[openstack-dev] [Stable][Nova] VMware NSXv Support

Gary Kotton gkotton at vmware.com
Tue Aug 11 13:49:37 UTC 2015


Hi,
This is a stable issue - please see https://review.openstack.org/165750.
This is required to support the plugin developed in Liberty. It is a bug
too. This is now blocked in L and will also be 50/50 from landing in M. So
will be in the same position again. As a result of the mail sent yesterday
this was -2’ed.
So please how can anyone actually add anything to Nova?
There is the option of taking code and adding it out of tree to get
features in like this, for example
https://github.com/openstack/networking-vsphere/tree/master/networking_vsph
ere/nova/virt/vmwareapi, but that is not healthy and prone to breakage.
But this is what people have chosen as adding anything driver specific in
Nova is likely to be blocked due to process.
Thanks
Gary


On 8/10/15, 6:45 PM, "Kuvaja, Erno" <kuvaja at hp.com> wrote:

>> -----Original Message-----
>> From: Gary Kotton [mailto:gkotton at vmware.com]
>> Sent: Monday, August 10, 2015 4:18 PM
>> To: openstack-dev at lists.openstack.org
>> Subject: Re: [openstack-dev] [Stable][Nova] VMware NSXv Support
>> 
>> 
>> 
>> On 8/10/15, 6:05 PM, "Gary Kotton" <gkotton at vmware.com> wrote:
>> 
>> >
>> >
>> >On 8/10/15, 6:03 PM, "Gary Kotton" <gkotton at vmware.com> wrote:
>> >
>> >>
>> >>
>> >>On 8/10/15, 5:46 PM, "Matt Riedemann" <mriedem at linux.vnet.ibm.com>
>> wrote:
>> >>
>> >>>
>> >>>
>> >>>On 8/10/2015 9:17 AM, Gary Kotton wrote:
>> >>>> Hi,
>> >>>> I am not really sure what to say here. The code was in review for
>> >>>>over
>> >>>>8
>> >>>> months. On a side note but related - we have a patch for a plugin
>> >>>>developed in Liberty - https://review.openstack.org/#/c/165750/.
>> >>>>This has  been in review since March. I really hope that that lands
>> >>>>in Liberty.
>> >>>>If
>> >>>> not we will go through the same thing again.
>> >>>> Working in Nova on code that is self contained within a driver is
>> >>>>difficult - terribly difficult. Not only is this demotivating, it
>> >>>>also  effectively does not help any of the drivers actually add any
>> >>>>features.
>> >>>> A sad day for OpenStack.
>> >>>> Thanks
>> >>>> Gary
>> >>>>
>> >>>> On 8/5/15, 4:01 PM, "Ihar Hrachyshka" <ihrachys at redhat.com> wrote:
>> >>>>
>> >>>>> -----BEGIN PGP SIGNED MESSAGE-----
>> >>>>> Hash: SHA256
>> >>>>>
>> >>>>> Hi,
>> >>>>>
>> >>>>> I think Erno made a valid point here. If that would touch only
>> >>>>>vmware  code, that could be an option to consider. But it looks
>> >>>>>like both  patches are very invasive, and they are not just
>> >>>>>enabling features  that are already in the tree, but introduce new
>> >>>>>stuff that is not even  tested for long in master.
>> >>>>>
>> >>>>> I guess we'll need to wait for those till Liberty. Unless
>> >>>>>nova-core-maint has a different opinion and good arguments to
>> >>>>>approach  the merge.
>> >>>>>
>> >>>>> Ihar
>> >>>>>
>> >>>>> On 08/05/2015 12:37 PM, Kuvaja, Erno wrote:
>> >>>>>> Hi Gary,
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> While I do understand the interest to get this functionality
>> >>>>>> included, I really fail to see how it would comply with the
>> >>>>>> Stable Branch Policy:
>> >>>>>> https://wiki.openstack.org/wiki/StableBranch#Stable_branch_policy
>> >>>>>>
>> >>>>>> Obviously the last say is on stable-maint-core, but normally new
>> >>>>>> features are really no-no to stable branches.
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> My concerns are more on the metadata side of your changes.
>> >>>>>>
>> >>>>>> Even the refactoring is fairly clean it is major part of the
>> >>>>>> metadata handler.
>> >>>>>>
>> >>>>>> It also changes the API (In the case of X-Metadata-Provider being
>> >>>>>> present) which tends to be sacred on stable branches.
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> The changes here does not actually fix any bug but just
>> >>>>>> implements new functionality that missed kilo not even slightly
>>but
>> by months.
>> >>>>>> Thus my -1 for merging these.
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> -          Erno
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> *From:*Gary Kotton [mailto:gkotton at vmware.com] *Sent:*
>> Wednesday,
>> >>>>>> August 05, 2015 8:03 AM *To:* OpenStack List *Subject:*
>> >>>>>> [openstack-dev] [Stable][Nova] VMware NSXv Support
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> Hi,
>> >>>>>>
>> >>>>>> In the Kilo cycle a Neutron driver was added for supporting the
>> >>>>>> Vmware NSXv plugin. This required patches in Nova to enable the
>> >>>>>> plugin to work with Nova. These patches finally landed yesterday.
