19:59:58 <blogan> #startmeeting octavia
19:59:59 <openstack> Meeting started Wed Apr 29 19:59:58 2015 UTC and is due to finish in 60 minutes.  The chair is blogan. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:03 <openstack> The meeting name has been set to 'octavia'
20:00:08 <rm_work> lol
20:00:09 <TrevorV> o/
20:00:12 <rm_work> dat timing
20:00:18 <blogan> 2 seconds off!
20:00:23 <madhu_ak> hi
20:00:27 <johnsom> o/
20:00:27 <ptoohill> 0/
20:00:30 <ajmiller_> o/
20:00:48 <blogan> #topic Announcements
20:00:57 <dougwig> o/
20:01:12 <blogan> midcycle for *aas repos and Octavia etherpad
20:01:19 <blogan> #link https://etherpad.openstack.org/p/LBaaS-FWaaS-VPNaaS_Summer_Midcycle_meetup
20:01:49 <blogan> surprise san antonio hasn't gotten many votes :)
20:01:58 <ptoohill> :(
20:02:48 <blogan> anyone have any idea when this will actually be decided? or at least should be?
20:03:02 <blogan> obviously before the dates
20:03:03 <dougwig> i put making a decision on the date later on the agenda.
20:03:19 <blogan> ah okay that works then
20:03:49 <blogan> we also have the lbaas vancouver planning etherpad
20:03:56 <blogan> #link https://etherpad.openstack.org/p/lbaas-vancouver-planning
20:04:38 <blogan> mestery has actually gotten Octavia a fishbowl session! and we'll have a working group for neutron lbaas v2 to sort through the features/issues that should get priority for Liberty
20:04:49 <dougwig> we have two sessions... lbaas features and octavia.  please put content for both into that etherpad, and we'll split them before the summit.
20:06:00 <blogan> Oh looks like german has a brown back for Octavia as well
20:06:02 <blogan> http://openstack.prov12n.com/vbrownbag-techtalks-at-vancouver-summitthe-schedule/
20:06:12 <TrevorV> That brown back though
20:06:25 <blogan> johnsom: do you know if that is more of a demo?
20:06:42 <johnsom> I think his intent is to do demo
20:07:27 <blogan> #topic Neutron process changes
20:07:30 <blogan> dougwig this is you
20:07:34 <dougwig> another announce, something you all should be aware of:
20:07:38 <dougwig> neutron - Adopt new spec process for Liberty
20:07:38 <dougwig> #link https://review.openstack.org/#/c/177342/
20:08:06 <dougwig> (done)
20:08:07 <blogan> dougwig can you give a quick summary as to what the process is now?
20:08:32 <blogan> at least what it currently looks like it will be since that doc is still evoloving
20:09:07 <dougwig> sure.  there is no spec deadline anymore, nor is there a hard requirement to submit a spec to neutron-specs.  if you want something in any neutron project, file an RFE, at any time.  a neutron driver will approve it for a particular milestone, and you can proceed with a blueprint/deisgn doc in neutron/docs/devref (optional), and code.
20:09:50 <blogan> does the devref doc have to be in the same review as the code?
20:10:33 <dougwig> it has to be with the code, so it can be in the same review as the code for simpler things, likely better as a dependent review for bigger stuff.  that's just my take.  the docs are required; sorting out how much and when is something to chat with the cores reviewing your stuff.
20:11:09 <dougwig> that spec hasn't gone in yet, so if you have reservations, comment in the above gerrit review.
20:11:42 <blogan> yes as it does have implications for us obviously
20:11:47 <blogan> well us in neutron-lbaas :)
20:11:52 <blogan> which is us
20:11:57 <dougwig> depends on the next topic.  :)
20:12:07 <blogan> nice segue
20:12:17 <blogan> #topic Joining neutron as an openstack project
20:12:26 <blogan> #link http://lists.openstack.org/pipermail/openstack-dev/2015-April/062310.html
20:12:28 <dougwig> #link http://lists.openstack.org/pipermail/openstack-dev/2015-April/062310.html
20:12:28 <dougwig> #link https://review.openstack.org/#/c/175952/
20:12:43 <blogan> that the new one?
