant controller role and
'-n'
with network_data_ganesha.yaml (1) enables the new feature to
work correctly while (2) making the new feature entirely optional so
that current CI runs without being affected by it.

I think the three outstanding patches here are ready to merge:

https://review.openstack.org/#/q/status:open+topic:bp/nfs-ganesha

I want to get them in so they'll show in downstream puddles for QE but
my full attention will immediately turn to upstream TripleO CI and doc for
this
new functionality.  In that regard I *think* we'll need Dan Sneddon's work
here:

https://review.openstack.org/#/c/523638

so that actual deployment of the StorageNFS network doesn't have to
involve copying and editing

network/config/*/{ceph,compute,controller}/.yaml

as done in the DNM patch that I've used for testing actual integration of
the
feature here:

https://review.openstack.org/533767

All said, this one seems to be a good poster child for composable roles +
composable networks!

-- Tom Barron


On Tue, Jan 23, 2018 at 2:48 PM, Emilien Macchi <emilien at redhat.com> wrote:

> I agree this would be a great addition but I'm worried about the
> patches which right now don't pass the check pipeline.
> Also I don't see any release notes explaining the changes to our users
> and it's supposed to improve user experience...
>
> Please add release notes, make CI passing and we'll probably grant it for
> FFE.
>
> On Mon, Jan 22, 2018 at 8:34 AM, Giulio Fidente <gfidente at redhat.com>
> wrote:
> > hi,
> >
> > I would like to request an FFE for the integration of nfs_ganesha, which
> > will provide a better user experience to manila users
> >
> > This work was slown down by a few factors:
> >
> > - it depended on the migration of tripleo to the newer Ceph version
> > (luminous), which happened during the queens cycle
> >
> > - it depended on some additional functionalities to be implemented in
> > ceph-ansible which were only recently been made available to tripleo/ci
> >
> > - it proposes the addition of on an additional (and optional) network
> > (storagenfs) so that guests don't need connectivity to the ceph frontend
> > network to be able to use the cephfs shares
> >
> > The submissions are on review and partially testable in CI [1]. If
> accepted,
> > I'd like to reassign the blueprint [2] back to the queens cycle, as it
> was
> > initially.
> >
> > Thanks
> >
> > 1. https://review.openstack.org/#/q/status:open+topic:bp/nfs-ganesha
> > 2. https://blueprints.launchpad.net/tripleo/+spec/nfs-ganesha
> > --
> > Giulio Fidente
> > GPG KEY: 08D733BA
> >
> > ____________________________________________________________
> ______________
> > OpenStack Development Mailing List (not for usage questions)
> > Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:
> unsubscribe
> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>
>
> --
> Emilien Macchi
>
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20180204/6f179d2d/attachment.html>


More information about the OpenStack-dev mailing list