14:01:15 #startmeeting tripleo 14:01:15 Meeting started Tue Jun 21 14:01:15 2016 UTC and is due to finish in 60 minutes. The chair is EmilienM. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:17 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:19 The meeting name has been set to 'tripleo' 14:01:20 thanks EmilienM 14:01:21 EmilienM: you run it 14:01:29 o/ 14:01:35 hey folks! 14:01:37 o/ 14:01:38 \o 14:01:38 hey 14:01:39 \o thanks EmilienM 14:01:41 hi 14:01:43 o/ 14:01:44 hi 14:01:48 hi! 14:01:54 #topic agenda 14:01:57 thanks EmilienM! 14:02:00 o/ 14:02:03 * bugs 14:02:05 * Projects releases or stable backports 14:02:07 * CI 14:02:09 * Specs 14:02:11 * open discussion 14:02:22 do we have other items for this week? 14:02:40 if no one-off items, let's move to bugs 14:03:13 #topic bugs 14:03:17 #link https://bugs.launchpad.net/tripleo/ 14:03:39 o/ 14:03:57 Does anyone have any bugs to mention this week in particular? 14:04:17 EmilienM: hey man, fyi/fwiw/as i mentinoed earlier, i am hitting a bunch of bugs for undercloud upgrade, like https://bugs.launchpad.net/tripleo/+bug/1593736 14:04:17 Launchpad bug 1593736 in tripleo "stable/mitaka to newton undercloud upgrade fails with 'Could not find class ::tripleo::selinux'" [Medium,Triaged] - Assigned to Marios Andreou (marios-b) 14:04:55 it sounds like a packaging issue in OPM 14:04:59 EmilienM: i think they are all package issues possible, though i see the fixes landed in master quite a while ago (different for each one, some few weeks old) 14:05:02 but I might be wrong 14:05:06 s/possible/possibly 14:05:23 marios: can you ping jayg (not here) and check if OPM/Mitaka is up to date? 14:05:29 I guess puppet-tripleo needs to be updated 14:05:45 EmilienM: ok i think i will file a tracking bug for each one i find (similar to that bug i pointed to, probably i have like 5 in total so far) 14:05:47 well 14:05:49 no 14:05:51 https://github.com/openstack/puppet-tripleo/tree/stable/mitaka/manifests 14:05:53 I don't see selinux.pp 14:05:58 so I guess something was not backported 14:06:21 https://github.com/openstack/puppet-tripleo/commit/8e533aaf447022c62865130f2ffc88690f06aef1 14:06:24 EmilienM: right so for some cases it may be backports and i should check mitaka 14:06:26 sounds like it was not backported ^ 14:06:35 also check with slagle, he did that patch 14:06:36 EmilienM: ack, just fyi cos is bugs /me end :) 14:06:40 to make sure we can backport it 14:07:04 dprince: do we have CI related bug? 14:07:14 everything is still super red on http://tripleo.org/cistatus.html 14:07:56 any other outstanding bug this week? 14:08:28 #action marios to investigate selinux backport to mitaka https://bugs.launchpad.net/tripleo/+bug/1593736 14:08:28 Launchpad bug 1593736 in tripleo "stable/mitaka to newton undercloud upgrade fails with 'Could not find class ::tripleo::selinux'" [Medium,Triaged] - Assigned to Marios Andreou (marios-b) 14:08:46 EmilienM, https://bugs.launchpad.net/tripleo/+bug/1594732 14:08:46 Launchpad bug 1594732 in tripleo "CI: No connected gearman servers" [Undecided,New] 14:09:12 EmilienM: ack thanks, i'll also file the rest of the issues individually for tracking after this like https://bugs.launchpad.net/tripleo/+bug/1593182 for example and a few more 14:09:12 Launchpad bug 1593182 in tripleo "failed openstack-nova-scheduler after updating undercloud from mitaka to newton packages" [Medium,Triaged] - Assigned to Marios Andreou (marios-b) 14:09:13 sshnaidm_: right, this one. Do we have any action on this one? 14:09:21 EmilienM, geard is restarted and running, also submitted a patch to renew it: https://review.openstack.org/#/c/332123/ 14:09:39 EmilienM, if it's installed from pip, it may to resolve the problem 14:09:47 #action team to review CI patch to deploy gear from pip https://review.openstack.org/#/c/332123/ 14:09:54 EmilienM, if not, will do something to restart it gracefully if possible 14:10:05 cool thx 14:10:09 can you assign the bug to you please? 14:10:14 EmilienM, yep 14:10:15 and tag it critical/confirmed 14:10:17 thx 14:10:22 auto triage FTW 14:10:27 anything else for bugs? 14:10:43 #topic Projects releases or stable backports 14:10:48 #link http://releases.openstack.org/newton/schedule.html 14:10:54 #link https://launchpad.net/tripleo/+milestone/newton-2 14:11:02 3 weeks until n-2 14:11:02 (I can't set its importance, only cores I think) 14:11:31 sshnaidm_: done. 14:11:33 sshnaidm_, done! 14:12:33 so 3 weeks before n-2, release is coming fast... Please take some time to do reviews on ongoing work for n2 14:13:01 If you think you'll miss the n-2 deadline, please move it to n3. 14:13:23 any question /blocker about release management? 