14:00:03 #startmeeting tripleo 14:00:04 Meeting started Tue Aug 16 14:00:03 2016 UTC and is due to finish in 60 minutes. The chair is EmilienM. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:05 bandini, hold on, he hasn't called roll yet :0 14:00:06 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:09 The meeting name has been set to 'tripleo' 14:00:09 #topic rollcall 14:00:11 o/ 14:00:13 o/ 14:00:14 o/ 14:00:15 hey 14:00:16 o/ 14:00:17 o/ 14:00:18 hello 14:00:18 hi 14:00:19 o/ 14:00:20 o/ 14:00:29 o/ 14:00:29 o/ 14:00:34 beagles: :) ops 14:00:38 hello again :P o/ 14:00:54 #topic agenda 14:00:58 o/ 14:01:03 * one off agenda items 14:01:18 * bugs 14:01:20 * Projects releases or stable backports 14:01:22 * CI 14:01:24 * Specs 14:01:26 * open discussion 14:01:29 o/ 14:01:39 #topic one off agenda items 14:01:44 #link https://etherpad.openstack.org/p/tripleo-meeting-items 14:01:44 o/ 14:01:49 d0ugal: hey, go ahead! 14:01:51 Hi 14:01:57 EmilienM: Hey, thanks 14:01:58 hi 14:02:14 So I just wanted to give everyone a bit of an update on the status of the transition to Mistral for deployment. 14:02:30 With the CLI patch here: https://review.openstack.org/#/c/342261/ 14:02:44 and the tripleo common series here: https://review.openstack.org/#/c/298732/ 14:02:44 o/ 14:02:47 We have it working! 14:02:56 d0ugal: nice 14:03:00 So, really looking for feedback on the approach, testing and reviewers. 14:03:00 woo 14:03:02 super cool 14:03:13 nice 14:03:20 d0ugal: that's a great news! thanks for all this work 14:03:26 The tripleo-common patches are fairly low risk - nothing will use them until the CLI changes lands. So I'd like to see that happen soon 14:03:35 (Then the CLI patch can actually test against CI properly) 14:03:52 If anyone wants more info to test or try things out, please ping me 14:03:56 great, so the patch in depends-on can be reviewed with low risk of breaking something 14:03:58 ... and that is all I've got. thanks :) 14:04:06 EmilienM: Yup, exactly. 14:04:07 o/ 14:04:19 o/ 14:04:27 #action tripleo team to review https://review.openstack.org/#/c/298732/ and https://review.openstack.org/#/c/342261/ 14:04:46 thanks d0ugal added to review queue fwiw 14:04:48 d0ugal: thx for the update! don't hesitate to use ML if you need more reviews or have more updates! 14:05:02 marios: thanks! 14:05:07 EmilienM: will do :) 14:05:22 #topic bugs 14:05:25 #link https://bugs.launchpad.net/tripleo/?orderby=-id&start=0 14:05:43 #link https://bugs.launchpad.net/tripleo/+milestone/newton-3 14:05:57 o/ 14:06:13 o/ 14:06:32 if you're working on a bug listed for newton-3, please make sure it's up to date 14:07:24 so, outstanding bugs :) 14:07:34 o/ 14:07:45 the one I can see is https://bugs.launchpad.net/tripleo/+bug/1613088 14:07:45 Launchpad bug 1613088 in tripleo "tripleo-ci: periodic nonha job failing to promote due to SSL" [Critical,In progress] - Assigned to Emilien Macchi (emilienm) 14:07:56 but it's under control afict 14:08:08 do we have other bugs to mention this week? 14:09:11 https://bugs.launchpad.net/tripleo/+bug/1613630 will be solved when our CI will be pin 14:09:11 Launchpad bug 1613630 in tripleo "CI: current-tripleo repo contains broken cinder and volume creation fails" [Critical,In progress] - Assigned to Sagi (Sergey) Shnaidman (sshnaidm) 14:09:28 here's the pin: https://review.openstack.org/#/c/354481/ 14:09:37 EmilienM, it's in progress, working with jpena from #rdo on this 14:09:51 sshnaidm: is there update on getting the repo switched back? 14:10:06 slagle, not yet, jpena is checking right now 14:10:41 might be a problem related to manual change of link yesterday 14:10:46 the patch is in gate now, hopefully it will merge in the next hour 14:11:06 do we have anything else about bugs this week? 14:11:19 please raise anything outstanding 14:12:09 ok, next topic 14:12:12 #topic Projects releases or stable backports 14:13:03 I haven't checked if shardy actually propose stable branch releases last week 14:13:20 #action EmilienM to check & release stable branch this week if needed 14:13:27 any question about releases? 