17:04:43 <kiall> #startmeeting designate
17:04:44 <openstack> Meeting started Wed Sep 18 17:04:43 2013 UTC and is due to finish in 60 minutes.  The chair is kiall. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:04:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:04:48 <openstack> The meeting name has been set to 'designate'
17:05:01 <kiall> Agenda: https://wiki.openstack.org/wiki/Meetings/Designate
17:05:18 <kiall> Apologies for starting late, a few of us got held up in another meeting..
17:05:27 <kiall> Who's about?
17:05:29 <eankutse> here
17:05:34 <vinodmr> vinod here
17:05:38 <msisk> here
17:05:42 <CaptTofu> here
17:05:43 <tsimmons> here
17:05:46 <mugsie> here
17:05:51 <simonmcc> here
17:05:56 <kiall> Okay :)
17:05:59 <betsy> here
17:06:06 <kiall> #topic Review action items from last week
17:06:10 <kiall> (as promised ;))
17:06:34 <kiall> Let's keep this sort each week "done/doing/blocked/etc"
17:07:00 <kiall> 1. BIND9 Docs (Kiall) - I've thrown up some basic notes, but we defiantly still need more.
17:07:23 <kiall> 2. MySQL BIND9 Docs (CaptTofu) - I assuming you Seattle trip has blocked this?
17:07:27 <kiall> Assume*
17:07:39 <CaptTofu> kiall yes, I need to carry that over
17:07:51 <CaptTofu> I have more time latter half of this week now to do it.
17:08:07 <kiall> 3. Pools/Agent Replacement (mugsie)
17:08:33 <mugsie> I have a working solution for pools about to by sent to gerrit
17:08:38 <mugsie> to be*
17:09:10 <mugsie> i ahve had to implent serves as a nested resource of pools, to allow for PowerDNS NS records
17:09:14 <betsy> code or blueprint?
17:09:19 <mugsie> code
17:09:44 <betsy> mugsie: You still planning on doing a blueprint too?
17:09:48 <kiall> Sorry - Yes that was BP for Pools/Agent Replacement (mugsie) copy and paste fail :(
17:10:02 <betsy> :)
17:10:17 <mugsie> yes, for the async, and how it will interact with pools...
17:10:37 <mugsie> should be easy enough to write out, now that I hjave worked out all the issues
17:10:58 <mugsie> (long story short, async is hard)
17:11:00 <kiall> Okay - Let's get that up in the next couple of days?
17:11:06 <mugsie> yup
17:11:29 <eankutse> async is hard; I agree :-)
17:11:41 <kiall> 4. CLI Docs (simonmcc) (Myself and simon keep passing this to each other)
17:11:50 <simonmcc> CLI Docs - no progress, sorry.
17:12:22 <simonmcc> I WILL get to it this week
17:12:22 <kiall> 5. Python Binding Docs (kiall) - Again, No progress.. Time is hard to find!
17:13:07 <kiall> (If anything, these reminders will make sure they get done eventually.. Even if it's not always on-schedule!)
17:13:19 <simonmcc> +1
17:13:33 <kiall> #topic API v2.0 Progress
17:14:20 <kiall> The RecordSets API is pretty functional at this point, another day or two on it and we should have the initial stab merged
17:15:09 <kiall> I see some of the RAX folks have left some comments in the RecordSet review, anyone have any questions/comments that they thought an interactive conversation would be better? :)
17:16:11 <vinodmr> kiall: I notice you mentioned that you would make some changes to v1 too for the recordset to work.  Is that some thing you plan to do over the next day or 2?
17:16:57 <kiall> So, the V1 API, as exposed to end users, won't be changing.. But, the V1 API code will need to change to account for the RecordSet changes in the central
17:17:14 <vinodmr> I meant the API code
17:17:32 <kiall> In that review right now, the V1 records API is totally and utterly broken :)
17:17:38 <kiall> As are the notification handlers/sink service
17:18:16 <kiall> Both of these will be fixed up, then tested to ensure they behave the same as pre-recordsets once I get another day or two to finish the review
17:19:29 <kiall> So, yes, the V1 API code will change in order to map from the V1 style to the V2 style, allowing us to keep both API versions around for long enough to wean users off them :)
17:20:02 <kiall> Make sense? :)
17:20:03 <vinodmr> Also just curious are there any other tests that you run for this code apart from the ones that you added in the code review?
