15:00:15 #startmeeting RDO meeting - 2016-09-21 15:00:15 Meeting started Wed Sep 21 15:00:15 2016 UTC. The chair is imcsk8. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:15 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:15 The meeting name has been set to 'rdo_meeting_-_2016-09-21' 15:00:23 o/ 15:00:24 #topic roll call 15:00:24 0/ 15:00:26 o/ 15:00:27 \o 15:00:29 ¯\_(ツ)_/¯ 15:00:44 (•_•) ( •_•)>⌐■-■ (⌐■_■) 15:00:52 Hmm. The openstack bot is awol. 15:00:53 o/ 15:01:03 oOo 15:01:04 #chair leifmadsen weshay coolsvap dmsimard rbowen jruzicka jpena apevec 15:01:04 Current chairs: apevec coolsvap dmsimard imcsk8 jpena jruzicka leifmadsen rbowen weshay 15:01:17 jruzicka wins this week's rollcall 15:01:17 \o/ 15:01:21 o/ 15:01:30 #chair chandankumar amoralej 15:01:30 Current chairs: amoralej apevec chandankumar coolsvap dmsimard imcsk8 jpena jruzicka leifmadsen rbowen weshay 15:01:48 ok, lets start 15:01:56 #topic Need volunteers/topics to speak at the RDO/Ceph meetup in Barcelona - https://etherpad.openstack.org/p/rdo-barcelona-meetup-schedule We need to finalize the schedule soon so that we can start event promotion. 15:02:14 We currently have just 2 people/topics on the schedule. 15:02:31 quick meetings are quick 15:02:42 Oh I got my ATC and ticket to the summit, so I can rant about something. 15:02:44 The austin schedule is https://etherpad.openstack.org/p/rdo-austin-bof for some idea of what we discussed there. 15:02:47 rbowen: I think it'd be silly to not take the opportunity to talk about rdo+ceph givent he people there 15:03:00 more user-oriented or contributors-oriented? 15:03:04 The format is rather different from last time. 15:03:15 5-10 minute slots, alternating with our friends from Ceph. 15:03:18 rbowen: so let's do something rdo+ceph (i.e, how we are setting up ceph with rdo deployments) 15:03:25 But we can do discussion items once presentations are over. 15:03:29 And, also, free drinks. 15:03:39 Javier Peña proposed openstack/packstack: Fix Neutron lbaas and add it to scenario002 https://review.openstack.org/374229 15:03:43 "free" 15:03:45 gfidente: are you going to barcelona ? 15:04:12 Anyways, if you just want to propose *topics*, we can have open discussion. You don't have to prepare slides, or even particular comments. 15:04:24 ok 15:04:40 o/ 15:04:43 15:04:47 #chair number80 15:04:48 Current chairs: amoralej apevec chandankumar coolsvap dmsimard imcsk8 jpena jruzicka leifmadsen number80 rbowen weshay 15:04:59 ok, next topic... 15:05:05 #topic Updates on tempest/rally 15:05:06 o/ 15:05:11 #chair trown 15:05:11 Current chairs: amoralej apevec chandankumar coolsvap dmsimard imcsk8 jpena jruzicka leifmadsen number80 rbowen trown weshay 15:05:33 dmanchad, rbowen /me would be happy to talk a little about rdo+ceph as well 15:05:33 I have created a draft blog on how to run tempest on packstack env with latest changes in tempest packaging 15:05:35 https://github.com/redhat-openstack/website/pull/702 15:05:43 gfidente: Excelent. Thanks. 15:05:50 Feel free to try it and have your comments there 15:05:57 maybe a walkthrough of http://giuliofidente.com/2016/08/ceph-tripleo-and-the-newton-release.html ? 15:06:20 if for end users 15:06:28 gfidente: That would be awesome. Thank you. 15:06:32 I have also written a magic script which will install all tempest test packages based on openstack environment 15:06:39 gfidente: has to be like 10 minutes tops iiuc :P 15:06:42 https://review.gerrithub.io/#/c/295082/ 15:06:59 dmsimard, even better, we just discuss the features there 15:07:08 instead of showing things from an actual live terminal session' 15:07:23 On rally side, the rally spec got cleaned up https://review.rdoproject.org/r/#/c/2237/ and new rally rpm will be available soon. 15:07:28 and collect feedback, which would be the most precious part of the discussion I think 15:07:35 that's it from my side. 15:08:02 ok, lets move on 15:08:14 #topic rdo-infra-list mailing list 15:08:25 gfidente, it's also joint meetup w/ Ceph folks 15:08:33 Hi 15:08:36 I'm sorry. 