16:00:01 #startmeeting OpenStack-Ansible 16:00:02 Meeting started Thu Jun 23 16:00:01 2016 UTC and is due to finish in 60 minutes. The chair is mhayden. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:05 The meeting name has been set to 'openstack_ansible' 16:00:12 #topic Roll call! 16:00:12 nice multithreading skills mhayden 16:00:14 o/ 16:00:17 o/ 16:00:23 evrardjp: trying my best :) 16:00:27 o/ 16:00:35 o/ 16:00:51 \o/ 16:00:55 o/ 16:01:26 o/ 16:01:28 o/ 16:01:33 yo 16:01:39 o/ 16:01:43 𝙒𝙊𝙊𝙏 16:02:17 looks like we have pretty good attendance today -- let's do the thing 16:02:24 #topic Review action items from last week 16:02:56 first one was midcycle related 16:03:05 but we will get to that in a minute in detail 16:03:25 next up was odyssey4me doing quite a few things around gating -- getting xenial/centos checks from non-voting to voting 16:03:31 and adding experimentals to openstack-ansible 16:03:43 i already saw some of this getting done 16:03:48 o/ 16:03:56 o/ 16:04:05 the security role has voting checks on trusty/xenial/centos, whee 16:04:27 okay, onto the other topics 16:04:34 #topic Newton Mid-cycle Meeting 16:04:43 #link https://etherpad.openstack.org/p/osa-midcycle-newton 16:04:47 the etherpad is coming along nicely 16:04:52 19 attendees so far 16:05:04 looks like 7 needs hotels 16:05:31 i'll see if we can get a hotel deal set up at the Aloft for that, but it might not be enough people -- then again, the deal was $9 off per night :/ 16:05:44 but they do offer a free shuttle to rackspace from what i understand 16:05:46 That’s some deal! 16:05:52 mhayden what about the regular Rackspace discount? 16:06:12 ah, do we have a discount code? let's not share it here, but if we do, that could be nice 16:06:41 mhayden: I know they booked me there for my interview so I'm sure they have something with them 16:06:59 the room is confirmed here at the Castle -- we should have large-screen TV hookups + audio 16:07:12 it's *possible* the room has VC -- but i need to walk up there and check 16:07:31 and the dates are set as Aug 10-12 16:07:53 if you haven't added yourself to the etherpad, please do that soon 16:08:06 and it would be cool to plan some type of social outing one of the nights 16:08:14 mhayden: Thx for all your footwork on getting this pulled together 16:08:17 so if anyone has excellent party planning skills, let me know 16:08:20 +++++ on the social outing 16:08:22 automagically: no problemooooooo 16:08:32 it will be toasty, so we could do something indoors 16:08:33 mhayden: I'll help with that 16:08:46 mhayden: 10 need hotels 16:08:56 #info Anyone interested in the social event at the mid-cycle -- please speak to spotz 16:08:58 thanks spotz 16:08:59 don't know if that changes something 16:09:28 thanks spotz, I'll be happy to attend 16:09:32 three of those are tentative 16:09:35 not sure i will attend social, but you can almost guarantee outside is going to be unpleasantly hot 16:09:46 i'll ask -- if we come up short on what we book at the hotel, rackspace has to foot the bill 16:09:54 :| 16:09:58 lol 16:10:11 #action mhayden to check on how many rooms would be in the minimum block 16:10:28 anything else to talk about on the mid-cycle? 16:10:35 i'm sure we will assemble the agenda a little later 16:10:44 FWIW, due to my corp travel policies I _may_ not be able to stay at the aloft anyway. I’ll let you know mhayden later today 16:10:49 i'll inquire about getting building tools for the folks flying in 16:10:55 automagically: oh okay 16:11:03 thanks 16:11:08 anything else on the midcycle? 16:11:16 so pushy 16:11:31 nothing, except do we have the priorities agenda yet? 16:11:47 evrardjp: not that i know of. that could be a good question for _ody 16:11:49 or we'll figure that out later? 16:11:50 ok 16:11:55 errr odyssey4me 16:12:01 Will wait next week for that 16:12:06 sounds good 16:12:07 fine 16:12:16 #topic Release Planning & Decisions 16:12:33 so 13.1.3, 12.0.15 and newton milestone 1 are out 16:12:39 that kilo one is still dangling 16:12:59 was it due to swift? 