20:12:49 <blogan> yep
20:12:50 <dougwig> this goes back to the discussion from last week, about promoting octavia to an openstack/ project.
20:13:17 <dougwig> and we can decide to: 1) stay in stackforge, 2) apply for independent status with the TC, or 3) join neutron.
20:13:30 <dougwig> my sense is that we could do #3 basically today if desired.
20:13:40 <TrevorV> dougwig is 3 the option where we'd be under that huge umbrella thing?
20:13:45 <blogan> i think #3 is the best, but anyone else feel differently?
20:13:46 <dougwig> TrevorV: aye
20:14:08 <ptoohill> are we doing a vote?
20:14:13 <TrevorV> I thought we had desire to make octavia the ref impl for Neutron LBaaS, is that not a thing anymore?
20:14:28 <blogan> i dont think we should do a vote today because I don't think all major stakeholders are here, we should really do an ML thread
20:14:38 <ptoohill> fair enough
20:14:43 <johnsom> I think ref impl is #3
20:14:48 <ajmiller_> +1 for discussing on ML.
20:14:55 <johnsom> +1 for ML
20:15:04 <TrevorV> +1 for +1 for ML
20:15:07 <madhu_ak> ML makes sense
20:15:26 <blogan> okay I think dougwig or I can start that thread
20:15:49 <blogan> anyone have any concerns about this we can talk about now?
20:16:01 <TrevorV> Aside from the ref-impl thing I just mentioned, nope
20:16:20 <blogan> the ref impl thing would still be #3
20:16:31 <TrevorV> But not so "done today" if it is, right?
20:16:35 <johnsom> I think we should have the basics working before we merge, but otherwise no additional comments
20:16:41 <blogan> really teh only thing to change is we'd have mesteyr as our PTL and we get to say we're openstack
20:17:06 <TrevorV> Alright, dead horse and all that, moving on yeh?
20:17:06 <blogan> johnsom: actually that was discussed in the thread and as long as we meet the requirements, it doesn't matter if it is working or not
20:17:30 <johnsom> Yeah, I remember that now
20:17:52 <blogan> but yeah i suppose all of this can be talked about on the ML
20:18:18 <blogan> ill come back to the next topic bc dougwig had to go afk for a second
20:18:23 <blogan> #topic Breif progress reports
20:18:35 <TrevorV> ssh_driver and PUT updates could use reviews
20:18:38 <TrevorV> I'll link shortly
20:18:52 <TrevorV> #link https://review.openstack.org/#/c/160964
20:19:13 <TrevorV> #link https://review.openstack.org/#/c/170989/
20:19:16 <johnsom> A fix for the keystone domain issue needs a +2/A
20:19:16 <blogan> i think the ssh driver is probably in a decent enough state to merge and then put up individual bug fix reviews, but ill go over it one more final time
20:19:20 <johnsom> #link https://review.openstack.org/#/c/178781/
20:19:39 <blogan> yeah that can be done fairly quickly
20:19:47 <dougwig> sorry, back
20:20:03 <blogan> dougwig: we'll circle back to your topics in a sec
20:20:07 <dougwig> np
20:20:08 <johnsom> I have been testing with the ssh driver.  The bugs I have come across so far are fixed.  I just want to get through a full create/update/delete cycle to confirm
20:20:41 <ajmiller_> The devstack plugin is basically ready to go, but am holding off on pushing for merge until the backend up and running more completely.
20:20:44 <johnsom> Controller worker is coming along.  Thank you for the reviews.  I have tried to incorporate those.
20:20:47 <ajmiller_> #link https://review.openstack.org/#/c/167796/
20:20:57 <ptoohill> Ive made a few updates to devstack plugin and other reviews. In process of testing drivers and updating the scriplt blogan had started on. Also, I plan to have a tempest test working soon.