15:08:47 I was the first one to tell Duck I didn't want a rdo-infra mailing list. 15:09:11 But the truth is we need somewhere that isn't someone's inbox to send potentially spammy emails (such as CI and monitoring related) 15:09:22 quack 15:09:22 apevec, well the walkthrough would only be useful to end users I suppose 15:09:27 I don't want to send those to rdo-list and expect everyone to filter them out 15:09:30 #chair Duck 15:09:30 Current chairs: Duck amoralej apevec chandankumar coolsvap dmsimard imcsk8 jpena jruzicka leifmadsen number80 rbowen trown weshay 15:09:44 dmsimard, thanks man :) 15:09:46 sorry I'm late, long previous conficall 15:09:53 +1. Mailing lists are cheap, and easy to filter. 15:09:56 Merged openstack/packstack: Change ironic identity to authtoken https://review.openstack.org/368232 15:10:22 So, what do we do ? Can we create a rdo-infra-list either on redhat.com (and migrate later?), waiting for rdoproject.org might take a while 15:10:25 dmsimard, if traffic would be spam only, that's fine 15:10:27 dmsimard: I think a new Ml is lightweight, and the tags you suggested could work but people don't get it 15:10:29 As I was mentioning earlier, I would like our lists to move over to @rdoproject.org eventually, but if that's not available now, we can create an @redhat.com one very quickly and easily. 15:10:34 but I'd set reply-to: rdo-list 15:10:38 It takes a day or less to spin up a new list. 15:10:44 so that no discussion happen there 15:11:08 apevec: I think that makes sense, at least to begin with 15:11:13 it's something that can be reconsidered later 15:11:21 as necessary 15:11:26 +1 on reply-to: rdo-list 15:11:29 +1 for mailing list 15:11:37 So, rdo-infra@redhat.com for now? I can get that requested today if so. 15:11:40 +1 for not spamming rdo-list 15:11:48 Can you set a reply-to at the list level? 15:12:01 If not, any automated mailings would need to set that themselves. 15:12:02 nice :-) 15:12:21 #action rbowen to request rdo-infra@redhat.com creation. 15:12:22 rbowen, there is somehting, we use it for upstream stable-maint 15:12:23 Good? 15:12:27 I'll have a look 15:12:29 rbowen: for the interest of consistency, I'd do with rdo-infra-list@ (we have rdo-list@) but yes I'm fairly sure we can set the reply-to 15:12:40 #action rbowen to request rdo-infra-list mailing list 15:12:51 #undo 15:12:51 Removing item from minutes: ACTION by imcsk8 at 15:12:40 : rbowen to request rdo-infra-list mailing list 15:12:52 we also have rdo-newsletter, so we're already inconsistent. :-) 15:12:52 #undo 15:12:52 Removing item from minutes: ACTION by rbowen at 15:12:21 : rbowen to request rdo-infra@redhat.com creation. 15:12:55 But, sure. 15:13:09 I'll get that requested right away. 15:13:11 now redo :) 15:13:14 rbowen: thanks 15:13:21 #action rbowen to request rdo-infra-list@redhat.com creation. 15:13:25 EmilienM: FYI ^ we'll have rdo-infra-list for CI stuff. 15:13:35 cool 15:14:03 should we continue with next topic? 15:14:08 yeah we're done 15:14:15 #topic RC testday Sep29/30 15:15:39 Last week we talked about ways to make test day more inviting. 15:15:46 That conversation is ongoing on rdo-list. 15:16:12 yeah, we have one week more time to improve it 15:16:16 The test day is just a week out, so probably most of that stuff will be things that we try to do for the next time around. 15:16:29 hrybacki, ^ 15:16:29 But better test case descriptions are always welcome and helpful. 15:16:51 on the repo level, we ( number80 / amoralej / me) are wrapping up RC builds today 15:16:55 Having test day be a way to 1) welcome new users and 2) give existing users a feature preview, are goals, as well as just bug finding. 15:17:04 yep 15:17:09 maybe have folks try to add scenarios to CI vs.. just test them 15:17:20 hguemar created rdopkg: Fix macro expansion when redefining macros http://review.rdoproject.org/r/2587 15:17:23 rbowen: did you get any feedback regarding the docs on images.rdoproject.org ? 