14:14:19 great let's move to CI 14:14:21 EmilienM, nothing blocker as such but i had some queries related to releases of some components 14:14:21 #topic CI 14:14:35 maybe later 14:14:54 coolsvap: go ahead? 14:14:57 #undo 14:14:57 Removing item from minutes: 14:15:18 #topic Projects releases or stable backports 14:15:25 coolsvap: what's up? 14:15:31 I was talking with shardy when he trigrgerred ther release 14:15:35 for instack-undercloud 14:15:48 and asked me to look for some other projects which also need 14:16:17 i didn't get how to do it for say diskimage-builder 14:16:21 or instack 14:16:28 which are independant 14:16:33 independent* 14:16:47 coolsvap: https://wiki.openstack.org/wiki/TripleO/ReleaseManagement not sure how up to date that info is, haven't dont it in quite a while 14:17:01 coolsvap: https://github.com/openstack/releases/commit/85bfe0c0488a1dfdd046e14b4d27cdd1d2e4c929 14:17:02 coolsvap: i hink it is missing instack-* 14:17:07 look what I did for our first Newton release 14:17:29 you can check in https://github.com/openstack/releases/tree/master/deliverables/newton 14:17:37 https://github.com/openstack/releases/blob/master/deliverables/newton/instack.yaml 14:17:58 coolsvap: did you have a particular question? 14:18:03 coolsvap this was independent 14:18:03 https://review.openstack.org/#/c/303986/ 14:18:06 EmilienM ^^ 14:18:08 EmilienM, but instack doesnt have release in mitaka 14:18:23 how to handle that 14:18:29 coolsvap marios EmilienM we link to that from the ReleaseManagement wiki 14:18:37 coolsvap https://review.openstack.org/#/c/303986/ 14:18:49 https://github.com/openstack/releases/blob/master/deliverables/_independent/instack.yaml 14:18:52 gfidente, alright 14:19:10 you can add it in independent 14:19:12 that was my question whether it will remain in independent i have some doubts 14:19:18 yes it will 14:19:19 but i will ask on #tripleo 14:19:23 cool 14:19:31 anything else about releases? 14:19:45 #topic CI 14:20:07 regarding liberty CI, upgrade job is still failing 14:20:20 I investigated a bit and I found out it was related to ipv6 on upgrade job 14:20:37 I tried to run ipv4 on upgrade job for liberty: https://review.openstack.org/329664 but it still fails 14:20:40 I have no progress since then ^ 14:20:50 if anyone volunteers to help, let me know 14:21:15 dprince: can you give an update about CI (derek is offline) 14:21:34 slagle: also your status on multi nodes please 14:22:40 hi 14:22:56 i'm still working on it (and other things) 14:23:10 it's been slow going, due to the networking setup (vxlan tunnels) 14:23:28 maybe i'll have something up this week 14:23:39 slagle: do you have blocker that you need solved from our side ? (composable roles, etc) 14:23:42 i'll at least push some updated patches 14:23:49 EmilienM: not that i know of 14:24:06 we had a CI outage this weekend 14:24:20 slagle: ok last week you mentionned " which may be possible when composable nova-compute lands" 14:24:25 slagle: so I was wondering 14:24:38 yes, an aio overcloud may be possible 14:24:49 dprince: it's the same as we had this morning/night right? 14:24:55 i'll give it a try, but want to make sure i get the networking correct first 14:25:00 The compute host running our "mirror" server died. After rebooting it and cleaning up a bit things are moving again. Although I forgot to run os-collect-config on the compute host after reboot so we need to fix that eventually too 14:25:02 slagle: right 14:25:13 EmilienM: no, different issue. Today was the te-broker server which runs gearman 14:25:22 ok 14:25:44 anyway, I think things should be running fine now 14:26:05 dprince: yeah, will see if using pip to install gear can help 14:26:06 I will watch the gearman until sshnaidm_ bumps the version as discussed earlier 14:26:26 and if yes, I would suggest to reach maintainer and try to update it in centos7 14:26:50 anything else about CI? questions? info? 14:27:21 #topic specs 14:27:49 https://blueprints.launchpad.net/tripleo/+spec/tripleo-sriov 14:28:07 just wanted to confirm if the above spec is accepted for newton ? 14:28:22 karthiks: it sounds like you addressed most of comments 14:29:08 yes ... we have uploaded the changes . 14:29:25 We are working on the patches. 14:29:41 karthiks: I have serious doubt about making it on time for Newton 14:29:44 but I might be wrong 14:29:57 :) 14:30:11 we are hopefull :) 14:30:43 freeze is ~ August 30th 14:30:48 I think sriov could land in newton just fine 14:30:52 so you have ~ 2 months 14:31:01 if dprince is confident, cool then :-) 14:31:05 EmilienM: the idea with composable services is actually that we can land things faster hopefully 14:31:11 right 14:31:14 :) 14:31:15 EmilienM: it is only CI which holds up up 14:31:19 us up 14:31:30 do we need to talk about other blueprints? 