Open Stack

On 08/11/2015 11:39 AM, Gary Kotton wrote:
> On 8/11/15, 6:09 PM, "Jay Pipes" <jaypipes at gmail.com> wrote:
>> Are you saying that *new functionality* was added to the stable/kilo
>> branch of *Neutron*, and because new functionality was added to
>> stable/kilo's Neutron, that stable/kilo *Nova* will no longer work?
>
> Yes. That is exactly what I am saying. The issues is as follows. The NSXv
> manager requires the virtual machines VNIC index to enable the security
> groups to work. Without that a VM will not be able to send and receive
> traffic. In addition to this the NSXv plugin does not have any agents so
> we need to do the metadata plugin changes to ensure metadata support. So
> effectively with the patches: https://review.openstack.org/209372 and
> https://review.openstack.org/209374 the stable/kilo nova code will not
> work with the stable/kilo neutron NSXv plugin.
>
> Now having said that I understand the issues with this maybe breaking the
> stable back port rules.

Not maybe. Definitely. :)

 > So the work around on my behalf was to update the
> wiki - https://wiki.openstack.org/wiki/Neutron/VMware_NSX_plugins - and
> whoever wants or needs this code can take the relevant patches. I guess
> that each distrobution can you their discretion if they want to take this
> or not - without it the plugin will not work.
>
> Now, it is not worthwhile crying over spilt milk for the kilo stuff. That
> boat has sailed. But for Liberty we need -
> https://review.openstack.org/165750. If this does not land in L we are
> going to be having the same thread again about 6 months time. This code is
> blocked with a -2.
>
> So what do you suggest?

Looking at https://review.openstack.org/165750, it looks like although 
you pushed the patch back in March, nobody other than Salvatore looked 
at it until two days ago. I think you probably could have notified folks 
that you needed reviews some time in between March and August.

Personally, I've reviewed a number of VMWare patches over the last 
couple months, and I'm still a little annoyed about the whole "Well, 
NSXv *requires* the (completely and utterly Nova-specific and 
non-ordinal) vNIC index to be specified in the VIF plug() request" 
thing. I just think that the fact that NSXv relies on Nova's view of the 
sequential order of vNICs during plug() operations is an indication that 
the NSXv APIs are totally b0rked-by-design.

But, you asked for my suggestion. And my suggestion would be more prior 
notification of critical dependencies like this.

Best,
-jay