20:21:23 <blogan> johnsom: well i'd like to get the ssh driver merged before the controller worker, and i'd like to get teh controller worker merged soon bugs and all so we can parallelize bug fixe reviews
20:21:24 <ptoohill> I need to make another patch there since were using api.py Al
20:21:27 <johnsom> I am doing testing by driving the controller worker directly.  I am working on an issue that impacts load balancer creation now.
20:22:00 <johnsom> blogan Agreed.  I put that on the agenda.
20:22:21 <blogan> johnsom: ah okay
20:22:27 <TrevorV> ajmiller_ ptoohill are you guys only testing devstack with the REST api?  Just curious
20:22:45 <blogan> ptoohill you were running into global-requirements issues with the devstack plugin right?
20:22:53 <TrevorV> By that I mean setup of octavia including the REST api not the ssh_driver
20:22:59 <ptoohill> its not because of the plugin blogan.
20:23:13 <ptoohill> and no, im testing the ssh driver if thats what youre talking about TrevorV
20:23:15 <blogan> ptoohill: no but it affects the plugin right?
20:23:19 <TrevorV> Oh okay, just checking
20:23:37 <ptoohill> Whatever is adding the ipaddr/that other one is causing the problems
20:23:51 <blogan> ptoohill: okay
20:24:08 <ptoohill> ill have to look which reviews those are in. I was trying to test everything and did a bunch of cherry-picking so im not sure which did which
20:24:19 <blogan> #topic Discuss merging the controller worker
20:24:31 <ptoohill> but removing those does resolve the req sync
20:24:37 <johnsom> So, when do we want to merge the beast?
20:25:04 <blogan> johnsom: have you figured out how to reload the variables passed in the store?
20:25:15 <johnsom> I have been through a create down to health monitors.  Some DB work is required, but it does complete in the end.
20:25:38 <johnsom> blogan I'm testing that today.  Plan A didn't work, trying out plan b....
20:25:40 <blogan> johnsom: bc i can't figure it out, and talkign to the taskflow people they say there's a hacky way to do it but there's no correct way as far as they coudl tell
20:26:42 <johnsom> Yeah, I think hacky way is the only way at the moment.
20:26:59 <blogan> im still down to merge it to be able to parallelize work on it since it is such a behemoth and what you have right now has set it all up well, just need to polish it up
20:27:28 <johnsom> I have added new flow elements that re-query and should overload the elements in the flow.  Worst case, I go back and switch to passing ids around.
20:27:47 <blogan> yeah all those db calls, but we may ahve to do it
20:27:48 <johnsom> I agree.
20:27:56 <ptoohill> i had a comment regarding the amp_network in this review that needs addressing at some point johnsom
20:28:33 <johnsom> ptoohill Remind me, which one?
20:29:00 <ptoohill> I made it lastnight, about it still being a list or not and the adjustment needed if not
20:29:08 <ptoohill> https://review.openstack.org/#/c/151496/36/octavia/common/config.py
20:29:36 <blogan> oh the config option
20:29:38 <johnsom> Ah, yeah, it should be a StrOpt
20:29:50 <ptoohill> Good deal ;)
20:29:54 <johnsom> We just switched it away from a list based on comments
20:30:09 <blogan> johnsom: think we can get this merged by tomorrow? :)
20:30:15 <blogan> bah by end of the week
20:30:18 <blogan> still need to merge ssh driver
20:30:24 <ptoohill> yea, just wanted to make sure i was on the right page since ih ave a change thats dependent on it
20:30:25 <blogan> though if we coudl do both...
20:30:47 <johnsom> We just need to pick a time, I will push in my latest changes and we can go.
