17:01:35 <Kiall> #startmeeting Designate
17:01:36 <openstack> Meeting started Wed Jan 28 17:01:35 2015 UTC and is due to finish in 60 minutes.  The chair is Kiall. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:01:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:01:40 <openstack> The meeting name has been set to 'designate'
17:01:40 <shivharis> thanks Sukhdev, bye all
17:01:45 <Kiall> Heya - Who's about?
17:01:51 <vinod> o/
17:02:10 <timsim> o/
17:02:10 <Kiall> mugsie just got nabbed by a phonecall, will be along in a minute..
17:02:21 <Kiall> rjrjr_ - about?
17:02:31 <rjrjr_> o/
17:02:49 <Kiall> Okay - Lets get started...
17:02:54 <Kiall> #topic Kilo Release Status (kiall - recurring)
17:03:01 <Kiall> #link https://launchpad.net/designate/+milestone/kilo-2
17:03:17 <Kiall> k2 is Feb 5th.. 7/8 days
17:03:56 <Kiall> Lot's of bugs still open, if you have one, and it's not going to land in k2, can we move them to k3 please?
17:04:10 <Kiall> (I started a few mins ago, but got distracted ;))
17:04:56 <Kiall> Okay - Other than that, we have the PM cache stuff rjrjr_is working on - The review is still marked WIP, is that intentional rjrjr?
17:05:30 <mugsie> o/
17:05:38 <rjrjr_> correct.  unit tests still pass (sqlalchemy) but i am working on getting the correct status out of mdns for create/delete.
17:05:59 <Kiall> rjrjr_: K, will keep an eye out for another patchset on that one :)
17:06:34 <Kiall> Okay - Other than that, k2 is starting to shape up.. So, let's move on!
17:06:44 <Kiall> #topic Pools - Where are we? (kiall - recurring)
17:06:51 <rjrjr_> i'd also like to know if 'force_check' is the right approach.  i posted in the main chat about it.
17:07:05 <Kiall> I think we already have the answer to this - and we can remove the recurring topic at this stage, thoughts?
17:07:21 <mugsie> i think so
17:07:25 <vinod> +1 to remove the recurring topic of pools
17:07:39 <Kiall> K - Consider it gone.. :)
17:07:40 <timsim> Yep
17:08:07 <Kiall> #topic #toipic
17:08:09 <Kiall> gah.
17:08:17 <Kiall> #topic Mid-Cycle Review (kiall)
17:08:19 <Kiall> there we go ;)
17:08:25 <Kiall> #link https://etherpad.openstack.org/p/designate-mid-cycle-jan-2015
17:08:55 <Kiall> So - Just wanted to figure out if we've left anything dangling after the mid-cycle..
17:09:07 <Kiall> i.e. things we need to follow up on etc?
17:09:32 <timsim> Kinda curious where we are with the Pools->DB stuff.
17:09:34 <vinod> IXFR and backup/restore
17:09:50 <Kiall> I know we've got the pools API stuff in progress, I'm hoping to find time to finish my piece of that this week.. But after being away from the office for 2 weeks it's hard ;)
17:10:13 <timsim> Cool
17:10:38 <Kiall> vinod: yes, IXFR and few topics were dropped :) rjrjr_, did you manage to finish that spec? (All good if you didn;t)
17:10:40 <Kiall> didn't*
17:10:48 <rjrjr_> no
17:11:08 <mugsie> batch actions / ns update as well
17:11:15 <Kiall> Okay, whenever you do, drop it on the agenda for the next meet :)
17:11:48 <Kiall> mugsie: Yep, and perf/testing/API microversions :) I was more hoping for stuff that we started, but left unfinished ;)
17:13:03 <Kiall> I'll take that as a now.. From my memory, we have the Pools Config -> DB stuff, and the Agent stuff we landed.
17:13:05 <Kiall> Any others?
17:13:13 <Kiall> The start of the agent stuff we landed*
17:13:30 <timsim> I don't think so.
17:13:38 <mugsie> nothing - the agent is tied up waiting for targets
17:13:45 <mugsie> which is part of the API->DB
17:13:51 <timsim> Yeah, the Pool Manager backend for it.
17:14:00 <mugsie> yeah
17:14:03 <Kiall> KK..
17:14:18 <Kiall> I'll move on so ;)
17:14:24 <Kiall> #topic Next Sprint?
17:14:28 <Kiall> link https://etherpad.openstack.org/p/designate-sprints
17:14:32 <Kiall> #link https://etherpad.openstack.org/p/designate-sprints
17:15:03 <Kiall> I figure we should pick+plan for the next monthly sprint by figuring out the what and when early..
17:15:19 <Kiall> My suggestion is to do a Docs sprint - thoughts?
17:15:27 <mugsie> yup
17:15:36 <timsim> Absolutely.
