16:00:00 #startmeeting openstack_ansible_meeting 16:00:01 Meeting started Thu Mar 9 16:00:00 2017 UTC and is due to finish in 60 minutes. The chair is mhayden. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:04 The meeting name has been set to 'openstack_ansible_meeting' 16:00:07 #topic Roll Call 16:01:07 o/ 16:02:04 its just me and you major ;) 16:02:30 i'm here 16:02:41 it is close to spring break in many parts ofthe US 16:02:58 o/ 16:03:27 o/ 16:03:30 o/ 16:03:41 o/ 16:04:00 * mhayden will wait til :05 after 16:04:05 I still think I should do \o 16:04:31 at least it's not o? this time 16:05:12 or maybe \o/ because I'm super enthusiastic! 16:05:25 #topic Review action items 16:05:25 it's :05 16:05:46 looks like we didn't have any 16:05:59 Hooray 16:06:02 or wait 16:06:10 we need a name for the OSA inventory python package 16:06:14 palendae: any updates there? 16:06:40 mhayden: None yet; I didn't get to sending an email to the mailing list. Last I checked there weren't a ton of votes for anything 16:06:51 Maybe 1-2 for each name https://etherpad.openstack.org/p/osa-python-package-name 16:07:21 i still think openstack-ansible-palendae-inventory-improvement-thing-2017 has a nice ring to it 16:07:36 hahahahahaha 16:07:40 yeah that is pretty good 16:07:44 ship it 16:07:57 Dashes aren't allowed ;) 16:08:01 palendae: could you send a few final candidates to the ML and we can sort it out? 16:08:05 with your name on it you'll still be fielding questions for many years to come :P 16:08:08 And I think that goes over the line limit ;) 16:08:08 palendae: replace with underscores 16:08:11 osai 16:08:12 really 16:08:14 who came up with osai? 16:08:20 * palendae shrugs 16:08:24 I proposed none of these fwiw 16:08:32 mhayden: Yes, I will try to get that done this week 16:08:39 palendae: woot 16:08:54 #action palendae to email a list of a few final names to the list for discussion/voting 16:09:09 the other action item was for logan- --> "add ceph_client version bump to development cycle checklist" 16:09:27 o/ 16:09:30 still on todo 16:09:31 sorry was trying to catch up on dev list:) 16:11:20 hah okay 16:11:26 we'll keep it then 16:11:36 #action logan- to add ceph_client version bump to development cycle checklist 16:11:51 #topic Ops Guide - asettle 16:11:55 * mhayden hands the mic to asettle 16:12:02 Not much! 16:12:03 REviews :) 16:12:04 Content 16:12:08 You konw the drill hoomans 16:12:17 Those who have signed up for content, add it in 16:12:20 We have until May 16:12:51 I think that's all mhayden :) 16:13:00 asettle: woot! 16:13:26 next up is naming the python package -- i think we've covered that 16:13:32 moving along... 16:13:37 #topic New periodic jobs - odyssey4me 16:13:41 * mhayden hands the mic to odyssey4me 16:13:49 howdy folks 16:13:51 excitement inc! 16:14:01 so, we have a periodic upgrade test for xenial which runs once a day 16:14:05 \o/ 16:14:09 it's the first of many periodics planned 16:14:43 I'd like to ask that we try to figure out how we expose the results a little more actively 16:14:49 the results are posted here: http://logs.openstack.org/periodic/ 16:15:02 #link http://logs.openstack.org/periodic/ 16:15:14 we only have one right now, but once there are more the folder is the job name 16:15:21 http://logs.openstack.org/periodic/periodic-openstack-ansible-upgrade-aio-master-ubuntu-xenial/ 16:15:32 odyssey4me: can we get a notification via IRC bot perhaps? 16:15:33 I was thinking maybe that we could add a dev-docs page which shows where they are 16:15:49 mhayden not to my knowledge, but I can ask 16:15:59 that's a good idea 16:16:09 make me an action maestro 16:16:22 YARLY 16:16:34 I've been bashing in a bunch of fixes to try and make it succeed 16:16:35 we could hack something up, but it wouldn't be as good most likely 16:17:00 ideally I'd like to see it exposed in the OpenStack Health Dashboard 16:17:18 http://status.openstack.org/openstack-health/#/ 16:17:22 I've asked cloudnull if he can try and make ARA output subunit streams instead of just junit 16:17:27 * mhayden hasn't looked at this fancy thing in a while 16:17:39 not sure if anyone has some time to sinnk into doing that 16:18:10 our automated upgrade script is a bit of a shambles, to be honest 16:18:14 we need an overhaul 16:18:48 any volunteers to work with me to work on a conprehensive improvement? 