14:13:44 o/ 14:13:57 ansiwen: what's up? 14:14:18 oh, I thought that's the way to say hello, sorry :-) 14:14:26 ok next topic 14:14:28 #topic CI 14:14:32 * ansiwen was late 14:14:59 derekh: I noticed you're around, do you have any update this week? 14:17:03 regarding upgrade/update jobs, I have no update, still working with matbu but really slow progress with all recent outages... 14:17:57 maybe you'll find it useful: https://etherpad.openstack.org/p/tripleo-ci-status 14:18:11 even when jobs are passing, are ha jobs are running close to the check timeout, and some are getting killed 14:18:31 * EmilienM bookmarks it 14:18:32 that will need to be investigated once the current issues are sorted 14:18:38 nice 14:18:59 slagle: we might reduce the number of services 14:19:15 slagle: ie: sahara, etc (even if I remember gfidente's concern about upgrades) 14:19:18 EmilienM, slagle ack, see dtantsur thread on subject on openstack-dev 14:19:26 EmilienM: yes i plan to do that in the multinode job at least 14:19:37 i think we can get it consistently under 60 minutes 14:19:40 EmilienM, yeah I was thinking about an experimental job deploying full overcloud as an alternative? 14:19:56 and trim down the gate job to a minimum instead 14:19:59 what do you mean by full? 14:19:59 slagle, for periodic jobs we added undercloud idempotent and pingtest deletion which take time 14:20:06 one with all services and that will timeout? 14:20:14 sshnaidm: this isn't for periodic though 14:20:15 EmilienM, yeah pretty much 14:20:20 gfidente: -1 14:20:35 EmilienM, eheh how do we test the optional services then? 14:20:36 experimental jobs are usually incubator jobs, I would rather see a new non voting job for now 14:20:39 sshnaidm: check-tripleo ha jobs are routinely taking 2h40m, or getting timed out 14:20:55 gfidente: by creating scenarios, like we did in puppet CI 14:21:03 they add more (not all) services 14:21:08 EmilienM: sorry, missed your ping, nothing to bring up at the moment, just back today and catching up 14:21:17 EmilienM, right but to do that we'd need a lot more resources 14:21:24 I know I pasted the link 1000 times but here's again https://github.com/openstack/puppet-openstack-integration#description 14:21:28 EmilienM, any permutation will run an entire undercloud/overcloud 14:21:33 gfidente: I kow 14:21:39 but the idea of experimental job won't help 14:21:54 well experimental can be triggered on request 14:22:00 1) it won't prevent breakages as it requires people to run "check experimental" 14:22:16 2) they'll timeout if you add all optional services 14:22:38 the idea of scenarios is to have multiple scenarios (1 with all telemetry for ex) 14:22:43 right so let's put it this way, with unlimited CI capacity, I agree permutations work better 14:23:02 derekh: ack 14:23:42 gfidente: one option is periodic job 14:23:51 scenarios in periodic 14:24:04 slagle, image build takes almost 45 mins, pretty long 14:24:14 so we cover everything, we don't timeout because we have multiple scenarios and we use periodic pipeline so don't consume too much resources 14:24:31 sshnaidm: yea, it takes half that time for me locally 14:24:34 maybe I can come up with an idea on the ML 14:24:53 EmilienM, yep that'd be nice 14:24:56 ok 14:24:59 we need similar support for ipv6 too 14:25:12 or for 'competing' services, as like ceph/rgw vs swift 14:25:13 how much effort is it to maintain the various scenarios? 