17:01:16 <Kiall> #startmeeting designate 17:01:21 <openstack> Meeting started Wed Jun 11 17:01:16 2014 UTC and is due to finish in 60 minutes. The chair is Kiall. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:25 <openstack> The meeting name has been set to 'designate' 17:01:25 <Kiall> Hey Guys - Who's here today? 17:01:29 <richm> here 17:01:30 <betsy> o/ 17:01:31 <rjrjr_> here 17:01:36 <timfreund> hello 17:01:37 <eankutse> here 17:01:44 <Kiall> mugsie is out today, so won't be here 17:02:03 <Kiall> #topic Review action items from last week (and the week before!) 17:02:08 <vinod> o/ 17:02:37 <Kiall> first was mugsie to translate etherpads over to launchpad blueprints - he asked me to push this till next week 17:03:03 <Kiall> first from the week before was "eankutse to get initial answers to jbrattons questions" .. I just saw a draft of this from eankutse 17:03:15 <eankutse> yes. it's here #link http://paste.openstack.org/show/83671/ 17:03:21 <Kiall> #link http://paste.openstack.org/show/83671/ 17:03:26 <eankutse> but will put it in bp 17:03:29 <Kiall> beat me to it :) 17:03:33 <eankutse> :-) 17:04:05 <Kiall> Great - We'll be able to iterate on the answers once it's in there to get a good idea of what we need to allow for in mdns 17:04:12 <eankutse> sure 17:04:24 <Kiall> Next was "eankutse to pick Jbratton's brain on ops side of things for minidns/pools" .. Seems to me the previous item covered it 17:04:37 <eankutse> that;s the intent 17:05:28 <Kiall> Next two were "kiall to write our various initial load scenarios" and "kiall to file BP on exposing the NS (and SOA?) record in the V2 API" - And I've not had time this week :( Every spare minute was getting the devstack gate running right 17:05:50 <Kiall> #action kiall to file BP on exposing NS/SOA records in V2 API 17:06:00 <eankutse> saw all the commits you did :-) 17:06:09 <Kiall> #action kiall to write out mdns initial load scenarios and add to eankutse's BP 17:06:22 <Kiall> (I think your BP and those will be pretty related) 17:06:28 <eankutse> yes they are 17:06:42 <Kiall> Okay - That was all the actions we missed last week 17:06:55 <Kiall> #topic Mini DNS - is 'mdns-designate-mdns-functional' a place holder or explicit functionality? (eankutse) 17:07:06 <Kiall> (what a BP name, I apologize for that ;)) 17:07:16 <eankutse> 1 sec 17:07:24 <Kiall> #link https://blueprints.launchpad.net/designate/+spec/mdns-designate-mdns-functional 17:07:31 <eankutse> yep 17:07:46 <eankutse> so 17:08:04 <eankutse> is this item standing on it's own? 17:08:12 <eankutse> or is it considered done when 17:08:15 <Kiall> So - My idea when I filed this BP was that, this would be where the skeleton designate-mdns service we now have turns from a skeleton to something useful - even if it's not all of mdns. 17:08:20 <eankutse> the items below it are done? 17:08:54 <Kiall> This would be where we actually get more than "SERVFAIL" and get real answers from mdns for simple queries - e.g. SOA lookups 17:09:10 <Kiall> (but not AXFR just yet, or TSIG sign requests) 17:09:20 <eankutse> ok. 17:09:45 <eankutse> so these would be things leading up to AXFR 17:09:48 <Kiall> If that doesn't make sense to anyone but me, we can add/remove BPs etc 17:09:50 <eankutse> and TSIG 17:10:02 <eankutse> ok. 17:10:03 <betsy> kiall: makes sense to me 17:10:07 <Kiall> eankutse: yea - that was my thinking when I split mdns up into a pile of BPs 17:10:27 <betsy> and leads to smaller patch submissions 17:10:33 <betsy> makes reviewing easier 17:10:54 <Kiall> Ideally, once this happens, and mdns can answer a simple SOA query, we can fork and have 3 people working on different things.. 