17:15:42 <mugsie> That was part of what we talked about on the thursday
17:15:49 <mugsie> we need docs, badly
17:16:01 <vinod> +1 on docs
17:16:02 <Kiall> Cool :) I missed that since I was gone at that point ;) I started a etherpad for it, https://etherpad.openstack.org/p/designate-documentation-sprint
17:16:29 <Kiall> Fill in any areas you think would be useful..
17:16:51 <Kiall> The last sprint was Jan 16th from memory - Should we pick a similar-ish date for the next?
17:17:09 <mugsie> yeah, mid month sounds good
17:17:14 <timsim> Sure, why not. After k2 I suppose
17:17:51 <rjrjr_> +1
17:18:36 <Kiall> So, say the 13th then?
17:18:43 <Kiall> Friday the 13th that is ;)
17:19:25 <timsim> Seems good. That's the day before Valentine's Day here (don't know if you guys celebrate the patron saint of overpriced greeting cards, flowers and chocolates)
17:19:49 <timsim> If anyone is going out of town with a significant other they may miss. Probably not a concern though.
17:20:08 <mugsie> timsim: crass consumerism? we have that here 100% ;)
17:20:22 <timsim> Ah, too bad lol.
17:20:27 <Kiall> Okay - Unless I hear shouting from someone who's unable to attend, we'll call it the 13th.. Ideally afternoon Ireland, Morning Texas, but exact times can be figured out closer to the day ;)
17:21:48 <Kiall> Sorry - Distracted for a sec.. Moving on since we're settled :)
17:21:58 <Kiall> #topic Strawman: Set a McCabe complexity threshold
17:22:01 <Kiall> #link https://review.openstack.org/#/c/149885/
17:22:06 <Kiall> #link http://logs.openstack.org/85/149885/1/check/gate-designate-pep8/de167ea/console.html#_2015-01-25_17_15_51_132
17:22:27 <Kiall> So - rjrjr_ was the one who suggested this at the mid-cycle, so figured I'd throw it out there and see what people think?
17:22:41 <Kiall> I'm personally not a fan of yet anything thing to -1 me ;)
17:23:15 <timsim> I don't know about voting, but it'd be a nice thing to keep track of for new patches.
17:23:17 <Kiall> With our current code - and a threshold of 10 - we have ~10 places where we exceed the limit..
17:23:30 <Kiall> timsim: I'm not sure I can make it non-voting to be honest...
17:24:15 <Kiall> It's ran as part of the pep8/flake8 job.. So can't easily pick+choose what votes from that
17:24:22 <mugsie> ah
17:24:28 <Kiall> rjrjr_: thoughts, since you brought it up? :D
17:24:48 <timsim> Ah. Well, maybe we make it a point to fix the more problematic ones (bugs?) and run it periodically for new things on our own and file more for things that have gotten out of hand?
17:24:59 <mugsie> hum, personally I dislike metrics like ^ when they are a used in a hard rule type of way
17:25:04 <rjrjr_> one way or another, i have to address this to get designate through our internal processes.
17:25:30 <Kiall> rjrjr_: yea, it would be interesting to see if ^ results line up with what your tooling is complaining about :)
17:25:30 <rjrjr_> but, it is not an immediate thing.
17:25:51 <rjrjr_> that and code coverage are the big 2.
17:25:59 <Kiall> I think I agree with timsim.. Something we run periodically outside the gate + address outliers..
17:26:03 <rjrjr_> we addressed that at the last sprint mostly.
17:26:11 <Kiall> started to anyway :)
17:26:48 <rjrjr_> in the end, it will make for easier to read code, hopefully.
17:27:09 <Kiall> rjrjr_: yea, I looked at each of the methods it highlighed - some were mine - and I still got confused ;)
17:27:42 <Kiall> But - I'm not convinced about making it a hard rule in the gate..
17:28:09 <rjrjr_> i think if we fix the ones it complains about, we can make it a rule later.
17:28:33 <rjrjr_> and if we decide we need the complexity, then we don't need to make it a hard rule.
17:28:35 <timsim> Yeah I don't like that. You could do that and make it something ridiculously high, but otherwise, eh.
17:29:17 <rjrjr_> ultimately, the code problem is real.  a value of 10 is reasonable.
17:29:36 <Kiall> Yea, I'm not sold on it ever being a hard rule .. But I do see the value in the output it generates.. Not sure how I reconcile my two viewpoints on this ;)
17:30:21 <Kiall> Anyway - Let's leave it, for now, as a guideline.. Each of those methods identified should be considered a bug, but the gate won't enforce 10 or less - for now.. Sound good?
17:30:23 <mugsie> Kiall: set up a sonar isnatnce as a 3rd party build ? ;)
17:30:50 <rjrjr_> kiall: agreed.
17:30:57 <Kiall> mugsie: go for it - tell me when your done  ;)
17:31:05 <timsim> Sounds good
17:31:11 <Kiall> Ok - Moving on :)
17:31:11 <Kiall> #topic Open Discussion
17:31:16 <rjrjr_> mugsie: i can look into that.  sonar is what we use in either case.
