14:00:12 <russellb> #startmeeting nfv 14:00:13 <openstack> Meeting started Wed Jul 30 14:00:12 2014 UTC and is due to finish in 60 minutes. The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:14 <russellb> hello, everyone! 14:00:16 <openstack> The meeting name has been set to 'nfv' 14:00:18 <ian_ott> hi 14:00:24 <yamahata> hi 14:00:28 <cloudon> hi 14:00:31 <pczesno> hi 14:00:32 <russellb> #link https://etherpad.openstack.org/p/nfv-meeting-agenda 14:00:35 <russellb> #chair sgordon 14:00:36 <openstack> Current chairs: russellb sgordon 14:00:37 <yyywu> hi 14:00:37 <davidmck> hello 14:00:48 <russellb> #topic review actions from last week 14:00:49 <bauzas> \o 14:01:02 <russellb> sgordon: since you ran the meeting last week, would you like to cover these? 14:01:09 <sgordon> bauzas, quick update on the dashboard situation plz - we have a lot of other stuff to get to :) 14:01:15 <bauzas> sgordon: 14:01:19 <bauzas> sgordon: sure 14:01:48 <bauzas> sgordon: so, working on providing an HTML file with custom dashboard, about to finish today 14:02:24 <sgordon> k, cool - please update via the list when ready 14:02:26 <bauzas> sgordon: currently running tests, alpha version to be provided without CSS 14:02:32 <bauzas> sgordon: sure, willdo 14:02:34 <sgordon> #action bauzas to update list with link to new dash 14:02:50 <sgordon> sean mooney was going to refresh the dpdk enablement specs which i believe he did 14:02:52 <sgordon> speak of the devil 14:02:54 <sgordon> #link https://review.openstack.org/#/c/95805/2 14:03:00 <sgordon> #link https://review.openstack.org/#/c/107797/1 14:03:17 <sgordon> i did not see any further indication of yay/nay on the requests for spec exceptions though... 14:03:25 <sgordon> #link http://lists.openstack.org/pipermail/openstack-dev/2014-July/040660.html 14:03:30 <sgordon> #link http://lists.openstack.org/pipermail/openstack-dev/2014-July/040877.html 14:03:43 <russellb> getting exceptions at this point is pretty tough 14:03:53 <russellb> the projects have already been overwhelmed with how much was already approved 14:03:58 <russellb> i think it's probably unlikely ... 14:03:59 <sgordon> right 14:04:22 <sgordon> the other action was that yamahata was going to follow up with rkukura to see if he could look at ml2-ovs-portsecurity spec 14:04:31 <adrian-hoban> russellb: Is there an exception process end date? 14:04:32 <sgordon> i think maru was looking at it too but he is at the nova midcycle atm 14:04:41 <sean-k-mooney> hi sorry for being late. am form the last nova meeting they mentioned that an excetion would only be considered if there are two core supporting the spec 14:04:47 <russellb> adrian-hoban: for nova, i believe the end was supposed to be last week 14:04:51 <russellb> not sure about neutron 14:05:36 <sgordon> not sure 14:06:06 <sgordon> neutron had a spec approval deadline but exceptions were approved after that 14:06:10 <sgordon> requirements are the same though 14:06:11 <sgordon> two core 14:06:14 <sgordon> sponsors 14:06:50 <bauzas> russellb: that's unclear when the exception deadline is 14:06:52 <pczesno> the DPDK vif spec has sponsors on neutron side 14:06:52 <sgordon> i think we have failed to make our goals clear enough particularly on the neutron side to the wider community which is why we are struggling here 14:06:57 <bauzas> russellb: for Nova I mean 14:07:09 <bauzas> russellb: as some specs have been excepted end of last week 14:07:20 <russellb> bauzas: right, but i thought the intended end of that was friday 14:07:45 <bauzas> russellb: I guess it's this friday, but anyway, that's not impacting NFV IIUC 14:07:49 <russellb> as much as i want this, i didn't sponsor because i can only realistically review so much ... 