15:00:53 #startmeeting RDO meeting - 2017-01-18 15:00:53 Meeting started Wed Jan 18 15:00:53 2017 UTC. The chair is jpena. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:53 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:53 The meeting name has been set to 'rdo_meeting_-_2017-01-18' 15:00:54 Meeting started Wed Jan 18 15:00:53 2017 UTC and is due to finish in 60 minutes. The chair is jpena. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:55 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:58 The meeting name has been set to 'rdo_meeting___2017_01_18' 15:01:11 remember to add topics to the agenda: https://etherpad.openstack.org/p/RDO-Meeting 15:01:16 #topic roll call 15:01:45 o/ at least partially 15:01:48 #chair apevec rbowen 15:01:48 Current chairs: apevec jpena rbowen 15:01:48 o/ 15:01:53 \o I'm here but conflicting with another meeting 15:01:57 I can't figure how it's a week since the last one. 15:01:59 #chair dmsimard 15:01:59 Current chairs: apevec dmsimard jpena rbowen 15:02:38 o/ 15:02:47 o/ 15:02:55 #chair trown number80 15:02:55 Current chairs: apevec dmsimard jpena number80 rbowen trown 15:03:15 \o 15:03:29 #chair jruzicka 15:03:29 Current chairs: apevec dmsimard jpena jruzicka number80 rbowen trown 15:03:44 let's go! 15:03:45 #topic Ceph Hammer (EOL) with Mitaka 15:04:01 who added this? 15:04:02 that's mine 15:04:30 Ceph Hammer is EOL'd since November (although "actual date" has not been updated yet in http://docs.ceph.com/docs/master/releases/ ) 15:05:15 I just wanted to agree and formalize whether or not we're going to be officially shipping or testing jewel against Mitaka 15:05:23 yeah, iirc ktdreyer said there will be one more EOL release? 15:05:40 dmsimard, we are not shipping it, storage SIG is 15:05:44 Jewel is known to work with Mitaka but we are not testing it anywhere - not in puppet-openstack-integration, not in TripleO, or anything 15:06:01 so it's their decision 15:06:20 but we could get involved, there's cry-for-help on centos-devel 15:06:26 o/ 15:07:01 \o/ 15:07:19 dmsimard, re. testing - I think upstream puppet ci is not ready to move to Ceph Jewel 15:07:20 So, we're not going to be testing jewel on Mitaka in our various CI then - just wanted to make that clear. 15:07:25 mwhahaha, EmilienM: can you ack ^ ? 15:07:27 and tripleoci probably the same 15:07:42 yeah, we want their acks 15:07:46 right 15:07:57 Mitaka puppet CI will never test Jewel 15:08:12 alright then it's clear 15:08:12 trown|mtg, weshay: can you ack ^ 15:08:45 ya, dont see why we would do that at this point 15:09:07 ok 15:09:54 #agreed While Jewel is known to work with RDO Mitaka, RDO does not currently test it and has no plans on testing it until Mitaka is EOL 15:09:57 fair statement? ^ 15:10:09 er, hang on 15:10:10 we just need to "fix" storage sig to keep hammer around 15:10:11 #undo 15:10:11 Removing item from minutes: AGREED by dmsimard at 15:09:54 : While Jewel is known to work with RDO Mitaka, RDO does not currently test it and has no plans on testing it until Mitaka is EOL 15:10:18 ya, that reads to me that we will test it after Mitaka is EOL :P 15:10:55 o/ 15:11:14 +1 15:11:25 #agreed Ceph Hammer is currently what RDO tests with the Mitaka release. Ceph Hammer is all but EOL. While Jewel is known to work with RDO Mitaka, RDO does not currently test it and has no plans on testing it until Mitaka is EOL.à 15:11:27 damnit 15:11:30 #undo 15:11:38 #agreed Ceph Hammer is currently what RDO tests with the Mitaka release. Ceph Hammer is all but EOL. While Jewel is known to work with RDO Mitaka, RDO does not currently test it and has no plans on testing it until Mitaka is EOL. 15:11:38 3rd lucky 15:11:38 haha 15:11:58 ok we can move on 15:12:01 hold on 15:12:18 we need action to work with storage sig to ensure ceph hammer is available 15:12:31 also updated to that latest EOL release when it happens upstream 15:12:40 We're not going to get that *here* 15:12:40 ktdreyer, ^ is that still planned? 15:13:44 o/ 15:16:52 true, but since RDO has vested interest, we need to have some level of involvement 15:16:52 I'll take action to coordinate that 15:16:52 #action apevec to check Ceph Hammer repo w/ CentOS Storage SIG 15:17:09 should we move on? 15:17:11 yes 15:17:13 next! 15:17:22 #topic Any feedback from tripleo/puppet CI on deps udpates in openstack-common-pending? 15:17:48 apevec, is that yours? 15:17:55 oh yes 15:18:15 so we need to move on with at least OVS 2.6.1 and mariadb 10.1.20 updates 15:18:24 they are currently in common-pending repo 15:18:31 and we can run weirdo jobs on them 15:18:47 but we don't have feedback from puppet and tripleo CI if they are safe 15:19:01 EmilienM, mwhahaha ^ how can we test those pending updates? 15:19:23 we can have a poi job 15:19:31 next week is Ocata freeze, we should get them released 15:19:31 a periodic maybe? 15:19:55 EmilienM, can periodic be hacked to run with this additional repo? 15:20:00 everything is possible in poi :D 15:20:26 what about tripleoci ? 15:20:51 apevec: same 15:20:55 hack something in tripleo-ci repo 15:20:59 apevec, are we shipping libvirt 2.5.0 ? 