18:00:19 #startmeeting trove 18:00:20 Meeting started Wed Sep 14 18:00:19 2016 UTC and is due to finish in 60 minutes. The chair is amrith. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:21 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:25 The meeting name has been set to 'trove' 18:00:28 o/ 18:00:30 ./ yo 18:00:40 o/ 18:00:40 courtesy ping to attendees peterstac johnma dougshelley66 pmalik vgnbkr mvandijk trevormc aliadil spilla songjian trevormc apsarshaik 18:00:48 😝/ 18:00:55 if your name isn't in the list above, please add it to the list in the meeting agenda page ... 18:00:57 hello! 18:01:00 o/ 18:01:12 o/ 18:01:33 \o\ /o/ 18:01:38 let's give folks a couple of minutes to come in, and I'll go get some coffee 18:01:49 hi, this is the trove meeting 18:01:51 that's dbaas 18:01:53 not the video game 18:02:04 maybe you are lost cp16net ... are you new here? 18:02:05 o/ 18:02:07 o/ 18:02:09 o/ 18:02:29 i have a bug where i cant get treasures in trove 18:02:35 long time no see cp16net 18:02:38 i'm so poor i need more moneys 18:02:42 lol 18:03:01 howdy 18:03:18 so cp16net I found some issues with this particular commit ... 10d738aad29777ebdb31b286ce5cd750e298e2ad 18:03:24 thought i would see how things are going 18:03:24 could you please go fix it :) 18:03:42 not my fault! 18:03:46 :-P 18:03:47 FYI, that's was your first commit to Trove in case people want to go and see how long ago that was 18:03:50 everything is your fault 18:03:57 let's get started 18:04:03 thats true... 18:04:08 if someone in Canada would jog peterstac ... 18:04:16 #agenda https://wiki.openstack.org/wiki/Meetings/TroveMeeting 18:04:24 o/ 18:04:30 #topic Pulse Update 18:04:52 and for those of you who don't know, cp16net, he used to do some trove stuff (not a lot) a while ago. 18:05:07 so we had a bunch of activity over the past week 18:05:19 and we're close to pushing an RC1 out of the door for Newton 18:05:25 and we'll talk more about that in a bit 18:05:35 but the review count this past week was high 18:05:41 because of all the churn related to the release 18:05:46 and thanks to all who did reviews this week 18:05:59 I saw some new names, songjian spilla ... thank you 18:06:22 the graphs and stuff are at 18:06:23 #link https://docs.google.com/spreadsheets/d/1vJxNaoR3VVNS1Cpiz7U--1zyJRZ6ybMxzJoayrjdduo/edit?usp=sharing 18:06:50 if anyone has anything they want to share on the subject ... 18:07:20 I think we hit an all-time high on total reviews 18:07:30 (since we started keeping the graph that is) 18:07:58 yup 18:08:14 sure was a record of sorts 18:09:03 Also, I think I've messed up the formatting of the agenda and I'll fix that 18:09:07 but the next topic is 18:09:09 #topic Newton release schedule 18:09:25 So, today is the deadline for Newton RC1 18:09:33 and I think we're in a generally good place. 18:09:55 We've release our client about 10 days ago 18:10:04 and there are tags for both dashboard and trove at this point 18:10:23 and these releases meet all of the FFE's that we set forth two weeks ago 18:10:46 So, we now have the harder task of deciding whether we are good with RC1 18:10:54 or whether we want to accept more changes in 18:11:06 the majority of what I wanted to discuss is related to that 18:11:22 peterstac, put together a list of changes for consideration in RC1 18:11:27 so let me turn it over to peterstac 18:11:36 and he can walk us through this ... 18:11:49 peterstac, ... sorry I didn't tell you I would put you on the spot. no pressure 18:12:02 no problem 18:12:15 * amrith runs to look at peterstac's list 18:12:20 I went through the list of 'jenkins-ok' bugs and put together a list 18:12:34 that I thought could still make it into newton 18:12:48 most are fairly straight-forward 18:13:19 on the other hand, none are critical - but I thought we could still get them in unless someone objected 18:14:04 (but we'd have to decide quickly, since it'll take 1 1/2 hours to churn through the gate) 18:14:23 I don't believe any conflict with each other ... 18:14:51 amrith, did you want me to go through them one at a time? 18:15:02 fine with me 18:15:13 ok - first: 18:15:17 or if people have questions or objections about any one of them, we could talk abou tit 18:15:19 about it 18:15:25 your call 18:15:27 floor is yours 18:15:45 it would probably only take 10 min or so 18:15:58 and then we'll have it over with :) 18:16:03 initial chkin of pylint 18:16:10 #link https://review.openstack.org/#/c/367640/ 18:16:26 This is a change that amrith put up to run pylint on the code 18:16:53 it would be the first step to having it run in the gate (first as a non-voting job of course) 18:16:59 I think it is low risk 18:17:08 I think it would help keep our code clean 18:17:19 and yes, very low risk 18:17:36 cp16net, johnma your thoughts? 