>> >>>>>> I have back ported them to stable/kilo as the Neutron driver is
>> >>>>>> unable to work without these in stable/kilo. The patches can be
>> >>>>>> found at:
>> >>>>>>
>> >>>>>> 1. VNIC support - https://review.openstack.org/209372 2. Metadata
>> >>>>>> support - https://review.openstack.org/209374
>> >>>>>>
>> >>>>>> I hope that the stable team can take this into consideration.
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>> Thanks in advance
>> >>>>>>
>> >>>>>> Gary
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>>
>> >>>>>>____________________________________________________
>> ______________
>> >>>>>>___
>> >>>>>>_
>> >>>>> ____
>> >>>>>>
>> >>>>>>
>> >>>>> 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
>> >>>>>>
>> >>>>> -----BEGIN PGP SIGNATURE-----
>> >>>>> Version: GnuPG v2
>> >>>>>
>> >>>>>
>> iQEcBAEBCAAGBQJVwgkjAAoJEC5aWaUY1u57NacIALsJ8oo6eJKqJIidBSFzwxv
>> g
>> >>>>>
>> zqJXHE56Lpg62/afRF94B2edfhm791Mz42LTFn0BHHRjV51TQX4k/Jf3Wr22CEv
>> m
>> >>>>>
>> zFZkU5eVMVOSL3GGnOZqSv/T06gBWmlMVodmSKQjGxrIL1s8G1m4aTwe6P
>> qs+lie
>> >>>>>
>> N+cT0pZbcjL/P1wYTac6XMpF226gO1owUjhE4oj9VZzx7kEqNsv22SIzVN2fQcc
>> o
>> >>>>>
>> YLs/LEcabMhuuV4Amde3RqUr0BkB+mlIX1TUv5/FTXT/F4ZwzYS/DBH9MaBJ5t
>> 8n
>> >>>>>
>> hgCTJzCeg598+irgOt3VJ3Jn3Unljz6LNzKIM8RnBG0o51fp8vfE/mODQQaUKOg
>> =
>> >>>>> =ZYP8
>> >>>>> -----END PGP SIGNATURE-----
>> >>>>>
>> >>>>>
>> >>>>>_____________________________________________________
>> ______________
>> >>>>>___
>> >>>>>_
>> >>>>>_
>> >>>>>__
>> >>>>> 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
>> >>>>
>> >>>>
>> >>>>
>> >>>>______________________________________________________
>> ______________
>> >>>>___
>> >>>>_
>> >>>>_
>> >>>>_
>> >>>> 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
>> >>>>
>> >>>
>> >>>https://review.openstack.org/#/c/165750/ is a feature add but it's
>> >>>not targeted against a blueprint, so it's just running as a random
>> >>>thing outside any tracking mechanism for features (launchpad).
>> >>>
>> >>>Salvatore made some comments back in March but otherwise no one
>> from
>> >>>the VMware development team has even commented on this.  As I've
>> said
>> >>>in some other VMware patches in Nova lately, I expect the VMware
>> >>>sub-team to be doing a better job of reviewing each other's code
>> >>>first since they are supposed to be the subject matter experts here.
>> >>>
>> >>>I know Gary reviews pretty much all of the changes that go into the
>> >>>VMware driver in Nova but I don't see the same reciprocated from
>> >>>other members of that team which I think also slows down development
>> >>>- and it impedes building a trust relationship between nova-core and
>> >>>the sub-team to be self-reviewing.
>> >>
>> >>Matt, I understand that ask. In the past when the team has reviewed
>> >>the patches, Trivial fixes have remained in review for months.
>> >>
>> >>For example: https://review.openstack.org/#/c/186716/.
>> >>
>> >>We are really between a rock and a hard place here. You yourself have
>> >>also stated that sometimes the review from sub-team members are
>> >>ignored as they are not ³trusted².
>> >>
>> >>So please advise.
>> >
>> >Here is another fine example - https://review.openstack.org/#/c/163831/
>> 
>> I agree that the sub team(s) need to review more.
>> 
>> The question is how do the team member feel like they are making
>> progress?
>> That is, do they see patches
>> Land. Do they receive postive feedback from cores that things are good,
>>bad
>> or ugly?
>> 
>> I think that the PTL should assign at least 2 cores to each sub team.
>>Let the
>> team have accountability. Without that there is no way of getting
>>anything
>> done and we are back in the same spot.
>> 
>> Without that we are just doing more of the same.
>> 
>
>Sorry, but is this really a [Stable] issue or would it be appropriate to
>get own thread for this?
>
>- Erno
>
>> >>
>> >>
>> >>>
>> >>>--
>> >>>
>> >>>Thanks,
>> >>>
>> >>>Matt Riedemann
>> >>>
>> >>>
>> >>>_______________________________________________________
>> ______________
>> >>>___
>> >>>_
>> >>>_
>> >>>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
>> >>
>> >
>> >_________________________________________________________
>> ______________
>> >___ 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
>> 
>> __________________________________________________________
>> ________________
>> 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
>
>__________________________________________________________________________
>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



More information about the OpenStack-dev mailing list