15:21:12 ah no, that was experimental build, I'll remove it 15:21:22 we use base EL7 libvirt 15:21:25 libvirt-daemon-kvm-2.0.0-10.el7_3.2.x86_64 15:21:32 amoralej: that's what I have 15:21:36 yeah that is base EL7 15:21:47 I'll untag 2.5.0 from pending 15:21:49 there is a new 2.5 in -pending, that's why i asked EmilienM 15:21:51 yeah 15:21:57 oh ok 15:22:04 we can easily test pending stuff with poi 15:22:11 we could do it in RDO CI though 15:22:28 ok, so who can hack in tripleoci to add common-pending? 15:22:31 dmsimard already did this process for our promotion pipeline 15:22:37 apevec: yes kindof 15:22:55 yeah, weirdo has passed w/ pending 15:22:59 dmsimard, ^ right? 15:23:15 actually that was before mariadb update, I tagged 10.1.20 today 15:23:30 weirdo tested ovs 2.6.1 successfully 15:23:36 the link to the job is somewhere in that thread 15:23:38 ship it then 15:23:46 dmsimard, I'll untag libvirt 2.5.0 then ask you to rerun it 15:23:47 there was one failure and it was unrelated, some cinder volume thing iirc 15:23:48 I am not sure how relevant the weirdo jobs are for that repo though... seems like every time we push changes there tripleo gets broken 15:23:54 mathieu bultel created rdoinfo: Enable ansible-pacemaker module package https://review.rdoproject.org/r/4351 15:24:05 trown|mtg, that's why I'm asking for tripleoci feedback w/ pending 15:24:13 trown|mtg: I agree, that's why I've asked for help on testing tripleo with it multiple times :( 15:24:21 trown|mtg: ovs in poi or ovs in tripleo would have same results I think 15:24:46 Also, I'm going to sound like a broken record and repeat myself but installing from scratch is one thing but upgrading is another thing entirely and worries me 15:24:49 dmsimard: well, i think it's a 5 lines patch in tripleo-ci 15:25:16 hmm, ya I didnt think about testing it with a tripleo-ci review 15:25:22 I can take an action for that 15:25:35 #action trown submit a review to tripleo-ci to test pending repo 15:26:07 excellent, thanks! 15:26:29 I'll ping you when pending is cleaned of experimental builds 15:26:29 we do need a longer term solution that is easier to integrate into some automation for changes to that repo, but that is the part I don't have cycles to personally work on at the moment 15:26:55 trown|mtg, yes. We need something ad-hoc to unblock current updates 15:27:12 we'll be adding automation for deps updates, details TBD 15:27:23 but number80 is already on it :) 15:28:35 next? 15:28:51 yes 15:29:12 #topic Still looking for talks for CentOS Dojo in Brussels 15:29:14 If 1) you're going to be in Brussels on Friday, February 3rd and 2) you like talking, the CentOS dojo (link in agenda) is still a little short on content. 15:29:48 Talks should be CentOS-centric, rather than OpenStack centric, but that still leaves a lot of topics. 15:30:02 Contact me if you have something in mind. 15:30:05 /EOL 15:31:13 #topic Mitaka: Should CI be using KVM from virt sig (>=2.6) ? 15:31:31 I just came across this so I sneaked it in 15:31:35 yes, are we not? 15:31:43 Newton was the first release to *enforce* KVM >=2.6 15:31:52 virt sig was not bundled in rdo-release 15:31:53 yes, we should. Anything else is asking for trouble with CentOS 7.3 15:31:54 for mitaka 15:32:14 ok, I'll update weirdo to use it for Mitaka. 15:32:25 it was using base OS 15:32:35 qemu-kvm-1.5.3-126.el7.x86_64 15:32:39 I can prepare a PR for the rdo-release rpm 15:33:06 Implying users might have to ugprade from <2 to >=2.6 15:33:12 I don't know 15:33:54 maybe we can communicate on the different MLs that users are encouraged to use >=2.6 with mitaka on 7.3 15:34:02 without necessarily enforcing it 15:34:26 [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-master-current @ http://tinyurl.com/zqowcwb |#| Build failure on centos7-master/current: tripleo-ui: http://trunk.rdoproject.org/centos7-master/report.html 15:35:01 I have nightmares about upgrading qemu-kvm in production but that was from experience a while back from icehouse to juno on Ubuntu :) 15:35:11 yeah, let's not change rdo-release-mitaka at this point 15:36:21 ok, that's it for me 15:36:41 #action dmsimard to update weirdo to use KVM from virt sig in Mitaka jobs 15:36:52 #topic chair for next meeting 15:36:57 any volunteer? 15:37:03 off topic but gfidente can you join #centos-devel to chat about Ceph storage SIG involvment 15:37:09 chandankumar: 15:37:17 i will volunteer for the same 15:37:27 thanks chandankumar :) 15:37:38 #action chandankumar to chair the next meeting 15:37:41 #topic open floor 15:37:41 dmsimard ack thanks 15:37:53 Jakub Ruzicka proposed rdopkg: patch: refactor to properly handle all cases https://review.rdoproject.org/r/4291 15:38:01 need eyes on that ^ 15:38:18 abregman|mtg, it's now noninteractive per your request ^ 15:39:01 much needed `rdopkg patch` fix including unit tests 15:39:59 * number80 is sorry looking for a lost ISP technician and concierge 15:41:26 Ken Dreyer proposed rdopkg: guess: handle "VX.Y.Z" Git tags https://review.rdoproject.org/r/4345 15:42:31 if there are no more topics, we can close the meeting (15 minutes more life!) 15:43:26 3 15:43:27 2 15:43:27 1 15:43:34 #endmeeting