14:00:03 #startmeeting sahara 14:00:03 Meeting started Thu Sep 8 14:00:03 2016 UTC and is due to finish in 60 minutes. The chair is vgridnev. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:07 The meeting name has been set to 'sahara' 14:00:09 o/ 14:00:14 hi 14:00:30 #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda 14:01:46 o/ 14:01:53 o/ 14:02:24 #topic News / updates 14:03:05 hi 14:03:17 hi 14:03:23 so, finally https://review.openstack.org/#/q/status:open+branch:master+topic:bp/initial-kerberos-integration is finished I think, waiting for some reviews on it 14:04:19 additionally fixed bug in hadoop-swiftfs that prevents correct job execution with Swift when kerberos enabled 14:04:50 also working on coordinating bug fix work 14:04:59 so, current status of RC1 is 14:05:02 #link https://launchpad.net/sahara/+milestone/newton-rc1 14:05:17 7 bugs left, 2 blueprints are still in open state 14:05:35 on this week I added rack awareness for CDH 5.5 and CDH 5.7, added event log for HDP, updated docs for Designate 14:06:04 The CLI core is done and functioning; we may want to add additional BPs for each plugin implementation. 14:06:12 writing tests for scenario fraemwork, it includes cluster verifiacation and added the ispublic and isprotected flags to cluster 14:06:42 egafford, so we can mark that as implemented also 14:07:24 I don't think we will need specs for these blueprints 14:07:27 vgridnev: Yeah, I'll do that after the meeting and create the new BPs. 14:07:36 vgridnev: Nah; we get how those work. :) 14:08:30 ok, seems like Kerberos is only open blueprint now :) :( 14:10:33 egafford, tosky is on PTO, isn't? 14:10:47 can't find him on IRC 14:11:00 vgridnev: Tjat 14:11:08 That's correct; he'll be back next week. 14:12:26 ok, thanks then. I was thinking about release for sahara-tests, we will need to check what we have to land before release. 14:13:08 #topic Open discussion 14:15:03 so, work on bugs is in progress, so what about docs? we actually can continue working on docs even after release, we will just land docs in master, but it will not be in stable/newton 14:16:04 I will prepare documentation about Kerberos shortly, also I saw a updates about pagination, and designate 14:17:08 vgridnev: Are we doing bug week, then doc week, or together? 14:17:25 vgridnev: all updates about designate are already merged 14:17:30 If we're doing doc review now, I'll send a mail to the ML to try to get that review process started. 14:19:00 ok, I think that we should finishing work on bugs (RC1 is around Wednesday) and also we can start working on docs now 14:20:30 vgridnev: Great; giving myself an action to send doc ML post kickoff thing. 14:23:48 Ok, about Ocata release, it's going be super short (only 16 weeks) 14:23:57 #link https://review.openstack.org/#/c/357214/7/doc/source/ocata/schedule.rst 14:25:02 vgridnev: o.O wow, that's tiny. 14:25:11 it is very short 14:25:28 then we should have strictly defined goals to reach, because there is no so much time to implement huge features 14:26:59 vgridnev: Yeah, we need to prioritize very aggressively. 14:27:11 agreed, for this first experience I believe we should have a really short spec deadline 14:27:49 egafford, tenobreg additionally we should keep in mind that there are always a LOT of weekends around New Year and Christmas, so we should remove 2 weeks for schedule 14:27:57 Do we want to implement an O1 spec post deadline? An O2 spec post deadline? 14:28:03 vgridnev: +2. 14:28:15 vgridnev, true 14:30:24 egafford, my vote is for O1 spec post deadline 14:31:43 agree for O1 spec deadline. This leaves spcecs a chance to be implemented in time. 14:31:50 tenobreg, I don't think we really need deadlines for specs, but probably we have to do so for large features. If something is really needed, we always can consider exception 14:33:06 I would say R-12 is better 14:33:31 vgridnev: I like the O1 spec post deadline with an option for exception until O2. 14:33:52 vgridnev, the spec deadline is exactly that. Since specs are usually for larger features, we should consider limiting the submission. This will give time to review and implement without surprises 14:34:03 since the time frame is changing 14:34:14 egafford, +2 14:35:40 additionally we will have some community wide goals to reach, and one of that is support of Python3 in sahara, but I think that since we have passing python3 jobs for a while it should not big problem for us 14:36:02 vgridnev: Yeah we're actually not behind OpenStack on that. 14:36:28 egafford: I'm not actually shure that eventlet runs correctly on py3 14:36:38 some time ago there were troubles with it 14:37:00 NikitaKonovalov: Ouch, that would be quite awful. 14:37:04 but if they are solved, than we in fact are in good shape 14:38:19 there was something like that: https://review.openstack.org/#/c/260999/ 14:38:58 if this run was correct, so then story is good 14:40:39 vgridnev: Well, if there are bugs deep in eventlet, those can be quite occasional and awful (debugging coroutine/threading issues is quite awful and the bugs are often hard to repro.) 14:40:51 NikitaKonovalov: Do you remember anything more specific about the bug? 14:41:36 egafford: afaik the eventled threads could simply hang forever 14:42:00 Yup, that's quite awful. 14:42:34 quite awful for sure 14:44:17 NikitaKonovalov: http://eventlet.net/doc/changelog.html it looks as though some py3 fixes were implemented in 17.3: /me researches release of that. 14:44:52 anyway right after newton is done we should start moving our jobs to python3 to validate everything is ok 14:45:07 +1. 14:46:06 +1 14:47:14 anything else that we can cover? 14:47:31 not from my side 14:48:01 Nothing here. 14:48:14 egafford, about replacing some artifacts from sahara-files.mirantis.com, you can track progress on https://review.openstack.org/#/c/367271/ 14:48:33 Awesome! I'll watch that. 14:49:25 ok, finishing meeting 14:49:29 #endmeeting