14:07:51 <sgordon> right - i think there was potential for some further discussion at the midcycle 14:07:58 <russellb> i'm already signed up to review too much 14:08:57 <sgordon> seguing over to the next topic here 14:09:00 <sgordon> #topic blueprints 14:09:56 <sgordon> #link https://wiki.openstack.org/wiki/Meetings/NFV#Active_Blueprints 14:10:07 <sgordon> i tried to update the table to reflect current state of all of these y/day 14:10:13 <sgordon> #link https://blueprints.launchpad.net/nova/+spec/multiple-if-1-net 14:10:26 <sgordon> #info multiple-if-1-net approved, code up for review 14:10:41 <sgordon> this had been languishing for a week or so with two minor pep8 issues 14:10:52 <sgordon> i corrected those late yesterday and the check jobs now pass it 14:11:32 <sgordon> danpb, since you arrived...you had actually +2'd this code submission previously btw - would appreciate if you could take another look 14:11:33 <sgordon> #link https://review.openstack.org/#/c/98488/ 14:12:36 <sgordon> there are also a number of vlan related specs in the table covering related but slightly different use cases (we probably need to split them out) 14:12:56 <sgordon> they are all unapproved still but i did get some feedback from sponsors for this one: 14:12:59 <sgordon> #link https://blueprints.launchpad.net/neutron/+spec/nfv-vlan-trunks 14:13:17 <sgordon> #link https://review.openstack.org/#/c/97714/ 14:13:36 <sgordon> again though as we get further past the deadlines these become less likely... 14:14:05 <sgordon> last comment from a core on the 24th 14:14:57 <russellb> seems nobody has much to say on these neutron bps :) 14:15:12 <sgordon> #link https://blueprints.launchpad.net/neutron/+spec/ml2-ovs-portsecurity 14:15:18 <sgordon> is in a similar situation to the previous one 14:15:19 <russellb> which is a problem we need to address ... need to get some neutron folks engaged more closely with this group 14:15:47 <sgordon> did get some feedback from cores late last week on an exception but have not heard anything further 14:16:06 <sgordon> (per the spec have a +2 from kyle, maru was still looking at it) 14:16:31 <sgordon> but really the above brings me to the next topic in the etherpad.... 14:16:33 <sgordon> #topic Team goals and structure 14:17:25 <sgordon> i have received a lot of feedback in the last week or so, particularly around the spec deadlines, that the wider development community still dont fully understand what NFV is and why they should care about it 14:17:39 <sgordon> which is i would say a failing of this group (myself included) and something we need to work at 14:18:23 <russellb> i've been scheming on a blog post discussing some of the things that have been approved/implemented 14:18:24 <sgordon> as while we dont necessarily need core contributors in this group directly we need to be able to clearly illustrate what we are trying to achieve so that when they come to our specs they can see where we are coming from 14:18:45 <russellb> to show some concrete examples of work being done 14:18:56 <adrian-hoban> sgordon: With the telco strategies track identified for the next summit, I think some presentations to the wider community on NFV topics should start to address the issue. 14:19:01 <sgordon> so here i think our use cases are with a few exceptions largely too high level and disconnected from the actual proposals 14:19:23 <russellb> one tough thing with the summit is that with the design summit overlapping, most of the dev community misses most of the main summit 14:19:23 <sgordon> adrian-hoban, certainly i think that helps my concern is much of the development community end up missing those as they are hunkered down in session 14:19:40 <sgordon> particularly for the nova and neutron peeps who have design sessions for almost the entirety of the allocated days 14:19:49 <russellb> right 14:19:52 <russellb> that's always been my problem 14:20:02 <russellb> i can count the main sessions i've been to over several summits on one hand 14:20:04 <adrian-hoban> sgordon: russellb: Agreed 14:20:19 <russellb> and it's also the reason lots of devs don't even bother submitting talks ... 