[openstack-dev] [nova][vmware] Future of Vim and PBM in Nova: summary of IRC discussion

Gary Kotton gkotton at vmware.com
Fri Jun 20 11:21:17 UTC 2014


Hi,
Thanks for the updated mail. I have a number of comments:
1. I agree with the proposed changes regarding the SPBM. I think that the
proposal and changes put forward by Radoslav are great.
2. It would be nice to see the integration of the oslo.vmware code. This
is pending the spawn rewrite, and a number of other patches in review. So
hopefully with a little help and eyes on those reviews we will be able to
move forwards with this.
3. Regarding the VIM rewrite proposal. I am not 100% sure what you want to
achieve here. I think that it would actually rock the boat and cause a lot
of instability. Over the course of the last 3 cycles we have worked very
hard to stabilize the Vmware driver. The initial idea was to do a forklift
to oslo and then address issues there (supporting the existing API¹s -
currently used by Glance and Ceilometer, and a WIP Nova patch). If the new
proposal can be be implemented under the existing API¹s then great. If not
I think that it will be a wrong direction at the moment. Sadly we are in a
mode where we are all stalled on the spawn rewrites. I think that once we
have upgraded to oslo then we can open the discussion again. In addition
to this there is also the idea of upgrading to pyvmomi in the future,
which may also require a rewrite of the code in oslo. My two cents here,
lets wait. Regarding the proposed spec it would be great if you could
actually add in the proposed API¹s and their comparison to the existing
ones. That would give the community a clear picture on what is required.
Yeah, things can be improved, but lets at least have a concrete plan on
how to do that and not go on a wild journey and have to revert everything
a few months down the line.
Thanks
Gary

On 6/20/14, 1:47 PM, "Matthew Booth" <mbooth at redhat.com> wrote:

>For anybody who missed it, we discussed the following 2 outstanding
>reviews yesterday:
>
>vmwareapi oslo.vmware library integration
>https://review.openstack.org/#/c/70175/
>
>VMware: initial support for SPBM
>https://review.openstack.org/#/c/66666/
>
>The issue is that oslo.vmware already contains nascent support for SPBM,
>so these are really 2 patches trying to achieve the same thing by
>different, incompatible means.
>
>After some discussion, we agreed that we would abandon the Nova SPBM
>patch to concentrate on the oslo.vmware patch. This patch has
>languished, but Vui is going to bring it up to date. It also has an
>approved BP.
>
>We also agreed that we only want 1 refactor review queue. As the
>oslo.vmware patch touches so much code, it inevitably conflicts with the
>spawn refactor. Therefore we will either rebase oslo.vmware on to the
>spawn refactor, or vice versa. As both patch sets are primarily on Vui,
>he will make the call about which is least disruptive.
>
>Radoslav has identified some non-disruptive cleanup work which he
>originally put into the Nova SPBM patch. He will now move this into
>oslo.vmware. This cleanup will be 100% backwards compatible, requiring
>no client updates whatsoever to continue working.
>
>We also need to add some additional features to SPBM support in
>oslo.vmware. These will be added, ensuring they don't impact any
>existing Vim users, and the oslo.vmware version bumped.
>
>I am continuing to advocate a significant rewrite of the Vim API.
>However, as we aren't proposing any backwards incompatible changes at
>the moment there is no current incentive to bring this forward.
>
>Matt
>-- 
>Matthew Booth
>Red Hat Engineering, Virtualisation Team
>
>Phone: +442070094448 (UK)
>GPG ID:  D33C3490
>GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490
>
>_______________________________________________
>OpenStack-dev mailing list
>OpenStack-dev at lists.openstack.org
>http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev




More information about the OpenStack-dev mailing list