15:00:38 #startmeeting RDO meeting (2016-02-03) 15:00:38 Meeting started Wed Feb 3 15:00:38 2016 UTC. The chair is jpena. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:38 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:38 The meeting name has been set to 'rdo_meeting_(2016-02-03)' 15:00:42 o/ 15:00:48 o/ 15:01:00 #chair jschlueter elmiko trown 15:01:00 Current chairs: elmiko jpena jschlueter trown 15:01:04 #topic roll call 15:01:26 o/ 15:01:30 o/ 15:01:38 o/ 15:01:41 #chair imcsk8 15:01:41 Current chairs: elmiko imcsk8 jpena jschlueter trown 15:02:06 o/ 15:02:15 #chair mflobo 15:02:15 Current chairs: elmiko imcsk8 jpena jschlueter mflobo trown 15:02:25 #chair chandankumar 15:02:28 #chair chandankumar 15:02:28 Current chairs: chandankumar elmiko imcsk8 jpena jschlueter mflobo trown 15:02:51 apevec, number80, you around? 15:04:02 ok, we'll start without them and they'll catch up later 15:04:08 #topic Packaging Delorean 15:04:11 o/ 15:04:15 #chair apevec 15:04:15 Current chairs: apevec chandankumar elmiko imcsk8 jpena jschlueter mflobo trown 15:04:19 just in time :) 15:04:36 yes, it's time to package delorean! 15:04:58 so we are planning to create a Delorean package. It doesn't make a lot of sense not to package a packaging tool 15:04:59 we just need to declare what do we want to have in 0.1 or 1.0 15:05:22 until now Delorean didn't have any release yet! 15:05:35 I put https://review.gerrithub.io/#/c/254505 there, but I would not block release on it. it is just a feature I have wanted since starting to work on delorean 15:05:36 so maybe 0.1 asap 15:05:45 +1 to 0.1 15:05:54 also +1 to 0.1 15:05:57 +1 15:06:09 trown, I have that review on my list 15:06:20 ok so 0.1 it will be 15:06:30 apevec: still have to solve jpena's concern 15:06:33 there's one more feature open from flepied 15:06:33 #agreed The first Delorean release will be 0.1 15:06:55 what is jpena's concern? 15:07:13 gnocchi package is actually a good example of no explicit epoch, but requires on subpackages 15:07:13 I'm doing some test package, so far I've found a couple issues, one of them specifically with python-sh 15:07:26 ah yes, we need patched sh 15:07:46 unless upstream merged Derek's patch 15:08:13 apevec: it's a different issue. We have Derek's patch, and the fact that running sh.env() works in sh 1.11 (from pypi) but not in 1.08 (from Fedora package) 15:08:32 I'm preparing a workaround for delorean, so it works on both 15:08:32 .whoowns python-sh 15:08:32 apevec: grover 15:08:52 btw, do we expect to support Delorean running on CentOS? 15:09:09 jpena, let's talk to python-sh maintainers 15:09:21 apevec: ack 15:09:24 ralph from fedora infra is co-maint 15:10:15 jpena, yes, delorean should run on centos 15:10:34 what's breaking on centos? 15:10:48 I'd expect the python-sh package is not any newer 15:11:15 it's the same 1.08 in epel7 15:11:34 but I don't see why maintainers wouldn't want to update it 15:11:45 it's just been neglectware 15:11:59 ok then 15:12:13 #action jpena to follow-up with python-sh maintainers on version bump to 1.11 15:12:26 jpena, best to prepare patch and put it in bz, to make it easy for them :) 15:12:41 or even ask for commit in pkgdb 15:12:59 mmm, that might work 15:13:01 #undo 15:13:01 Removing item from minutes: ACTION by jpena at 15:12:13 : jpena to follow-up with python-sh maintainers on version bump to 1.11 15:13:25 #action jpena to make sure python-sh is bumped to 1.11 and it includes Derek's patch 15:13:55 #info Derek's sh patch == https://github.com/amoffat/sh/pull/237 15:13:59 ^ not merged yet 15:15:19 ok, I'll try to have a version to test later this week. Anything else to discuss here? 15:16:33 nice 15:17:23 ok, for 0.1 we need topic:delorean-packaging 15:17:32 Epoch: 15:17:50 and Fred's --recheck and --status 15:18:15 and yes, rdoinfo tags support from me (to be sent today) 15:18:24 anything else? 15:18:52 there are few older patches, but all need to be rebased so I'm ignoring them for now 15:19:15 oh ya, recheck would be great 15:20:01 I think that's all 15:20:33 moving on? 15:22:01 yes 15:22:06 #topic Migrate the f22 Delorean worker to f23 15:22:35 So we spoke about it last week I think. I've started tests on a separate VM, it looks good so far 15:22:50 http://209.132.178.4/f23/status_report.html 15:22:56 host is still f22 ? 