16:04:18 <rkukura> #startmeeting networking_ml2
16:04:19 <openstack> Meeting started Wed Feb  1 16:04:18 2017 UTC and is due to finish in 60 minutes.  The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:04:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:04:24 <openstack> The meeting name has been set to 'networking_ml2'
16:04:38 <rkukura> dasanind, want to repeat that for the log?
16:04:48 <dasanind> The db patch https://review.openstack.org/#/c/404293/ for the live migration got merged :)
16:04:57 <rkukura> dasanind: nice work!
16:05:06 <dasanind> rkukura: thank you
16:05:35 <Sukhdev> Here is the agenda from last week - https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_January_25.2C_2017
16:05:52 <Sukhdev> one patch merged - the other one is not ready yet
16:05:59 <dasanind> rkukura: and Sukhdev thank you for the help and reviews on the patch
16:06:28 <dasanind> I am working on the extension patch https://review.openstack.org/#/c/414251
16:06:37 <Sukhdev> dasanind : no worries - I will review the other one as well
16:07:26 <dasanind> As I am introducing a new extension I am facing some errors. Debugging them hopefully will have a fix today
16:08:25 <dasanind> After that my next step is to test the api's
16:08:57 <dasanind> Do you all think that I should a start an api-ref patch in parallel so that the documentation does not fall through the cracks?
16:09:26 <rkukura> dasanind: probably makes sense, and may help with review
16:09:39 <rkukura> will there be a 3rd patch that adds this to ML2?
16:10:57 <dasanind> rkukura: I think it's better to separate that in the 3rd patch so that it's easy to review them
16:11:28 <rkukura> dasanind: agreed, just wanted to confirm that a 3rd code patch was planned for that
16:12:30 <Sukhdev> dasanind : I just looked through the patch - it looks good to me
16:12:44 <dasanind> rkukura: but I will need some information on how to add this to ML2. If you can explain that it will be helpful.
16:12:52 <dasanind> Sukhdev: thank you
16:15:11 <rkukura> dasanind: I expect you will need to mix in a base class to the ML2 plugin class, add this extension to its supported extensions list, and make whatever changes are needed to ML2’s port binding implementation, etc.
16:15:49 <dasanind> rkukura: makes sense
16:16:24 <rkukura> I have not been tracking the spec or patches as closely as I’d like, so I don’t have anything more concrete. But I will try to get back up to speed on them so I can provide more useful advice
16:16:44 <dasanind> ok
16:17:04 <rkukura> thanks dasanind!
16:17:05 <rkukura> anything else on these patches today?
16:17:11 <dasanind> But the existing port binding implementation should remain unchanged correct?
16:17:25 <Sukhdev> dasanind : correct
16:17:37 <dasanind> Sukhdev: ok
16:17:52 <rkukura> dasanind: Is that correct? Might be, but I thought some integration would be needed.
16:18:33 <rkukura> I’m not talking about the portbinding extension itself, but about the ML2 plugin’s logic that binds ports.
16:19:13 <rkukura> I thought that would need to change to allow the new binding to be created while the old one still exists, etc.
16:20:32 <dasanind> I have to look into the existing logic in little more detail
16:21:50 <dasanind> but if we change the implementation then will not affect the users who are not using the new apis?
16:24:07 <rkukura> dasanind: you do need to avoid breaking existing functionality
16:24:17 <dasanind> My understanding is the integration should be done in a way that the existing portbinding methodology works in addition to the new methodology.
16:24:43 <dasanind> rkukura: thats what I understood
16:25:14 <Sukhdev> right - we need to keep it backward compatible
16:25:18 <rkukura> There needs to be some code somewhere that calls the ML2 mechanism drivers to bind the port and manages the state
16:25:43 <rkukura> That code needs to support both the old and new APIs for accessing that binding state
16:26:00 <rkukura> hopefully there is a plan that addresses that
16:27:12 <dasanind> rkukura: do you mean there is already existing code that manages the state or we have to add new logic for that
16:28:31 <rkukura> dasanind: There is existing code in the ML2 plugin that binds ports. I’d assume this code needs to be modified, but maybe not.
16:30:04 <rkukura> dasanind: I need to take a closer look at the spec and existing patches - I have not been tracking those as closely as I’d like, so I may be missing something entirely here.
16:30:54 <dasanind> rkukura: sounds good. I will dig in through the code to get a better idea.
16:31:35 <rkukura> dasanind: I hope to be much better prepared for this discussion by next week ;)
16:31:45 <dasanind> rkukura: :)
16:31:56 <Sukhdev> dasanind : the new code as-is looks good, however, the plumbing part to fit it in with the older one needs to be worked out clearly -
16:31:57 <rkukura> anything else on the live migration work?
16:32:06 <rkukura> Sukhdev: exactly
16:32:28 <dasanind> Sukhdev: correct. That part needs to planned properly
16:32:45 <dasanind> That's all for the live migration work
16:32:55 <Sukhdev> dasanind : andreas_s has done some ground work on that - you may want to ping him as well
16:33:14 <dasanind> Sukhdev: sure will do that
16:34:11 <rkukura> OK, is ther anything else to discuss today?
16:34:29 <dasanind> nothing else from my side
16:34:45 <rkukura> Sukhdev? Anyone else?
16:34:53 <Sukhdev> not really
16:35:01 <rkukura> OK, thanks!
16:35:16 <rkukura> #endmeeting