14:20:25 <russellb> anyway, not this groups problem to solve 14:20:27 <sgordon> adrian-hoban, but yes i think being able to more clearly illustrate a link between what *has* been achieved and use cases will help 14:20:29 <yyywu> I think explain the user case is important, and also comparing with IaaS model because most people in nova and neutron are familiar with Iaas model well. 14:20:41 <sgordon> adrian-hoban, but also more communication both via the M/L and nova/neutron meeting attendance 14:21:04 <sgordon> i should probably say attendance *and* contribution :) 14:21:52 <adrian-hoban> sgordon: Also a fair point. 14:22:14 <sgordon> i am guilty of falling down in this regard on the neutron side myself 14:22:22 <cloudon> Do you think the devs are at least aware there is a thing called NFV, it's just that they don't particularly understand it/are interested in it? 14:22:27 <sgordon> so not pointing fingers, just highlighting that it sounds like this is still an issue 14:22:50 <sgordon> cloudon, tbh i think the impression is "NFV turned up and is a big thing but i still dont know why and it seems to be doing a lot of uncloudy things" 14:22:59 <russellb> sgordon: yes, that. 14:23:00 <sgordon> that is just based on some of the feedback i am getting 14:23:02 <beagles> cloudon - I think there are general misconceptions about it 14:23:18 <beagles> and the uncertainty leads to hesitation etc.. 14:23:18 <cloudon> sgordon: uncloudy in what ways? too hardware-specific? 14:23:43 <sgordon> cloudon, typically specific proposals exposing too much of the detail to the end use 14:23:59 <sgordon> cloudon, it's a tough balance to strike because to a degree that is what NFV applications want in many cases 14:24:18 <danpb> sgordon: some things to bear in mind are that this is the first cycle we've done the spec process 14:24:25 <sgordon> danpb, true enough 14:24:33 <danpb> sgordon: one thing I think we did wrong is that we didn't even start reviewing specs until Juno1 was already open 14:24:38 <sgordon> danpb, and i think it has been overwhelmingly positive 14:24:40 <sgordon> fwiw 14:24:46 <danpb> so we basically lost an entire dev stage for doing work 14:24:51 <sgordon> danpb, right that goes to something i mentioned on the m/l last week 14:25:00 <bauzas> some feedback came yesterday in the Nova mid-cycle sprint about how to manage specs in the next cycle 14:25:02 <danpb> we need to front-load review for Kilo-1 to be during Juno-3 14:25:02 <sgordon> danpb, even less approved in -1 and -2 this cycle than previous ones 14:25:16 <sgordon> huge +1 on that 14:25:20 <danpb> so people actually have a chance to do useful work in Kilo-1 14:25:25 <bauzas> long story short, there will be probably less specs to be on focus for Kilo 14:25:37 <ian_ott> danpb: agreed 14:25:54 <bauzas> but on the other hand, some work don't require a spec while it has been posted 14:26:11 <adrian-hoban> sgordon: Completely agree, and we need a considerable amount of time to deep dive on BPs at the Kilo design session 14:26:14 <sgordon> #info Broader communication with wider development (nova, neutron, etc.) communities required to illustrate what NFV is and is not, current state and progress 14:26:50 <sgordon> #info Feedback from nova midcycle is that more needs to be done to front load the earlier release milestones late in the previous cycle 14:26:59 <danpb> bauzas: that sounds mad, approving even fewer specs in Kilo would be making the bad situation even worse 14:27:05 <sgordon> right 14:27:09 <ian_ott> do we need a dev only session on NFV at next summit or some other venue/vehicle to help educate wider dev community on NFV? 14:27:11 <bauzas> danpb: I'm just the messenger :) 14:27:22 <sgordon> really it should be putting stuff that has been ironed out or is noncontroversial too the front of the queue 14:27:31 <bauzas> danpb: https://etherpad.openstack.org/p/juno-nova-mid-cycle-meetup 14:27:37 <sgordon> plenty of good candidates abandoned last week that were well on their way 14:27:45 <sgordon> ian_ott, we certainly would intend to do something 14:27:54 <sgordon> ian_ott, design summit sessions not open for proposal yet 14:28:01 <sgordon> ian_ott, ideally attempt to get a cross-project session 