15:17:36 Meanwhile, https://dashboards.rdoproject.org/rdo-dev is all green so we're in good shape for the test day on that front. 15:17:40 code vs. test, then we could use those additions for future releases 15:17:41 hrybacki: I didn't see any, no. 15:18:26 rbowen: we're expanding them to include upgrade and update scenarios right now. Aside from more coverage what can I do to increase visability/encourage use of these for the next test day? 15:18:29 But those are linked from the test day pages now as examples of how to test, so that's goodness. 15:18:39 * hrybacki nods 15:18:46 We should link them from the right places in the test matrix 15:19:05 Everything in the test matrix just links to Quickstart, which is not very helpful in most cases. 15:19:11 #chair weshay hrybacki 15:19:11 Current chairs: Duck amoralej apevec chandankumar coolsvap dmsimard hrybacki imcsk8 jpena jruzicka leifmadsen number80 rbowen trown weshay 15:19:33 maybe it's a crazy idea, but what about having some pre-made pool of VMs to test with? Just OS and basic repos, and have a way to handle them to testers during the test days 15:19:47 That would be awesome. How do we make that happen? 15:19:51 jpena: sounds nice 15:19:57 Is that possible for this time, or something to work towards for next time? 15:20:10 rbowen: okay. Closer to the next test day I'll link up with you, discuss scenarios we have covered and update the matrix at that point? 15:20:14 jpena, hmm, that gives them root acess? 15:20:30 hrybacki: That would be great. Thanks. 15:20:35 np! 15:20:43 apevec: yeah, it'd require some controls on who gets the VMs 15:21:25 rbowen: we'd need some public cloud to help with the resources, so I think it could be a longer-term goal 15:21:44 not sure how much we could use of TryStack 15:21:49 jpena: what about using trystack resources for trying out? 15:21:54 jpena, not sure about trust model here, also if it is ok w/ OS1 fineprint to share account 15:22:01 jpena: Probably not for this time, right. 15:22:23 trystack needs upgrade, which is not scheduled before summit iiuc 15:22:38 we could have ansible playbooks for test cases 15:23:39 yup 15:23:52 ok, let's collect those ideas in the thread started by rbowen 15:23:57 We're looking at outages from ci.centos, review.r.o, rcip-dev and trystack in the near future 15:24:28 anything else on this topic? 15:24:30 let's move to that as a next topic? 15:24:31 apevec: they want to do trystack before summit last I heard from wfoster 15:24:53 #topic Announcements 15:24:54 There was a mention of that on the Facebook group. I was just looking for the date ... 15:25:10 They'll have a parallel cloud up soon (i.e, they're not "upgrading" afaict) 15:25:12 It says "mid september" 15:25:22 "We'll be planning an upgrade to the Mitaka OpenStack release in mid-September" 15:25:37 I think one could possibly say we're still in mid-September. 15:25:40 dmsimard: we'll be deploying a separate stack for trystack and then migrating the older compute nodes into it later (we actually need to install 10GbE NICs in the older gear because we have a 10GbE switch for backend traffic now). 15:25:56 wfoster++ 15:25:56 dmsimard: Karma for wfoster changed to 1 (for the f24 release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:26:07 so far as you guys are concerned you shouldn't be impacted until we migrate rest of existing stuff back in 15:26:29 and we'll be giving you plenty of notice 15:26:40 Yeah I think we'll be fine on that front 15:26:59 just not at the same time w/ rcip outage 15:27:19 It's also worth mentioning that the rcip-dev cloud (that hosts review.r.o) will have a period of downtime at some point, still not clear on the date but I've asked for >= oct 15 at the very least 15:27:39 Turns out CentOS are not the only ones having to physically move servers around 15:28:17 all the community cage is moving, or the whole RDU datacenter? 