1 getting AXFR working, 1 adding TSIG support, 1 writing the supporting code to send DNS notifies 17:11:12 <Kiall> (fork != github style fork ;)) 17:11:26 <eankutse> cool 17:11:37 <Kiall> Oh .. There is 1 other BP blocking this one 17:11:42 <Kiall> #link https://blueprints.launchpad.net/designate/+spec/mdns-designate-mdns-dnspy-translations 17:12:24 <Kiall> That was intended to be where we would write out the "converter" methods to translate from our objects, over to dnspy's objects 17:12:27 <eankutse> i was thinking this would happen in pieces 17:12:34 <eankutse> as we implement the other functionality 17:12:39 <eankutse> eg NOTIFY, 17:12:41 <eankutse> SOA 17:12:43 <eankutse> AXFR 17:13:01 <eankutse> these would 'build up' this bp 17:13:07 <eankutse> as we go 17:13:24 <Kiall> e.g. translating an "A" designate.objects.Record to dns.rdata.a.A() (or w/e the dnspy class is called) 17:13:32 <vinod> Is the blue print mdns-designate-mdns-dnspy-translations different from https://blueprints.launchpad.net/designate/+spec/mdns-designate-mdns-notify? 17:13:38 <vinod> I would assume so 17:13:50 <vinod> The description for blueprint mdns-designate-mdns-dnspy-translations talks about NOTIFY 17:13:56 <Kiall> vinod: dooh - description is wrong 17:13:58 <Kiall> title* 17:14:10 <Kiall> copy and paste fail while filing all them 17:14:44 <Kiall> refresh please :) 17:14:45 <Kiall> https://blueprints.launchpad.net/designate/+spec/mdns-designate-mdns-dnspy-translations 17:15:08 <Kiall> That was my original intent for that BP - does that change things? 17:15:50 <tsimmons> (They're chatting) 17:16:09 <Kiall> lol - I was wondering ;) 17:16:41 <eankutse> so Kiall: what do you think about the piecemeal approach? 17:16:43 <vinod> I would think that it would be easier to do the translations as we use dnspython objects 17:17:05 <Kiall> vinod / eankutse sure - we can cross that BP out and do it bit by bit 17:17:13 <eankutse> I like that 17:17:36 <Kiall> and .. marked as obsolete. 17:17:46 <vinod> +1 on that 17:18:09 <Kiall> Anyone disagree before we move on? 17:18:42 <rjrjr_> i'm good. 17:18:45 <betsy> kiall: I liked your approach better, but I’m not the one doing the code. 17:18:46 <betsy> :) 17:18:49 <Kiall> lol 17:18:56 <betsy> So, I’m good 17:19:04 <richm> +1 17:19:28 <Kiall> I think getting 1 done before the meat of the mdns-functional done would be a good idea, all is likely not necessary 17:19:41 <Kiall> getting all done is likely not necessary* 17:20:00 <Kiall> Okay .. Moving on so :) 17:20:03 <Kiall> #topic DevStack (kiall) 17:20:20 <Kiall> I think everyone noticed the binge of DevStack patches ;) 17:20:28 <vinod> #action kiall Mark https://blueprints.launchpad.net/designate/+spec/mdns-designate-mdns-dnspy-translations as obsolete 17:20:31 <eankutse> :-) 17:20:32 <betsy> 3 cheers for Kiall! 17:20:51 <Kiall> it's not all cheers - I had to revert some stuff too :( 17:21:02 <vinod> So is all the devstack stuff done now? 17:21:04 <Kiall> https://github.com/stackforge/designate/commit/f205633c6303dd209ece20237ed60fa5043689da 17:21:35 <Kiall> That added some nice functionality which, sadly, broke when ran in the gate. as does https://review.openstack.org/#/c/97628/ 17:22:18 <Kiall> We're going to have to figure out what exactly is breaking so we can get them merged.. I'm frankly stumped on the first one.. and just confused on why unbound is ending up installed, preventing the second one 17:22:50 <Kiall> richm: any chance you were able to identify how/why unbound was ending up installed? 17:24:14 <vinod> Is it being installed by one of the other projects? 17:24:40 <Kiall> vinod: "unbound" is nowhere to be found in devstack, so it shouldn't be 17:25:04 <Kiall> And - When ran locally in vagrant - it's not there... 17:25:12 <Kiall> (even with the same set of services/devstack options etc) 17:25:37 <Kiall> I'm starting to wonder if the infra folks are pre-installing it on all the single use slaves 17:26:00 <Kiall> maybe fungi / clarkb know? 17:26:43 <richm> Kiall: no, I'm not an ubuntu guy - I'll need to install a vm and do some investigating 17:26:44 <fungi> Kiall: yes, we install it because we need to cache local dns queries on job workers 17:27:18 <richm> ok - so looks like, for testing, we'll need to run the designate DNS server on another ip/port 17:27:25 <Kiall> fungi: ooo. Okay - I can chat with you after then.. We might be able to make them co-exist so long as neither bind to *:53 17:27:49 <fungi> Kiall: perhaps, get up with us in #-infra when you're free 17:28:12 <clarkb> Kiall: importantly we do it because the cloud providers DNS servers don't work... 17:28:14 <Kiall> richm: or we pick another IP - 127.