20:31:00 <johnsom> So, this afternoon, tomorrow, etc. I'm open
20:31:02 <blogan> johnsom: okay lets coordinate after meeting
20:31:20 <blogan> #topic Finalize meetup dates
20:31:29 <blogan> for the midcycle
20:31:52 <ptoohill> NEVER, for me #supersadpandaprobably
20:32:03 <blogan> fyi we (Rax people) dont know yet whether we will be able to go
20:32:22 <johnsom> 7/15-7/17 is the only window that works for me.  It also has the most votes
20:32:48 <dougwig> based on the etherpad, can we just decide on that window and plan from there?
20:32:51 <blogan> its actaully the only one with any votes, all hte others have -1's
20:33:02 <blogan> dougwig: +1
20:33:07 <ajmiller_> +1
20:33:27 <jwarendt> +1
20:33:32 <blogan> deleting in 3
20:33:34 <blogan> 2
20:33:37 <blogan> 1
20:33:43 <blogan> beat me to it
20:33:45 <blogan> must have been dougwig
20:33:47 <madhu_ak> nice
20:33:51 <ajmiller_> I need to sign off, take care of some family stuff..
20:33:51 <dougwig> done.
20:34:02 <blogan> ajmiller_: take care man!
20:34:02 <dougwig> ok, next week we can joust over location
20:34:13 <blogan> im the only one to vote for boise
20:34:15 <johnsom> Maui
20:34:23 <dougwig> johnsom: +1
20:35:04 <dougwig> (i think we can move on)
20:36:05 <blogan> dougwig: anything you wanted to add to the discussion about moving octavia under neutron?
20:36:14 <blogan> #topic Dougwig's Playland
20:36:40 <dougwig> no, i'm fine with where it was left.  i like #3 as well.  can we have a gerrit vote or something?  the ML is as lossy as these meetings.
20:36:49 <ptoohill> is  it open discussion? I'm curious what is needed to get the new reqs into global-reqs so it doesnt fail anymore. Not sure how others arent running into this. Am I missing something?
20:36:51 <dougwig> i'm fine with posting either one.
20:37:07 <johnsom> Let's have a ML discussion, then gerrit vote.
20:37:13 <dougwig> ptoohill: pull down openstack/requirements and submit a patch.
20:37:21 <ptoohill> ah
20:37:24 <dougwig> johnsom: ok, let's move it there.
20:37:30 <ptoohill> Thank you dougwig
20:37:43 <johnsom> We have hit the actdiag issue with openstack/requirements too
20:38:10 <blogan> i think the two entries are requirements for pecan, which is surprise pecan doesn't pull them down
20:38:15 <blogan> surprising
20:38:18 <ptoohill> I'm not alone, though I only seen the ipaddr and the pecan one
20:38:28 <blogan> singledispatch
20:38:31 <blogan> and ipaddr
20:38:36 <ptoohill> Yea, those two
20:38:53 <blogan> anything else?
20:38:57 <blogan> from anyone?
20:38:59 <ptoohill> actdiag?
20:39:04 <blogan> sphinx
20:39:06 <ptoohill> ah
20:39:11 <johnsom> The doc-requirements has actdiag which isn't upstream
20:39:20 <ptoohill> Guess i dont have that review picked
20:39:37 <dougwig> doc-requirements exists as my way of getting around the proposal bot, fyi.
20:40:00 <blogan> i think its a good idea
20:40:01 <dougwig> anything in there is *not* checked, but tox does pull it in.
20:40:02 <ptoohill> oh, i see
20:40:18 * blogan watches for a hole to open up underneath for saying dougwig had a good idea
20:40:27 <johnsom> We have a team that does builds and uses a local pypi that only load openstack/requirements.  May not be an issue outside our strange team.
20:40:49 <blogan> strange is normal
20:40:55 <ptoohill> :)
20:40:57 <dougwig> i didn't mean for anything that the egg requires to go in that file.
20:42:19 <blogan> alright i think its time to end
20:42:25 <blogan> have a good day!
20:42:28 <blogan> #endmeeting