15:23:09 yes, still f22 15:23:29 I'm running delorean locally on my laptop with f23 and it goes just fine anyway 15:23:31 ok, let's plan c7 for the next rebuild 15:23:56 btw, what would be good timing for Delorean rebuild, using puppet-delorean? 15:24:21 we did the last a week after summit (it's usually a quiet time) 15:24:35 hmm... we have a window nowish, but after that, I agree after summit 15:25:08 yeah, after summit looks good 15:25:15 actually nowish is bad if something goes wrong since so much upstream is relying on delorean 15:25:48 yes, now is probably not a good idea 15:26:00 I was thinking just in terms of RDO test days 15:26:00 actually, I'm not going to Austin so we could do it *during* summit :) 15:26:39 f22 eol is few weeks after f24 ga in June, so we're good to stay on it for a while 15:26:47 for host 15:26:59 I'd move replace f22 worker w/ f23 asap 15:27:27 jpena, during summit is even better, provided you keep old running 15:27:30 right, so once I finish all my checks, the next step will be to stop the f22 worker and create the f23 one in the Delorean instance 15:27:32 then just switch IP 15:28:02 apevec: yes, that's always the plan, only switch IP once the new one is ready 15:28:04 jpena, in anycase, time to trello it, with a checklist 15:28:20 ack 15:28:30 I have a delorean instance running on centos if you need it 15:28:54 flepied: cool! Any gotchas? 15:28:55 jpena, that's a good point ^ 15:29:18 jpena, maybe we should consider moving to the minicloud where rpmfactory is running? 15:29:28 flepied, is that where you have your Delorean? 15:30:20 apevec: if we have resources, sure! 15:30:25 apevec: yes http://46.231.133.253/delorean/status_report.html 15:31:38 flepied, can you give access to jpena for testing there? 15:32:15 apevec: yes but we need to coordinate as I'm experimeting patches there 15:32:40 flepied, I mean cloud access, so he can setup new production delorean 15:33:28 is the "minicloud" able to handle the traffic load that our current delorean gets? 15:33:44 that's what we need to test :) 15:34:22 apevec: yes it can be done. I have an ansible script to reprovision a delorean instance when needed if you want. 15:34:26 flepied, what size of internet pipe do those machines have? 15:34:47 flepied, heh, and jpena just finished puppet-delorean ;) 15:34:58 cfg warz... 15:35:29 apevec: I don't know the details but it's a good network connection. I can get the details if you need. 15:36:07 apevec: what do we do with the bootstraping issue? 15:37:13 flepied, yeah, sphinx/debcollector... do we have more or only that one? 15:37:34 apevec: I don't think so 15:38:10 flepied, what about providing explicitly a list of packages which need to be built w/ bootstrap macro enabled? 15:38:24 to avoid parsing spec 15:39:15 btw tl;dr for others: we have build dep loop 15:39:16 apevec: the problem like I said is to compute the right order we need to parse the spec files 15:39:57 flepied, right and you ignore those explicitly specified and build dep first 15:40:02 s/dep/them/ ! 15:40:13 e.g. python-oslo-sphxin with bootstrap 1 15:40:29 after that you can compute order from the rest of the specs 15:40:54 apevec: ok we can try something like that 15:41:23 so this would one more Delorean fix before tagging 0.1 ? 15:42:00 apevec: oh no I don't know. what's the target of 0.1? 15:42:20 we didn't set ETA but soon 15:42:36 I mean what's the objective? 15:42:39 it was first topic today, we identified few patches to merge 15:42:44 flepied, to get it packaged 15:43:20 ok no needed to wait for this cycle issue if you need a package soon 15:43:34 ok 15:43:53 but knowing you, there'll be patch review ready in no time ;) 15:44:10 jpena, sorry we went OT ! 15:44:27 np, but we should move on. We have more topics and only 15 minutes to go 15:44:27 put us back on track 15:44:41 #topic Cleanup trello 15:44:52 not sure who added this to the agenda 15:44:59 that was number80's but he's on the bus to Brno 15:45:00 number80: ^ 15:45:01 number80 has been rocking that 15:45:12 ack to all what he proposed: 15:45:17 obsolete cards => delete 15:45:21 cards without feedback => marked for deletion 15:45:25 merge redundant cards 15:45:57 ack. Who's going to do the effort? 