*.*.* is all "ours" .. just have to make sure we're doing it in a sane way 17:28:31 <clarkb> Kiall: if you happen to have any sway in how dns is run at one of them ;) 17:28:44 <Kiall> clarkb: lol - recursive dns is not my area ;) 17:28:50 <Kiall> That's a whole other team 17:29:00 <Kiall> thanks fungi / clarkb - at least that solves 1 mystery :) 17:29:32 <Kiall> Okay - Anyway - I just put that on the agenda so people could think about/come up with solutions ;) 17:30:34 <Kiall> Also - You may have noticed I added a Vagrantfile to the contrib/devstack folder which will spin up DevStack with out plugin in either ubuntu or fedora 17:30:54 <Kiall> richm: that would be the quickest + easiest way to get a ubuntu VM :) 17:31:48 <Kiall> Unless there's any Q's on all the Devstack changes / what's needed going forward we'll move on.. 17:31:52 <Kiall> So .. Any Q's? :) 17:32:02 <richm> Kiall: ok - will take a look 17:32:04 <rjrjr_> i'm good. 17:32:11 <eankutse> good work! 17:32:15 <vinod> are there more devstack changes to be done? 17:32:43 <Kiall> vinod: richm has a patch up that's blocked on the unbound thing that extends the testing we do 17:33:19 <Kiall> and we'll want to try and write some tempest tests if we can figure out how to hook into that, as that's going to be expected of us once we incubate 17:33:34 <rjrjr_> did we incubate? 17:33:39 <Kiall> That's next ;) 17:33:40 <vinod> Is this https://review.openstack.org/#/c/98968/ needed for incubation? 17:33:56 <Kiall> No - Nothing is blocking incubation anymore 17:34:03 <Kiall> #topic Incubation progress (kiall) 17:34:07 <Kiall> may as well ;) 17:34:16 <Kiall> #link https://review.openstack.org/#/c/97609/ 17:34:19 <Kiall> ^ the voting .. 17:34:31 <Kiall> 6 Yay's, 0 Nay's 17:34:37 <betsy> It’s looking good 17:34:45 <eankutse> :-0 17:34:48 <eankutse> :-) 17:34:53 <Kiall> There's also 3 others who voted +1 before the updated patchset 17:35:03 <Kiall> markmc asked us to update the mission statemen 17:35:04 <Kiall> t 17:35:05 <vinod> So is somebody expected to do a +2 before we can get incubated? 17:35:36 <Kiall> myself and graham came up with an alternative wording which made people happy - https://review.openstack.org/#/c/97609/3/reference/programs.yaml 17:36:16 <Kiall> "Enable operators to meet their Domain Name System needs in a stable and scalable manner." <- original wording 17:36:16 <eankutse> That looks good! 17:36:26 <Kiall> Any concerns before it's too late to change it again? 17:36:50 <vinod> As long as the TC is happy, I am happy 17:37:01 <Kiall> :) 17:37:02 <tsimmons> in a* technology-agnostic? 17:37:13 <Kiall> tsimmons: NOOOOO 17:37:26 <betsy> tsimmsons: +1 17:37:27 <tsimmons> :O 17:37:36 * Kiall closes his eyes and pretends he didn't hear that 17:38:04 <vinod> eyes or ears? 17:38:07 <tsimmons> http://2.bp.blogspot.com/-UUBBVvfBUcM/T1t8Zi673mI/AAAAAAAAA0c/uEdZ-PA4IpI/s1600/homer-simpson-bush-gif.gif 17:38:09 <Kiall> both 17:38:15 <Kiall> ^ that 17:38:30 <tsimmons> I'm sure it'll be fine. 17:38:47 <Kiall> :) 17:38:54 <tsimmons> If Anne Gentle (technical writer) didn't care I don't think we have to ;) 17:38:58 <Kiall> Okay - Anyway, We'll be back on the agenda for the TC meet next week 17:39:03 <tsimmons> Sweet 17:39:14 <Kiall> If there are no -1's, I believe that's the end of it 17:39:31 <Kiall> (or the start, whichever ;)) 17:40:06 <vinod> So after incubation do we move to openstack from stackforge? 