14:28:10 <ian_ott> happy to help with that for sure 14:28:14 <sgordon> ian_ott, that said needs specific goals 14:28:14 <cloudon> +1 14:28:29 <sgordon> as one was proposed last time and not actually accepted 14:28:33 <sgordon> (think alank proposed it) 14:28:47 <danpb> sgordon: the design summit sessions were not dealt with well last time 14:29:03 <sgordon> for sure 14:29:04 <danpb> sgordon: people proposed lots of sessions for very narrow topics which inevitably got cut 14:29:23 <sgordon> yes there were several competing proposals for this 14:29:27 <danpb> we need to be co-ordinated on proposing a broad session for outstanding NFV work 14:29:31 <sgordon> none of which had a lot of detail about the goals of the session 14:29:46 <danpb> so that people see the big picture importance 14:30:02 <sgordon> i think for the most part they were cut in favor of specific sessions like sr-iov, libvirt roadmap, etc 14:30:05 <sgordon> right 14:30:20 <ian_ott> sgordon: danpb: agreed, maybe next week have an agenda item and come with topics we would include in a session 14:30:32 <sgordon> #info Cross-project design session proposal for Kilo summit needs to be co-ordinated and have specific goals to be accepted 14:30:33 <adrian-hoban> Should we have one thread in the design sessions where we work through all of the identified gaps, and if possible prefix that thread with an update from the ETSI-NFV community? 14:31:10 <sgordon> adrian-hoban, i think at a higher level we just need to communicate what NFV actually looks like on OpenStack, current state, goals for future state 14:31:25 <sgordon> not sure if ETSI-NFV update will necessarily help with the target audience 14:31:32 <sgordon> though for sure is of interest to this group 14:33:44 <sgordon> #action Bring possible topics/goals for cross-project session to next week's meeting. 14:33:45 <smazziotta> are dev aware of the nfv wiki ? 14:33:57 <sgordon> smazziotta, painfully i think is the feedback 14:34:35 <sgordon> i think we perhaps should move the mission statement up above the who we are though to make that clearer 14:34:54 <smazziotta> because we 've spent some time to document. but if they are not aware of it... it is perhaps the issue and we might want to include more 101 material in the wiki 14:36:22 <sgordon> smazziotta, i think the pov of the average reviewer outside of this group is that the spec should outline what/why to enough detail that they dont need to seek that out 14:36:54 <sgordon> smazziotta, though certainly if we can do a better job of explaining on the wiki that helps too 14:37:17 <sgordon> as i said earlier though i think the bigger issue is active engagement on our part (collectively) within the relevant projects 14:37:51 <sgordon> otherwise NFV is perceived as something being done too openstack rather than something being done organically within the projects 14:37:54 <sgordon> if that makes sense 14:38:06 <sgordon> again i am just going off the feedback i am getting at the moment 14:39:06 <sgordon> i see ijw wandered in which brings me to the other topic i had on the list... 14:39:10 <sgordon> #topic meeting times 14:39:16 <sgordon> i created a poll here: http://whenisgood.net/exzzbi8 14:39:28 <sgordon> as i want to confirm whether or not we have outgrown our meeting slot 14:39:54 <sgordon> most of the responses so far indicate for same time or earlier but i am going to keep it open till end of the week 14:40:00 <sgordon> #link http://whenisgood.net/exzzbi8 14:40:10 <sgordon> #info please vote on future meeting times by EoW 14:40:18 <sgordon> the alternative might be to have alternating meetings 14:40:28 <sgordon> if it seems like there are two distinct groupings of availability 14:40:55 <ulikleber> Whenisgood doesn't say which time zone it is talking. 14:40:55 <adrian-hoban> sgordon: Thanks, we need to make sure meeting times are not getting in the way of collaboration 14:41:15 <sgordon> ulikleber, i believe it let's you select at the top 14:41:40 <sgordon> ulikleber, the default it will appear in is UTC+5 because that is where i am but you can switch it in your response (i highlighted this in the email) 14:42:03 <adrian-hoban> ulikleber: sgordon: I could select the timezone at the top. 14:42:04 <ulikleber> Ah, understand. Sorry. 