17:31:27 <mugsie> I was half joking
17:31:40 <mugsie> I did have it on my longer term plan to do
17:31:52 <Kiall> rjrjr_: I was kidding too :) BUT - I don't see any harm with having a public sonar instance chugging over the code
17:32:07 <mugsie> but in the next month or i could look into it
17:32:12 <rjrjr_> it would help me be proactive instead of reactive.  i'll look into it.
17:32:18 <mugsie> kk
17:32:34 <Kiall> It's not going to be as useful as it would be for C/Java/.NET .. Cuz... Python, but I doub't it would be harmful ;)
17:32:41 <Kiall> dount*
17:32:44 <Kiall> doubt*
17:32:45 <Kiall> -_-
17:32:58 <Kiall> Okay - Any other topics?
17:33:03 <rjrjr_> so, for the open discussion, just a comment
17:33:51 <rjrjr_> i am going to be workiing on our internal rollout for the next few weeks. i am juggling quite a bit right now. so, if it looks like i'm slow, it's just because i'm doing other work too.
17:34:02 <mugsie> rjrjr_: np
17:34:12 <mugsie> I know how you feel ;)
17:34:14 <Kiall> hah - no problem :)
17:34:21 <timsim> Is there any way we can step up and help you out?
17:34:34 <Kiall> ++ to ^ ;)
17:34:38 <rjrjr_> test, test, test.
17:35:00 <rjrjr_> once i have this mdns stuff worked out, we need to test.
17:36:21 <Kiall> Okay - Any other topics from anyone? I'm all out.
17:36:38 <rjrjr_> i'm thinking about a test_service_sqlalchemy.py, test_service_noop.py, and test_service_memcache.py for pool manager.  thoughts?
17:37:10 <rjrjr_> or maybe just a _cache and _nocache since sqlalchemy and memcache should behave the same.
17:37:24 <timsim> Should probably test that they do though.
17:37:46 <Kiall> rjrjr_: It sounds like 1 suite to me.. with the cache sensitive methods mocked in a few of the single tests
17:38:10 <rjrjr_> all the methods are cache sensitive.
17:38:31 <rjrjr_> we only have 4 public methods in pool_manager.
17:39:02 <rjrjr_> and i figured we'll want tests for the periodic_recovery and periodic_sync too.
17:39:10 <rjrjr_> but ultimately, all cache sensitive.
17:39:24 <Kiall> Humm - Aren't we moving the cache pieces into only the update_status methods, rather than pre-creating etc?
17:39:51 <Kiall> (I'm probably forgetting something TBH - Been away for way too many hours already)
17:39:55 <rjrjr_> not that i'm aware.
17:39:56 <Kiall> awake*
17:40:50 <Kiall> I think I'll need to eyeball the code again to be able to say ^ with any amount of confidence ;) Let's leave it as whatever you have already for now and can change later if needs be?
17:40:54 <rjrjr_> unless we want the backend drivers calling update_status.  right now, we are making calls to the backend drivers.
17:41:46 <rjrjr_> but we decided that was a bad approach last week.
17:42:17 <Kiall> Yea, I think I remember - too tired to be able to think back that far though ;)
17:42:33 <rjrjr_> okay.  i'll get the code working and we can decide then.
17:42:38 <Kiall> ++
17:42:53 <Kiall> Okay, unless anyone else has something.. Let's call it a day!
17:42:53 <timsim> I thought it was more like, create the cache entries when you're creating/updating/deleting, but if they don't create it's cool because each piece that uses them can handle them not being there.
17:42:58 <Kiall> or not ;)
17:43:08 <rjrjr_> timsim, correct.
17:43:12 <timsim> Alright, I'm good then :P
17:43:33 <rjrjr_> all i was suggesting was a separate suite for each cache driver.
17:43:59 <rjrjr_> or one suite with all the test methods x3.
17:44:47 <Kiall> I think there's a seperate test suite for each driver, but only for the driver code rather than testing the driver 3 times through pool manager - Maybe that's where we got out of sync? ;)
17:45:27 <rjrjr_> each driver has a test.  but a noop driver will behave different than the memcache and sqlalchemy driver.
17:45:40 <rjrjr_> inside pool manager.
17:46:16 <rjrjr_> should have wrote 'each driver has a test suite of their own'.
17:46:26 <Kiall> I'm starting to remember ;) You're making more sense to me now :)
17:46:55 <rjrjr_> how about this.  i'll do what i want and everyone can -1 it if they don't like it. :)
17:47:06 <Kiall> lol - perfect.
17:47:11 <timsim> Sounds good
17:47:16 <rjrjr_> i'm good then.
17:47:22 <Kiall> Okay, let's call it a day then :) Thanks all!
17:47:30 <Kiall> #endmeeting