16:18:56 *comprehensive 16:19:22 * odyssey4me looks at logan- and jmccrory 16:19:30 sure 16:19:36 logan-: jmccrory when you are finished fighting over who gets to do it :P 16:19:39 :D 16:19:44 lol 16:19:44 voluntold! 16:20:02 internet just dropped here 16:20:05 back now 16:20:26 mhayden: we decided that jmccrory would rehaul the upgrade bits 16:20:26 ok, logan- and jmccrory I'll setup an etherpad to discuss what we want and break it into pieces 16:20:30 you can help 16:20:41 lol 16:20:47 ok thanks odyssey4me 16:20:58 evrardjp: success! 16:21:02 best ever. 16:21:03 no, thank you for volunteering logan- and jmccrory :) 16:21:16 * mhayden goes and orders upgrade fixes for pickup at walmart 16:21:36 okay, are we good on this topic ? 16:21:41 right, but back to the periodics 16:21:54 are there any other tests we want to add to the list of periodics? 16:22:08 odyssey4me: telemetry is the obvious one right now (since we removed it from the integrated build) 16:22:37 I was thinking of backporting the upgrade test to Newton->Ocata as well to prevent further bitrot 16:22:44 also Ironic 16:22:59 PErhaps some sort of scenario that involves Magnum, Sahara 16:23:19 yeah ironic. smh - i need to try find some time. because this: https://review.openstack.org/#/c/392959/ is now passing so we should be able to reverse engineer that into our ironic gates 16:24:04 ok, once the upgrade test is working I'll troll for ideas in channel 16:24:11 I think we're good here. 16:24:38 boom - and thanks for the work, its awesome to see the upgrade test up and running (and almost passing) 16:24:49 +1 16:25:07 speaking of - i will be fixing up our in role upgrade tests and making them voting 16:25:26 yeah, there are a bunch of tests that should be promoted 16:25:35 otherwise patches merge which break them 16:25:42 yeah basically 16:25:45 the ceph AIO job should be made voting IMO 16:25:51 currently i think neutron/nova are broken the others are mostly working so going to get that set up 16:25:53 the xenial one I mean 16:25:58 yeah happy with that too 16:26:50 ok i think we're good to move on mhayden! 16:26:53 woot 16:27:21 #topic Release planning and decisions - andymccr 16:27:24 * mhayden hands it to andymccr 16:27:50 ahh yeah 16:27:50 so 16:27:52 drum roll 16:27:57 Ocata 15.0.0 was released this week! 16:28:05 WOOT 16:28:11 i know, i know. 16:28:17 #success OpenStack-Ansible 15.0.0 Ocata released this week 16:28:46 but genuinely, thanks for all the work during the cycle - aside from the rc1 hiccups it really was pretty smooth (and those werent necessarily our messups) 16:28:51 so onward and upward to Pike! 16:28:54 gratz andymccr - it's looking pretty good, we fixed up a lot of upgrade things and should be doing plenty more 16:29:14 thanks andymccr 16:29:23 Newton/Mitaka are being released as usual, so next week will be another released, I'll SHA bump for Ocata but only do a release 15.0.1 2 weeks after (so we can keep them all released at the same cadence) 16:29:50 "give back to caesar what belongs to caesar" 16:30:07 * mhayden wants a caesar salad 16:30:24 ahhaha 16:30:27 yeah that! 16:30:38 give that back! 16:30:45 there are a few critical bugs up still, the keystone one and a high one around mysql upgrades - so we still have some bits to do 16:31:16 that's all i got on releases! 16:31:20 \o/ 16:32:48 mhayden: next! 16:32:59 #topic Blueprints 16:33:08 xgerman: want to give us a brief on octavia? 16:33:17 sure 16:34:16 Octavia is an operator grade load balancer… and I put together some playbooks to install it 16:34:56 the main advantage is that it is HA as opposed to the previous one which didn;t have redundancy 16:35:27 I set it up that if Octavia isinstalled Neutron will make it the default load balancer in the system 16:35:45 nice -- what do you need from us to finish up the work? 16:35:58 the missing two patches merged 16:36:12 #link https://review.openstack.org/#/c/428979/ 16:36:44 #link #link https://review.openstack.org/#/c/422062/13 16:37:01 the lat one has some jenkins problem I don't understand 16:37:34 Once I have it running on. a multnode cluster I will write up some installation guide 16:37:52 that will be great -- docs on octavia are a little sparse ;) 16:38:08 it’s a major push for the Octavia team this cycle 16:38:14 xgerman: feel free to tag me in the review, happy to watch your docs :) 16:38:17 'watch' sigh 16:38:18 