17:21:00 <kiall> Yea, at HP we have some JMeter test plans that we (ab)use to exercise the API, and our QA team has a whole pile of stuff
17:21:09 <vinodmr> That makes sense - also I think it is easier this way to have the same code at the backend for both v1 and v2
17:22:28 <kiall> yea, having the adaptions at the edge should keep things much cleaner.. and reduce as much duplication as possible
17:23:21 <kiall> re other tests, the other parts we have are fairly tied into HP stuff, and would be difficult the make generic enough to open up..
17:23:41 <betsy> Makes sense
17:24:07 <vinodmr> Just one other question - I am assuming that mugsie's change would cover the statuses more in depth rather than your change - is that correct?
17:24:43 <kiall> vinodmr: yea, there are a couple of things in there that were dropped in simply so I could continue :)
17:25:08 <kiall> My personal favourite being L34  of https://review.openstack.org/#/c/46094/10/designate/api/v2/views/recordsets.py
17:25:57 <tsimmons> Haha nice.
17:26:09 <betsy> :)
17:26:40 <kiall> Okay - So, as we discussed last week, the async work has ended up blocked on the (bare minimum) pools idea, and pools fits under the V2 API topic :)
17:27:25 <mugsie> yeah, so the code i have will implement the v2 api for both pools and servers
17:27:56 <mugsie> currently we will resitricted to single pools, as the backend is still very barebones
17:28:00 <mugsie> will be*
17:28:42 <mugsie> this is to get the pools concept into designate, so we can expand on it for the async stuff, and multiple pools
17:28:56 <betsy> mugsie: that makes sense
17:29:19 <mugsie> I want to have a WIP change on gerrit soon, so I can get comments from you guys
17:29:39 <kiall> It's going to conflict with eankutse1's "Update domains when servers are created, modified or deleted" review https://review.openstack.org/#/c/45078/
17:29:45 <eankutse1> mugsie: good.
17:29:55 <tsimmons> I'm excited to see that.
17:30:02 <kiall> So - Ideally, I'd like to see eankutse1's land first, as it's a much simpler change
17:30:02 <eankutse1> also the blueprint will help in folliowing the WIP code
17:30:14 <mugsie> eankutse1: yup
17:30:52 <eankutse1> cool :-)
17:31:08 <vinodmr> Will it be pools in the backend for v1 too?
17:31:30 <mugsie> vinodmr: no, we will have a config item to create a "default" pool, that v1 will use
17:31:40 <mugsie> s/create/define/
17:31:56 <mugsie> and v2 will have full functionality
17:32:43 <kiall> Yea, ideally we don't implement "new stuff" in the V1 API.. Ideally, V1 goes away at some point in the future, and maintaining feature parity reduces the incentive for people to upgrade
17:33:03 <kiall> All the while making more work for us :)
17:33:34 <vinodmr> No I meant will the create server api call in v1 go through the same code as in v2 or would it be totally separate piece of code
17:33:56 <mugsie> same piece of code, just with the pool_id set to the default set in the config
17:34:08 <kiall> vinodmr: ah, the v1 create server API was admin only.. So I'm of the opinion that it's ditched. But I don't think it's been discussed
17:35:30 <betsy> kiall: so are you saying there's no create server API call in v2?