15:28:37 number80: iiuc rcip-dev is actually hosted in enovance premises in Paris 15:28:38 rcip is in Paris? 15:28:53 rcip is supposed to be in Montréal 15:28:57 -dev 15:28:59 nope 15:29:06 rcip is and is not 15:29:11 * jruzicka enters zen mode 15:29:43 is there anything else on the topic? 15:29:58 * number80 not convinced 15:30:16 number80: nothing to be convinced about, it's a fact :p 15:30:37 dmsimard: it's a fact, it's not in Paris ... 15:30:49 lol 15:30:59 ok, next topic 15:31:00 we can take details about infra outages on the list/irc 15:31:04 ack 15:31:16 #topic Chair for next meeting 15:31:24 i can take it 15:31:24 We should have an infra list by tomorrow ... :-) 15:31:34 cool thanks amoralej ! 15:31:35 rbowen, NOT for discussions! 15:31:43 heh. Right. 15:31:49 #topic Open Floor 15:32:36 we still have 28 minutes, we can finish the meeting or keep going on this topic 15:32:36 thanks rbowen 15:32:53 about infra, there was some discussion about disaster recovery testing for review.r.o some time ago 15:33:01 please help us checking the inventory 15:33:04 Oh. It's already been created. Awesome. 15:33:04 did we test it? 15:33:12 Info will be sent to rdo-list momentarily. 15:33:56 amoralej, not that I know, SF team took it into their backlog iirc 15:36:25 open floor. we need a file server :) 15:37:03 hey it's not xmas yet! 15:37:12 hehehe 15:37:27 weshay: is swift server not good enough? 15:37:31 but I heard St Wfoster has one? 15:37:47 number80, adarazs tried it, reported it wasn't 15:38:09 ack 15:38:19 weshay, number80: wait, it's actually okay, I just didn't find os-loganalyze 15:38:22 we have a Netapp in RDU2 but it's not wired up yet. That will happen after Summit when we can get 10GbE nics installed in the older machines and pull together the current Liberty + new, parallel RDO Mitaka environment. 15:38:40 adarazs: maybe it's worth pinging tristanC about it 15:38:41 number80, weshay: I just added this comment: https://trello.com/c/yv58BCKt/170-log-server-avail-over-http-for-3rd-party-openstack-ci#comment-57e2a809ea9789881c9168fa 15:38:51 into one big, sweet, backend 10GbE networked super trystack (with hopefully openstackid auth) 15:39:01 adarazs, thanks.. 15:39:03 wfoster++ 15:39:04 weshay: Karma for wfoster changed to 2 (for the f24 release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:39:05 number80: I pretty much figured it out, works like a charm. 15:39:07 rbowen: wow that was fast 15:39:16 wfoster++ no more FB! 15:39:17 apevec: Karma for wfoster changed to 3 (for the f24 release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:39:48 number80: we just need an instance on the same cloud where the swift is to host the os-loganalyze, update our collect logs playbook to upload to swift and we're good. 15:39:51 weshay: ^ 15:40:09 which cloud is that on? rcip? 15:40:20 dmsimard: Yeah, there's a form to fill out, and it's all automated, apparently. 15:40:23 weshay: yes. 15:40:33 The form says 3 days to get it created, but it was instant. 15:40:37 adarazs: yeah I got the email, I'll get you a VM - it won't be beefy but should be enough 15:40:39 number80, weshay: and I'm talking about the log aspect now. :) 15:41:06 wfoster++ 15:41:07 jruzicka: Karma for wfoster changed to 4 (for the f24 release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:41:25 dmsimard: I don't think it needs to be too beefy, it needs to run apache with a wsgi and that's it. it needs a floating ip too :) 15:41:36 adarazs: well I mean, we're talking about *apache* here :P 15:42:05 dmsimard: but probably very few requests per day :) 15:44:07 I think we're done here 15:44:13 yup 15:44:23 everyone gets 15min life back 15:44:28 high speed! 15:44:30 hehehe nice! 15:44:40 ok, i'll end the meeting... 15:44:52 apevec: re: PBR I think the only thing to watch out for will be packstack spec file 15:45:01 #endmeeting