Review them 16:38:22 thx 16:38:32 also I am planning to keep up with the milestone releases 16:38:39 xgerman: looks like tempest is angry in the gate 16:38:44 xgerman getting reviews for those is a bit hard because not many people understand networking well enough to review properly 16:38:45 http://logs.openstack.org/62/422062/13/check/gate-openstack-ansible-openstack-ansible-aio-ubuntu-xenial/9d149ab/console.html#_2017-03-09_15_13_14_900543 16:38:46 yep 16:38:54 but I couldn;t see why 16:39:01 that formatting is crazy 16:39:34 i can try to build an AIO from that patch later and see what's happening 16:39:37 xgerman does it work if you test locally? 16:39:56 Binding failed for port 52dd82ba-6e57-45c0-9ae5-bbe3cfe13682, please check neutron logs for more information. 16:40:03 mmh 16:40:30 I will run it locally and see if I have more luck 16:41:41 also do you need Octavia backported to Ocata? 16:42:07 xgerman: we wouldn't backport a feature 16:42:16 ok, that’s what I thought 16:42:21 a major feature 16:42:35 si si :) 16:44:19 especially with such a small user/dev base 16:44:50 it also is entirely unproven - we need to grow a user base and hopefully pick up more than one maintainer for the role 16:45:16 yeah ideally - that'll happen with time though, so its not really a stress 16:45:53 there's interest where i work, i'm pushing to get them to test the patches out there for review now instead of the manual install they've been doing 16:46:09 jmccrory: sounds awesome 16:46:11 +1 16:46:45 excellent :) 16:46:55 I also have some interest from other people in the community - once it works it could be quite popular 16:47:02 excellent, if we can get more than one org maintaining it then we can encourage its use 16:47:09 it can mature a bit and release in pike properly 16:47:16 woot! 16:47:26 yeah agreed 16:47:29 such collaboration, much happy 16:47:32 yeah, it’s in early… 16:47:34 "once it works" doesn't sound motivating 16:47:58 the glass is half-full, darnit! 16:48:25 yeah, my installations work but I haven’t tried with everything merged properly 16:48:27 in other news, cloudnull and i have been working on the monitoring plugin PoC as part of the monitoring spec 16:48:31 ahh yeah 16:48:32 https://github.com/major/monitorstack 16:48:32 awesome 16:48:45 mhayden: glass is always full! 16:48:57 test coverage is increasing and it's easily buildable in pypi 16:49:17 i'd like to bring it under the openstack namespace at some point 16:49:24 mhayden: yeah i like that idea 16:49:26 i've been talking to mrhillsman about that 16:49:35 he said the ops folks would be interested 16:50:12 would it make sense to have it under openstack-ansible's responsibility for a bit until it finds a forever home? 16:50:13 mhayden: would end goal to test it using OSA (and perhaps vice-versa?) 16:50:19 andymccr: yeah 16:50:23 mhayden: yeah im happy with that 16:50:30 not sure if there are any objections 16:50:34 the aim would be to have it be generic 16:50:40 but we are maintaining it for now, similar to the security role 16:50:46 i just called it monitorstack because i couldn't find anything big in Google for that 16:51:36 mhayden: in OpenStack-Ansible we have a rich tradition of inventing smart and fanciful names - i expect you to uphold that tradition 16:52:13 I'd just suggest that we don't prefix the repo name with openstack-ansible 16:52:30 It'll need to be a more universal name. 16:52:48 hah 16:52:49 Perhaps it should be its own project name, but we curate it initially. 16:52:52 odyssey4me: current name is 'monitorstack' 16:53:21 i dont mind either way - if its more work to create a separate project now, then happy to just include it in the openstack-ansible namespace for now 16:54:00 okay, we can talk on that one later in the channel 16:54:05 #topic Open floor 16:54:08 anything else? 16:54:13 all good from my side 16:54:57 mhayden andymccr I'm suggesting that it just be another repo in our deliverables, but not with our prefix 16:55:06 ahh that works 16:55:09 odyssey4me: totally agreed 16:55:17 agreement has been reached! 16:55:17 therefore it can move out of our deliverables at another time without the hassle of renaming the repo 16:55:45 woot 16:55:50 okay, i'll close this meeting up 16:55:53 it could possibly actually fit under the oslo umbrella, but let's not even try 16:55:53 thanks everyone! :) 16:56:01 #endmeeting