17:40:30 <Kiall> vinod: yes, usually during the next scheduled Gerrit downtime 17:40:40 <Kiall> (you can't rename gerrit projects without downtime) 17:41:04 <betsy> kiall: oh, interesting. I thought it wasn’t until after integration 17:41:33 <Kiall> I don't think that's the case! 17:41:41 <Kiall> https://github.com/openstack/barbican suggests it's not :) 17:41:55 <tsimmons> Fun. I'm sure that won't mess any scripts up :P 17:42:13 <betsy> cool 17:42:16 <Kiall> tsimmons: you have no idea how much pain that's going to cause me ;) 17:43:25 <Kiall> Anyway - once accepted, we'll be expected to start doing things like A) submit our devstack plugin to devstack proper and fix all the issues they find, B) start writing tempest tests, C) start writing a horizon plugin etc 17:43:45 <Kiall> (HP has C covered, we have one we're happy to open source - when we can put it into openstack/horizon) 17:43:57 <tsimmons> Cool. 17:44:31 <Kiall> We'll also have to deal with the fun (aka pain) of the "integrated gate" 17:45:00 <Kiall> Approved patches will take much longer than land, but, they'll still land :) 17:45:22 <Kiall> http://status.openstack.org/zuul/ the center column gives you an idea of the integrated gate queue times 17:45:56 <betsy> I know. I’ve looked at that before and been glad we were in that queue 17:46:01 <Kiall> ;) 17:46:08 <betsy> The price of incubation 17:46:20 <richm> we are cloud developers - we should just horizontally scale out the gate :-) 17:46:32 <Kiall> richm: lol .. that's exactly what happens ;) 17:46:47 <Kiall> there are currently 900 instances acting as test slaves 17:46:53 <richm> egads 17:46:58 <Kiall> ;) 17:47:10 <Kiall> The first graph at the very bottom of the page shows it 17:47:16 <Kiall> #topic Open Discussion 17:47:30 <Kiall> (Since we're basically onto the open discussion topic anyway) 17:47:37 <eankutse> I have a quick follow up 17:47:40 <eankutse> regarding #link http://paste.openstack.org/show/83671/ 17:47:49 <Kiall> Internal Server Error 17:47:49 <Kiall> The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application. 17:47:51 <Kiall> ;) 17:48:07 <Kiall> there we go - 3 refreshes later and it came up 17:48:25 <eankutse> Can I add it to the current bp where it is? 17:48:45 <eankutse> since it will all be moved to designate-spec? 17:48:46 <Kiall> rather than in designate-specs? yea, for existing BPs moving them over seems pointless 17:48:52 <eankutse> yea 17:49:00 <eankutse> ok 17:49:17 <Kiall> Any other topics? 17:50:24 <Kiall> Going once? 17:50:26 <tsimmons> I have a quick question 17:50:52 <tsimmons> That's probably not relevant to everyone :) So it can wait till after, just fyi 17:51:07 <Kiall> lol .. sure? 17:51:28 <tsimmons> Alright, as far as extending the v2 api, should we use a similar method to v1, stevedore etc? 17:52:36 <vinod> I just wanted to check which patches were still waiting for reviews? 17:52:37 <Kiall> Yea, i think the old pattern worked well - using stevedore (well, entry points i.e the stuff behind stevedore) let's you write out of tree API extensions 17:52:49 <tsimmons> I mean you could just write more views/controllers and hack them into the root.py thing, but that seems messy. 17:52:57 <Kiall> I know we have some very HP-specific extensions for the V1 API 17:53:27 <Kiall> tsimmons: yea, hacking them in means forking when you want to add an private extension, and once you have a fork, things go wrong fast :) 17:53:49 <Kiall> (Just ask a certain cloud provider how that went down with Nova ;)) 17:54:07 <tsimmons> Yeah that's no fun. So I guess I'll look at the other projects (none of which seem to use pecan in a similar way) and try and figure out an extension mechanism with pecan. 17:54:27 <Kiall> vinod: mugsie came up with this for review backlogs http://bit.ly/1u7Zvjy 17:54:36 <betsy> very bad idea 17:55:10 <Kiall> vinod: It'll be tailored to you, showing you the patches which you should "probably" look at 17:55:44 <vinod> Thanks kiall and musgsie for the link 17:56:19 <Kiall> tsimmons: yea, the others use WSME on top of Pecan - Which at the time, was simply not ready to be used in any way. 17:56:51 <tsimmons> Kiall: Should we switch to that now? 17:57:26 <Kiall> <PersonalOpinion>It's still not ready for prime time</PersonalOpinion> 17:57:45 <Kiall> But - It is something we should consider at some point.. Just not this cycle. 17:57:56 <tsimmons> Fine by me. I guess I'll just see what I can figure out with Stevedore 17:58:01 <tsimmons> and the current state. 17:58:15 <Kiall> Okay - Anything else before we call it a day? 17:58:22 <eankutse> none 17:58:44 <vinod> nothing else 17:58:58 <Kiall> Okay - Thanks all :) 17:59:00 <Kiall> and .. all yours SlickNik :) 17:59:01 <Kiall> #endmeeting