15:46:01 and yes, number80 already started pruning, 15:46:04 I'll join that 15:46:16 #action apevec to join trello prunning 15:46:29 btw if you want the trello bot for this chan, EmilienM is running one for freenode 15:46:31 #undo 15:46:31 Removing item from minutes: ACTION by apevec at 15:46:16 : apevec to join trello prunning 15:46:36 #action apevec to join RDO trello pruning 15:46:49 what is that bot doing? 15:46:59 telling all actions on a trello board 15:47:10 go on #rdo-puppet, you'll see 15:47:16 apevec: abstract: http://paste.openstack.org/show/485861/ 15:47:17 ok, I'll have a look 15:47:24 every action on a board is sent on IRC 15:47:30 "bob commented on XX" 15:47:36 "dan added bob to XX" 15:47:37 etc 15:47:42 hmm, that might be a bit too noisy 15:47:52 yeah but I'm sure we can adjust it 15:47:54 not sure our trello is all that noisy :) 15:47:56 but lemme watch it few days in #rdo-puppet to make up my mind 15:48:17 it's usually not that much, although there are some noisy moments 15:48:34 I'm sure we can add filters to the code 15:48:42 so we can only see new/inprogress/close things 15:48:56 ok, let's put that on the agenda next week 15:48:56 that'd be great, I think 15:49:44 ok, so let's move on 15:49:52 #topic kilo 2015.1.3 rebase 15:50:18 I think number80 added this one, too 15:50:20 that's also number80 - afaik he did it and put on testing 15:50:55 #action apevec send call-for-testing 2015.1.3 on rdo-list 15:51:20 only missing horizon for the rebase but needed to check w/ mrunge 15:51:43 #action apevec to ping mrunge about 2015.1.3 rebase 15:51:46 number80, sorry, I what rebase? 15:51:51 number80, or you grab him in Brno 15:51:56 ah there he is 15:51:56 ack 15:52:00 * mrunge runs 15:52:13 :P 15:52:35 mrunge, 2015.1.3 i.e. f23 branch for now 15:52:39 (Kilo) 15:52:42 2015.1.3 rdo? 15:52:46 yes 15:52:55 #undo 15:52:55 Removing item from minutes: ACTION by apevec at 15:51:43 : apevec to ping mrunge about 2015.1.3 rebase 15:53:03 sigh 15:53:45 kilo per se does not support django-1.8 15:53:48 number80, anything else for 2015.1.3 ? 15:54:22 mrunge, didn't you patch it in rdo? 15:54:50 yes, I did that, but honestly, I don't know the state in rdo 15:55:18 rdo kilo is still f23 branch distgit 15:55:38 oh, ack. makes things way easier 15:55:39 we just rebuild it from there in CBS 15:55:44 I can't 15:55:56 but I'll ping number80 15:56:10 yep 15:56:30 Hey folks, wondering if anyone knows how to determine which controller node is acting as the VRRP master in a 3 controller HA config 15:56:47 mrunge, btw, you can apply for CBS account now - ask number80 for details 15:56:53 cdearborn: please wait a few minutes, we are in the middle of a meeting 15:57:07 there's accounts.centos.org 15:57:07 apevec, don't really know if I should 15:57:13 mrunge: we can sit down or do it half way, rebase the patch branch for me 15:57:13 mrunge, up to you 15:57:23 cdearborn, the one who pings instances from router 15:57:25 (and i'll finish) 15:57:43 ack number80 , let's talk about this in Brno 15:58:03 yup, no pressure :) 15:58:40 ok, done w/ rebase 15:58:49 ? 15:59:06 #topic open floor 15:59:18 Any volunteer to chair the next meeting? 15:59:24 flepied, re. your ansible for delorean - please check https://github.com/javierpena/puppet-delorean 15:59:48 apevec: ok but I'm not a puppet guy ;-) 15:59:51 flepied, do you have your playbook published ? 15:59:52 flepied would be perfect as a chair 15:59:58 * number80 runs 16:00:11 jpena, flepied, let's decide which one to continue using 16:00:20 maintaining both is probably a waste 16:00:32 number80: I'm chairing enough meetings, thx but no. 16:00:39 :) 16:00:41 :) 16:01:02 if no volunteer, I'll take it 16:01:19 ack, i'll take the following week 16:01:19 apevec, i can volunteer for next meeting 16:01:41 chandankumar, you get the 3rd :) 16:01:51 thanks everyone! 16:01:51 ack 16:01:55 o/ 16:02:07 if there's nothing else to discuss, we're out of time... 16:02:14 3 16:02:15 2 16:02:22 1... 16:02:25 #endmeeting