18:17:40 yes I think that would be good to get it in for RC1. 18:18:35 johnma, I made your recommended correction 18:18:42 I did find something in the process (the tool did) 18:18:47 that I wanted to bring to your attention 18:19:05 It flagged this as an error 18:19:06 ERROR: trove/guestagent/datastore/experimental/db2/service.py 256: E1101 no-member, DB2App._update_dbm_config: Instance of 'str' has no 'is_warning' member 18:19:10 yes I saw that amrith 18:19:18 I'm not sure whether it is a real error 18:19:20 or benigh 18:19:24 benign 18:19:27 I ran it on my local env and am fixing it 18:19:45 but want to highlight it because it isn't one of the normal "only-known-at-runtime" kinds of issues that cause pylint to give false positives 18:19:49 so, your call 18:20:07 as peterstac said the tool has no associated tests (in the CI) though you can tox it. 18:20:14 I didn't rebuild the errors file 18:20:36 * amrith shuts up and looks to peterstac (the MC) to continue 18:20:52 right, so it will probably output a couple of errors now, but I think we can fix them after we get a nv check job running 18:21:21 +1 18:21:34 (in other words, I'm good with it :) ) 18:21:42 * amrith turns and looks to the rest of the attendees 18:22:23 me too 18:22:33 ok, sold! 18:22:42 (anyone else??) 18:23:03 next: Deprecate 'guest_log_long_query_time' 18:23:13 #link https://review.openstack.org/#/c/311237/ 18:23:33 my only question about this is whether we are really 'deprecating' the variable 18:23:50 or are we just not going to honor a legacy setting and use the new one instead? 18:23:54 This basically allows the end user to specify the slow query time, instead of having it hard-coded by the operator 18:24:13 if it is set by operator, with this change, what happens? 18:24:17 is that value ignored? 18:25:05 * peterstac runs off to look at the code again 18:25:39 my suspicion is that it is ignored 18:25:56 see my comment from 18:25:56 May 6 9:15 PM 18:25:58 Well. The way it is now: the value from the config group takes presedence over the value from Trove config. 18:26:26 pmalik, if there is no value in the config group will the value in trove config be used? if yes, I have no objections 18:26:36 if no, I have reservations but see the value of the change. 18:26:55 and realize that there is no universally "right" way around this. 18:27:23 Yes. I made it this way as a reaction to your comment from May 6 18:27:46 OK, I will re-review; I missed that part. 18:27:49 thanks pmalik 18:27:51 So if there is not value in the config group the value from Trove config gets used. 18:28:03 I have no objections (subject to that re-review). 18:28:06 that's fine 18:28:09 that is ideal 18:28:30 well, ideal is a strong word. that is justifiable, rationalizable, ... 18:29:00 peterstac, you already have a +2 on it, yes? I'll re-read it shortly 18:29:04 I will review this as well 18:29:10 I kind of suspect as well that no operators have changed the default 18:29:59 yes, I put a +2 on it 18:30:01 pmalik, the value is now in *all* datastore specific configs or only the ones where we use it now? 18:30:16 it would only be where there are user logs 18:30:25 so mysql variants and postgresql 18:31:09 thx pmalik, +2 18:31:46 ok, next: Use common methods for cluster validation 18:31:57 #link https://review.openstack.org/#/c/320687/ 18:32:30 this consolidates the validation for cluster stuff (like flavors, etc.) into one spot 18:32:49 pmalik, any comments? 18:33:40 i wouldn't say that is worth it for N 18:33:44 vote punt to o 18:34:18 It has been there quite a long time. If the community feels like it is ok to go now... 18:34:21 the large file seems to be an add 18:34:27 the rest seem to be reasonable merges 18:34:45 it is cleanup only 18:34:59 not worth adding risk at this point 18:35:01 i'm leaning towards mvandijk's view 18:35:49 I'm ok with that - we'll put it on the agenda for next week then 18:35:50 unless there's a strong reason to get it into newton, or there is low risk, I'm inclined to say no 18:36:17 to be clear, the deprecation thing we just talked about has a strong reason in my mind, it starts the deprecation clock 18:36:17 next: Stop adding ServiceAvailable group option 18:36:20 of one cycle .. 