16:13:16 i'm not sure what's hanging that one up -- i assume someone hasn't done kilo-eol tags yet 16:13:20 I don't remember who hasn't given feedback yet 16:13:35 mhayden: yes, that's what I think too 16:13:45 I don't recall which one, and I didn't follow that 16:13:50 FWIW, i am still hacking on a script to help with better diffs on those releases -> https://review.openstack.org/#/c/332858/ 16:14:16 that won't work terribly well on newton yet since our role requirements are still set to 'master' 16:14:54 we can sync up with odyssey4me on that kilo release when he's back from holiday 16:15:08 so that was the end of the agenda for today... 16:15:09 master is still WIP, so I think it shouldn't be a worrying target 16:15:12 #topic Open floor 16:15:37 anyone need help or want to share something? 16:16:05 Hey, wanted to ask this here so in front of the larger group. Is there a reason why OSA is using MariaDB 10.0 instead of 10.1? 16:16:50 I ask because currently 10.0 isn't available for ppc64el Xenial, only 10.1 is. 16:17:03 not that I am aware of, we did a bump a while back from another thing 16:17:11 maybe 10.1 is possible 16:17:14 At least in the official mirrors (including the RAX one) 16:17:21 For those of you using OVS or have used OVS or want to use OVS with OSA, I’d appreciate you keeping an eye out on https://review.openstack.org/#/q/topic:bp/support-openvswitch+status:open 16:17:26 I’d love help and love reviews 16:17:32 that is a good question -- maybe cloudnull would have some input on the mariadb version? 16:17:36 adreznec: do you think it will be a big work? 16:17:52 automagically: sounds good 16:17:59 automagically: starred, will do whenever possible 16:18:02 adreznec: no reason in particular that I'm aware of 16:18:07 evrardjp: which way? To get 10.0 working on ppc64el Xenial, or to move forward to 10.1? 16:18:15 I think there were forks that played with 10.1 a while back 16:18:16 to move forward 16:18:24 but I've not seen any movement on that in a bit 16:18:43 I know that the galera bits are included with 10.1 so it may be a good time to change 16:18:48 That's not 100% clear to me. Both are stable, but I'm not sure what the forward movement/upgrade impact would be 16:18:50 cloudnull: was there a reason to not merge at that time? 16:18:55 however IDK what the state is at this time / 16:19:00 evrardjp: no. 16:19:20 I dont remember ever seeing a PR 16:19:20 ok 16:19:20 I think it was only discussed 16:19:44 adreznec: I've not given it a go myself, so IDK what the upgrade impact would be at this time. 16:20:02 adreznec: do you think it would be possible for you to try it and give us feedback? 16:20:04 if you have some cycles to give it a try in a lab it'd be appreciated for sure. 16:20:19 Ok. Not sure if anyone else has more experience there, but I could throw together a test environment quick on the side running 10.1 to see what the impact is 16:20:28 ++ 16:20:32 https://mariadb.com/kb/en/mariadb/upgrading-from-mariadb-100-to-101/ 16:20:36 cool 16:20:48 Thanks mhayden 16:20:50 #link https://mariadb.com/kb/en/mariadb/upgrading-from-mariadb-100-to-101/ 16:20:53 it looks like some options changed defaults, but nothing too nuts 16:21:04 ^ just adding so it'd be in the logs 16:21:11 the upgrade process looks standard, too -- upgrade and run mysql_upgrade 16:21:18 thanks, cloudnull :) 16:21:19 Yeah, looks pretty straightforward 16:21:31 our current role should support the upgrade process 16:22:24 Sounds good, I'll take an initial stab at running with 10.1 here and see how things go then 16:22:47 #action adreznec to look at upgrading to MariaDB 10.1 and report back 16:22:52 thank you, adreznec 16:23:45 also, i'll try to assemble that big WHOA report monthly for OSA, so if you know something that should go in the report, please don't hesitate to send me something 16:23:49 thank you indeed 16:24:01 it's hard to keep up with all of these moving pieces :) 