14:25:17 #action EmilienM to propose periodic scenarios CI jobs testing on ML 14:25:41 beagles: unknown at this time, we don't have it, but shouldn't be that hard with our composable services 14:25:42 not that it's a reason not to do do it.. just thinking about delegation to devs, etc 14:25:48 good point 14:25:49 gfidente: exactly 14:26:04 anything else about CI this week? 14:26:53 #topic Specs 14:27:01 #link https://review.openstack.org/#/q/project:openstack/tripleo-specs+status:open 14:27:54 if you have a spec and need a review, please go ahead 14:28:05 it sounds like some of them have positive agreement 14:29:39 but some of them are still under discussion 14:29:48 please update / rebase your spec if you have one 14:30:22 any question about specs? 14:30:59 #topic open discussion 14:31:54 we're heading to end of newton-3, and final approach is in few days, please continue to review as you can 14:32:15 please let us know blockers and what can't be done for n3 14:32:21 hmm, when is the deadline for tripleo? the same as other projects? 14:32:25 * feature deadline ofc 14:32:31 marios, EmilienM is it fair to say that upgrade process from mitaka is still hackable after freeze? 14:32:42 Aug 29-02 -> N3 14:32:49 gfidente: yes 14:32:56 marios, EmilienM I think we consider upgrade a feature so it's fair game to improve it until release 14:32:59 so, 2 more weeks essentially... 14:33:05 we might want RC1 during Sep 12-16 14:33:22 yes, we have 2 more weeks than other projects 14:33:35 but in my experience, I like keeping short deadlines, for impediments :-) 14:33:49 I'm sure we'll have plenty of things to do during those 2 weeks 14:33:53 gfidente: EmilienM yeah i think we have always lagged on the upgrade after the release 14:34:00 our ultimate deadline is Oct 17-21 14:34:08 gfidente: EmilienM we have backport exception for upgrades related bugs right 14:34:16 marios: right 14:34:33 any help to bring upgrade / update CI jobs upstream would be great 14:34:56 ping me or matbu - we're looking at it recently, but 0 progress since outages 14:34:58 EmilienM: I can't stress enough the need for the mistral transition stuff to get eyes 14:35:03 and reviews 14:35:18 we need it to land for n-3 for the UI or we're in deep water 14:35:29 jrist: right 14:35:40 thanks 14:35:43 jrist: please dress a list and share it with our team 14:36:00 maybe a gerrit topic or LP # 14:36:03 EmilienM: my understanding is that asking for reviews on the ML is frowned upon? 14:36:18 I think listing project priorities is ok 14:36:19 jrist: what? 14:36:25 dtantsur: ok thanks 14:36:25 just appearing out of blue with your patches is not 14:36:28 dtantsur: got it 14:36:33 jrist: using ML for anything is IMHO good 14:36:37 ha 14:36:37 it keeps things open and async 14:36:38 ok 14:36:58 EmilienM, there were bit objects for using for 2 things: asking for reviewing and 3rd party CI status updates 14:37:00 same goes for the validations patches -- /me will email the ML 14:37:01 don't ask me why please :) 14:37:08 related, tripleo-ui launchpad would like to migrate to tripleO launchpad 14:37:10 * big objections 14:37:15 * dtantsur has no clue what he's typing 14:37:20 do I need special access for adding stuff? 14:37:25 anything else for this week ? 14:37:28 some sort of admin rights on tripleo launchpad? 14:37:41 jrist: for what? 14:37:47 I don't think so 14:37:53 * beagles imagines all of the frowning upon that he's been provoking lately 14:37:56 adding/changing/deleting launchpad items for tripleo-ui in the tripleo launchpad 14:37:56 let me know if something blocks you 14:37:59 beagles: heh 14:38:02 EmilienM: ok thanks 14:38:17 I'm closing the meeting in 30s, go ahead now if you have anything 14:38:33 * EmilienM preparing his notes for puppet meeting in 20 min 14:38:58 thanks everyone and keep rocking folks! 14:39:00 #endmeeting