18:36:22 sorry peterstac 18:36:25 go ahead 18:36:35 np, I agree with that 18:36:48 if we deprecate something, we should put out the warning asap 18:37:09 this is test only, yes? https://review.openstack.org/#/c/366963/ 18:37:09 #link https://review.openstack.org/#/c/366963/ 18:37:31 This tweaks the tempest stuff and is test only 18:37:34 and makes the tempest people happy 18:37:37 so low risk 18:37:39 I'm for it 18:37:57 me too. 18:38:07 ok, in that case I'll approve it 18:38:08 thx 18:38:30 johnma, you ok with https://review.openstack.org/#/c/311237/? 18:39:02 oh, way ahead of me 18:39:03 yes amrith. I approved it 18:39:07 you already +1'ed it:) 18:39:53 next: Call GuestError with proper options 18:40:04 #link https://review.openstack.org/#/c/363796/ 18:40:13 This is just cleanup code as well 18:40:27 I'm ok with punting it to Ocata 18:41:01 (it's been there a long time, and just fixes the omission of part of the message) 18:41:53 next: last: Use http_proxy_to_wsgi middleware 18:42:02 #link https://review.openstack.org/#/c/337063/ 18:42:39 i dont know what that does tbh 18:42:54 here's the bug describing it 18:42:57 #link https://bugs.launchpad.net/keystone/+bug/1590608 18:42:58 Launchpad bug 1590608 in OpenStack DBaaS (Trove) "Services should use http_proxy_to_wsgi middleware" [Undecided,In progress] - Assigned to Masaki Matsushita (mmasaki) 18:43:20 before you move on 18:43:25 https://review.openstack.org/#/c/363796/ 18:43:46 yes? 18:44:12 how risky is this? 18:44:16 I don't see much 18:44:28 not much 18:45:19 we could still put it in if there are no objections ... 18:45:40 i don't see a good reason not to ... 18:45:49 anyone? 18:46:05 how do we currently handle HTTP_X_FORWARDED_PROTO 18:46:43 sorry johnma, we switched back to ' 18:46:43 Call GuestError with proper options' :) 18:46:44 I dont see any problem adding this in 18:46:50 peterstac, I've +2'ed it. 18:47:11 put mine back too 18:47:22 as for the wsgi thing, I've tried to understand it, like when the other guy mkrotschek (name?) did some of that 18:47:27 and I didn't quite grok it 18:47:46 yeah, I'm a bit fuzzy on the details too, but Trove still seems to work after :) 18:48:14 well, I think we've got to set the bar a bit higher now ... 18:48:34 johnma, cp16net, how does it look to you? 18:49:02 I am lost. which one are we talking about now peterstac 18:49:42 https://review.openstack.org/#/c/337063/ 18:49:47 i think 18:49:51 sorry, we're back to HTTP_X_FORWARDED_PROTO 18:50:27 The change looks the same as was done in keystone 18:50:45 I'm not in favor of doing this one now; I don't know what it'll break that we haven't thought of 18:50:59 but others who understand this better than me ...you should decide 18:51:48 since this is just configuration, operators could always set this up themselves 18:52:03 so I'm ok to push this to Ocata 18:52:38 any other opinions? 18:52:54 +1 for ocata 18:53:03 mine is a non-opnion in favor of ocata 18:53:10 does that count? 18:53:31 counts for me :) 18:54:21 ok, so that's the list, yes? 18:54:51 just one clarification on https://review.openstack.org/#/c/363796/ 18:54:56 my reckoning is that we said yes to 1, 2, 4, and 5 but not 3 and 6 18:55:04 should we push that over? 18:55:25 that was my question 18:55:35 #5 18:55:36 that's #5, so ok 18:55:42 well, is it 18:55:47 is that what we agreed to? 18:55:54 1, 2, 4 and 5? 18:55:54 I'll push it over the line, so we're done 18:56:00 yes 18:56:07 johnma, is that OK? 18:56:15 cp16net, if you are still here ... 18:56:27 yes amrith. that sounds good to me 18:56:48 ok, peterstac ... 1, 2, 4 and 5 18:56:51 and we're at RC1 18:57:07 wait, johnma is working on one fix 18:57:16 for the err.is_warning() 18:57:20 yes? 18:57:20 amrith, all done 18:57:31 ... not so fast ... 18:57:34 yes I will push that out 18:57:45 ok, so 1, 2, 4, 5 + is_warning() 18:57:53 (all done wrt the six in the agenda) 18:58:01 ok 18:58:17 anyone here for Open Discussion? 18:58:18 I'll look out for the is_warning fix too and test asap 18:58:20 #topic open discussion 18:58:36 sounds good, thanks peterstac, amrith 18:59:01 np johnma 18:59:11 I don't have anything else amrith 18:59:16 so for ocata 18:59:29 can we get https://review.openstack.org/#/c/236082/ in ? 18:59:51 mvandijk, yep, that's on my list of things to push next 18:59:51 have to wait for the split 18:59:58 ok good thanks 19:00:50 thx all 19:00:51 #endmeeting