16:24:34 indeed that's a whoale lot of changes in a month 16:24:49 mhayden, https://review.openstack.org/#/c/332595/8 will be important 16:25:13 the june report is here -> https://major.io/2016/06/15/whats-happening-openstack-ansible-whoa-june-2016/ 16:25:14 I agree with palendae it's kind of a big deal IMO 16:25:30 palendae: interesting 16:25:35 Still needs docs work before final submittal 16:26:10 * mhayden starred it 16:26:49 also, if darrenc was awake now, he'd probably ask for a hand on reviewing his doc patches 16:26:58 he's got some good stuff going for the install guide 16:27:59 okay, anything else to talk about? 16:28:30 that was the fastest community meeting ever? 16:28:38 our agenda was a little light 16:28:43 :) 16:28:44 : - ) 16:28:46 I need to get a blueprint in for a reworked inventory system 16:28:47 https://review.openstack.org/#/c/331786/ thinking of handling conversion of tables in neutron, aodh, horizon databases in upgrade playbook 16:28:54 I’ve been reviewing Darren’s doc patches. Definitely need more eyes there 16:29:08 it looks like the others are all done within project db migration scripts 16:29:08 I think I'm mostly done with the current stuff, and I want thoughts on my thoughts before I proceed 16:30:12 palendae: that's a good idea, I'd be happy to give my opinion on this :p 16:30:23 jmccrory: good job for the investigation 16:30:39 jmccrory: i like the idea of consistency -- do we have a plan to convert existing tables over to the right collation? 16:30:44 evrardjp, Yeah, I saw your mailing list thread. I want to consolidate my thoughts and note where I think there's going to be some issues 16:31:11 mhayden: that's what i was hoping to clear up. i was thinking maybe a playbook upgrade for newton 16:31:28 upgrade playbook* 16:32:03 palendae: when this mail was written, ara didn't even exist, so there are maybe experiences to share/take/give to build a larger community too 16:32:08 jmccrory: alrighty 16:32:15 evrardjp, Yeah 16:32:29 I know OSCI is working on Craton, which may feed in too 16:32:32 OSIC* 16:32:50 I big thing I wanted to do was lay out what the current inventory does and why, then address the problems with it 16:32:56 jmccrory: the upgrade playbook seems the most readable approach, we just need to make sure everything is idempotent if possible 16:33:05 ^ yes 16:33:19 I'm a big supporter of upgrade playbooks and them being idempotent 16:33:38 evrardjp: yep, just worried about turning off foreign key constraint for that moment 16:33:55 Oh yes... :/ 16:34:01 jmccrory: so the issue is that some tables are *explicitly* created as latin1 by the projects for some reason? 16:35:17 oh, the problem was we default to different collation than charset 16:35:55 jmccrory: i wonder if this would make a good ML discussion 16:36:23 i'm not sure if i totally understand the problem 16:36:36 it has been a while since i dealt with collation shenanigans 16:37:52 so our my.cfg hardsets the default collation to utf8-unicode, but mariadb's default for utf8 charset is utf8-general 16:39:11 okay, i see 16:39:39 so your patch addresses 1) new deployments and 2) new tables added to existing deployments after the patch is applied 16:39:52 but the question is: what to do about those other collations that are hanging around? 16:39:57 so it was breaking a few days ago after cinder changed their initial db migrate script and there were different collations between tables 16:40:16 yep 16:41:05 nothing's broken at this point, just a consistenty thing and avoid this issue in the future 16:41:25 jmccrory: i gave you a +2 -- this stops the bleeding and makes new deployments less broken :) 16:41:32 but we need to figure out the follow up for sure 16:41:59 but automagically still has the W-1 hammer on that one 16:43:27 I can be reasoned with 16:43:27 okay, do we want to take this discussion to the main channel and close up the meeting? 16:43:29 ++ mhayden 16:43:45 alrighty, well thanks everyone! :) 16:43:48 #endmeeting