17:35:32 <kiall> If it's simple - It's no harm to map
17:35:55 <kiall> betsy: there is (via pools)
17:36:03 <betsy> ah
17:36:04 <mugsie> betsy: it will be /pools/<id>/servers/
17:36:19 <mugsie> like record sets within zones
17:36:22 <betsy> pools instead of servers, right
17:36:32 <kiall> In the V2 API we go from "servers" being an admin only resource, to "pools" being a public resource (at least, in terms of the API spec)
17:37:06 <kiall> pools will eventually allow customers to create pools, which might trigger some nova instances to be booted and configured as private DNS servers etc
17:37:29 <kiall> While the V1 servers API had no use at all for anyone but the cloud operator
17:39:07 <kiall> I think we've managed to stray into the next topic "Async State Transitions (blocked on framework for Pool functionality) (mugsie)", which was basically indented to be ^ conversion :)
17:39:25 <kiall> Oh well :)
17:40:23 <kiall> Any more on the V2 API/Pools/etc?
17:40:57 <kiall> <rant>It's impossible on IRC to tell if someone is typing.. or if everyone is waiting for another message</rant>
17:41:30 <kiall> #topic Open discussion
17:41:57 <kiall> Okay .. I'll start this one with - both the RAX reviews up on Gerrit look good :)
17:42:10 <tsimmons> :D
17:42:16 <vinodmr> I see some changes by DirkMueller
17:42:36 <msisk> Anyone from HP going to http://openstack.onales.com to hear me blunder though a talk on Designate?
17:42:38 <vinodmr> So is Designate on the radar for all openstack related changes
17:42:41 <kiall> Myself and eankutse1 had a conversation last night about 1 change to his one - but it's pretty good
17:42:56 <eankutse1> Cool
17:43:35 <kiall> Anyway - tsimmons for some reason Jenkins has ignored your review ;)
17:43:51 <eankutse1> kiall: you want to get rid of the "old_server_name" arg?
17:44:06 <kiall> eankutse1: Well, I think there's got to be a better way :)
17:44:19 <vinodmr> kiall:  I had some questions on eankutse's changes that you might know better
17:44:29 <vinodmr> I just posted them
17:44:36 <kiall> I still don't know what it is myself! Other than maybe doing to SELECT from the PowerDNS DB before making the change..
17:44:51 <kiall> allowing us to determine the old_server_name in the backend itself, avoiding the need to pass it around
17:45:00 <kiall> vinodmr: let me have a look
17:46:25 <eankutse1> Kiall: ok will put some work into that
17:46:51 <kiall> vinodmr: let me follow up on those later actually :) Need to poke around and follow them though
17:47:01 <vinodmr> thanks kiall
17:47:19 <kiall> msisk: OpenStack on Ales - I'm afraid I'm not :(
17:47:44 <msisk> kiall: Bummer. But more beer for me. ;)
17:47:52 <kiall> Ireland -> Oregon is .. expensive .. too expensive for a few pints! :)
17:48:22 <vinodmr> Are some of you going to be in Seattle at that time?
17:48:54 <kiall> I think you'll just be missing CaptTofu by a few days :)
17:48:55 <msisk> kiall: We'll have more than a few pints available. :)
17:49:12 <kiall> Someone, us Irish got shafted out of a US trip.. and out US folks are coming to Ireland in Oct :)
17:49:19 <kiall> Somehow*
17:50:13 <kiall> Anyway - The other one I wanted to bring up was, welcome betsy to designate-core, I wanted to make give someone at RAX that.. and betsy emailed me :)
17:50:25 <kiall> s/make//
17:50:31 <betsy> Thank you
17:50:52 <tsimmons> Woohoooooo :)
17:50:54 <mugsie> betsy: congrats
17:50:55 <jmcbride> Thanks Kiall
17:51:12 <betsy> :D
17:51:15 <kiall> When we go incubated.. There will be rules to follow etc etc .. But.. We still make those for now :)
17:51:43 <kiall> And! Anyone have anything else? I'm out.
17:52:16 <vinodmr> Nope
17:52:30 <mugsie> nope
17:52:54 <eankutse1> no
17:53:01 <tsimmons> ls
17:53:08 <betsy> I'm good
17:53:09 <tsimmons> haha. Wrong window. I'm good.
17:53:16 <kiall> Okay - That's that then :)
17:53:30 <kiall> Thanks all :)
17:54:00 <kiall> #endmeeting