14:42:23 <sgordon> np it is easy to get confused with these :) 14:42:51 <sgordon> adrian-hoban, and yes i had some feedback from at least some people who would participate but cant so certainly something to evaluate 14:43:00 <sgordon> since this time was just something we settled on as a trial after summit 14:43:13 <sgordon> there is still lots of availability in the -3 meeting room so we have some flexibility 14:44:27 <sgordon> #topic open discussion 14:44:36 <sgordon> does anyone else have something to discuss? 14:44:51 <sgordon> ijw, we rattled through the BPs earlier but might be worth revisiting if you are available 14:44:57 <smazziotta> sorry on the topic of the poll. when is the deadline to decide for a change or not ? 14:45:24 <sgordon> in particular did you hear anything else since we last spoke about exceptions for the vlan trunking or port security proposals 14:45:36 <adrian-hoban> Also, can we discuss if there is a mechanism to help prioritise reviews of BPs that missed the Juno deadlines? 14:45:53 <sgordon> smazziotta, i suspect we may end up meeting same time next week to confirm the change (if there is one) 14:46:01 <smazziotta> ok 14:46:13 <sgordon> adrian-hoban, ack 14:46:22 <smazziotta> let's set a deadlin until next wed. ok? 14:46:24 <sgordon> adrian-hoban, so sounds like that was a topic at the nova midcycle meetup 14:46:35 <sgordon> smazziotta, sure 14:46:51 <sgordon> #action sgordon confirm future meeting times at next week's meeting 14:47:07 <sgordon> adrian-hoban, i guess for discussion on the M/L for those of us who werent there 14:47:26 <sgordon> danpb, has what you referred to earlier about trying to frontload Kilo been discussed on the M/L at all yet? 14:47:29 * sgordon may have missed it 14:49:00 <danpb> sgordon: some nova cores discussed it on irc about a week ago 14:49:30 <sgordon> tentatively definitely maybe then :) 14:49:38 <sgordon> fair enough 14:50:28 <adrian-hoban> sgordon: danpb: Was something agreed or still needs to be discussed on M/L? 14:51:00 <sgordon> adrian-hoban, sounds like it still need to be raised on the M/L 14:51:43 <danpb> adrian-hoban: amongst those discussing it there was general agreement we need to front-load more review approval to allow useful dev work during Kilo-1 14:52:00 <danpb> we didn't write any hard & fast rules around it though 14:52:20 <danpb> there is still going to be a balancing act because we'll have important bug fixing work todo at the same time 14:52:49 <danpb> so Kilo1 will still be lighter than Kilo2/3 in terms but should be better than what we had for juno1 14:53:32 <sgordon> right 14:53:58 <sgordon> i know originally it was suggested specs would be locked/not reviewed at this point 14:54:31 <sgordon> adrian-hoban, do you want to try raise this on the M/L to prod some more discussion? 14:54:49 <adrian-hoban> sgordon: Yes, give me that action :-) 14:55:24 <sgordon> #action adrian-hoban to follow up on M/L regarding potential for pre-loading the early kilo milestones 14:55:26 <sgordon> ok 14:55:35 <sgordon> so we're drawing on time for today 14:55:42 <smazziotta> FYI, currently in the ETSI meeting. we started discussion on liaison. I will be reporting to the team the outcome of the discsussion on ESTI NFV gaps next wednesday. 14:55:52 <sgordon> thanks 14:56:06 <sgordon> #action smazziotta to report on outcome of ETSI NFV gaps discussion at next week's meeting 14:56:14 <sgordon> thanks all for the discussion 14:56:19 <bauzas> thanks 14:56:48 <sgordon> let's look to make sure we are closing those gaps we talked about w.r.t. engagement with the specific projects we are working on 14:57:06 <sgordon> it's an ongoing issue that pre-dates this subteam but we need to continue to work on it 14:57:09 <sgordon> thanks all! 14:57:12 <sgordon> #endmeeting