14:31:34 i have uploaded a spec for iorc changes for kernel boot parameters #https://review.openstack.org/#/c/331564 14:31:37 last week we mentionned composable upgrades 14:31:39 *ironic 14:32:07 dprince: well, if they have some other specs in ironic, etc... we might not do it on time... 14:32:16 it is also related to SR-IOV and DPDK changes.. 14:33:09 no more spec to discuss? 14:33:35 #link https://review.openstack.org/#/q/project:openstack/tripleo-specs+status:open 14:33:59 marios, jistr: any update on https://review.openstack.org/#/c/319264/ ? 14:34:03 Composable Upgrades 14:34:29 this one is essential: https://blueprints.launchpad.net/tripleo/+spec/overcloud-upgrades 14:34:36 no update, other than the PoCs are being debated in terms of tool selection 14:34:54 debated in Gerrit? 14:34:55 I guess this would be a good place to bring it up, when do we want to meet again? 14:35:04 I don't see much discussion on https://review.openstack.org/#/c/319264/ 14:35:05 EmilienM: yeah, i think that will likely be updated after we finish the discussion about the overall direction 14:35:25 where is happenning the discussion? sorry for dumb question 14:35:35 EmilienM: we had a bluejeans call last week (I believe rhallisey extended the invitation to all during last week irc meeting) 14:35:40 #link https://etherpad.openstack.org/p/tripleo-composable-upgrades-discussion 14:35:54 ok excellent ^ 14:35:55 EmilienM: ^ 14:36:08 EmilienM: there will likely be another call this week, tbd afaik, possibly tomorrow shardy around? 14:36:16 are we going to reach a concensus by n-3 ? 14:36:24 EmilienM: hope so :) 14:36:27 shardy is travelling 14:36:29 EmilienM, ya I think so 14:36:34 EmilienM: so actually 14:36:35 great 14:36:38 EmilienM: i believe the conclusion was 14:36:52 EmilienM: yes it would be great to (and given we are talking about direction, not full working implementation) 14:36:59 EmilienM: but if mitaka..newton 14:37:13 EmilienM: sorry, but for mitaka... newton upgrades, we will use the existing workflow 14:37:27 bash scripts, right? 14:37:33 EmilienM: for which there is WIP 14:37:36 EmilienM: not quite 14:37:37 yea, and non-composable ones 14:37:41 ok 14:37:44 EmilienM: 3 heat stack updates 14:37:50 EmilienM: and some bash scripts :) 14:37:56 it makes sense, 14:38:03 EmilienM: what could possibly go wrong 14:38:04 so you're targetting this work for n..o 14:38:17 docs wip at https://review.openstack.org/#/c/308985/ 14:38:28 EmilienM: yes that was my understanding at least 14:38:43 cool 14:38:49 EmilienM: but anyway, at least the overall direction shouldb be decided in 2/3 weeks !! 14:39:05 "pray for upgrades" 14:39:05 marios, well depending on what we setting with and when, we could apply some of the work back for this 14:39:21 just a thought 14:39:33 may not be within the timeframe though 14:39:58 rhallisey: ack, just not concrete yet what we will do ... this is the blocker atm 14:40:13 do we have other spec to discuss today? I'll let 30 more sec and go to next topic 14:40:49 #topic Open Discussion 14:41:16 I was wondering where the use of the new pacemaker module in tht could fit in regarding specs ?(maybe Add initial next generation HA architecture spec) 14:41:48 chem: that's a question for bandini (not here) 14:42:01 oki 14:42:27 chem: honestly, I think you can consider the spec good for most of bits 14:42:37 chem: ie: most of openstack services won't be managed by pacemaker 14:42:48 chem: and take it in consideration in your work 14:43:15 chem: but needs to be confirmed with ha folks 14:43:27 Would like some more thoughts on dlrn-repo: http://lists.openstack.org/pipermail/openstack-dev/2016-June/097221.html 14:43:29 EmilienM: I know, I just had a meeting with bandini, I just wondered under which umbrella this effort is 14:43:43 chem: ha-lite 14:45:36 bnemec: that's a good proposal, did you have feedback from trown and rdo folks? 14:45:58 EmilienM: The only feedback I have so far is from derek on that thread. 14:46:12 Which is why I'm bringing it up here. :-) 14:46:18 great 14:46:37 I would get in touch with trown & other rdo folks for this one 14:46:51 ya I am less sure that is something needed in RDO 14:47:06 the convoluted repo setup is specific to tripleo-ci 14:47:40 I'll let this meeting open 30 more seconds unless something comes up 14:47:52 now, whether that mini python app is better than what is in tripleo.sh is up for debate, and I dont really have any opinion on it 14:49:40 ok thanks folks 14:49:43 thanks 14:49:45 have a great week 14:49:51 #endmeeting