21:05:12 <ttx> #startmeeting project 21:05:13 <openstack> Meeting started Tue Jul 29 21:05:12 2014 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:05:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:05:15 <mikal> Hi 21:05:17 <openstack> The meeting name has been set to 'project' 21:05:21 <ttx> stevebaker: you looked familiar 21:05:22 <mikal> But also at a meetup so not super paying attention 21:05:22 * devananda lurks, but is still semi-afk at the ironic meetup 21:05:32 <ttx> Agenda for today is available at: 21:05:35 <ttx> #link http://wiki.openstack.org/Meetings/ProjectMeeting 21:05:37 <stevebaker> ttx: this is not the PTL you're looking for 21:05:43 * stevebaker waves arm 21:05:45 <ttx> just added a topic based on the TC discussion 21:05:53 <ttx> #topic News from the 1:1 sync points 21:06:04 <ttx> We had 1:1 syncs today for everyone but Nova, here is the log: 21:06:27 <ttx> http://eavesdrop.openstack.org/meetings/ptl_sync/2014/ptl_sync.2014-07-29-11.48.html 21:07:13 <ttx> Some plans are ambitious, some others very reasonable. We'll see where we stand in two weeks 21:07:22 <ttx> #topic Other program news 21:07:25 <ttx> Infra, QA, Docs... anything you'd like to mention ? 21:07:36 <ttx> jeblair, mtreinish, annegentle ^ 21:07:57 <dhellmann> I'm behind on email, but there should have been an announcement for a new oslo.utils library earlier today. 21:08:21 <jeblair> have folks seen the message from sean and i about testing changes we're proposing? 21:08:38 * dhellmann makes a note to read up 21:08:38 <notmyname> jeblair: yup. sounds great IMO 21:08:45 <ttx> yep, read and commented 21:08:53 <ttx> (and hijacked) 21:08:53 <stevebaker> I've started the work to move tempest orchestration scenario tests to be the heat functional tests, which means heat-slow will need to become heat-functional 21:08:58 <jeblair> functional testing++ is a big takeaway there 21:09:16 <jeblair> #link http://lists.openstack.org/pipermail/openstack-dev/2014-July/041057.html 21:09:25 <eglynn> jeblair: /me wonders about the timescales envisaged for that switch to in-tree functional tests? 21:09:35 <stevebaker> I'm not sure if I'm jumping the gun there, but am more than happy for heat to be the first here 21:09:43 <eglynn> ... and the sequencing WRT the libification of tempest 21:10:31 <eglynn> jeblair: are we looking at this switch as a juno thing? (... or more with an eye to kilo?) 21:11:05 <jeblair> eglynn: that's a good question; some projects are already heading in that direction: swift and neutron for starters 21:11:34 <jeblair> i think we need to polish a bit on the infra side and a bit on the projects side to get a good pattern for others to follow 21:12:22 <jeblair> i'd guess that we'd expect to see a good pattern in place by juno, and maybe push harder for wide adoption in kilo 21:12:22 <eglynn> jeblair: well it would be good to have a small number of solid exemplars in place initially, to avoid too much divergence/duplication of effort when the other projects follow suit 21:12:24 <jeblair> but that's just a guess 21:12:29 <jeblair> sdague may have thoughts too 21:12:49 <eglynn> jeblair: that sounds like a reasonable enough goal 21:12:57 <markwash> so we keep being kind of interested in refactoring our functionalish testing in glance 21:13:03 <stevebaker> neutron has a functional job already? 21:13:37 <markwash> I'd be interested to have somebody 'splain me what kind of testing changes we ought to consider 21:13:53 <sdague> jeblair: yeh, I think that's the right cadence 21:14:17 <jeblair> stevebaker: yes, there's a neutron func test 21:14:51 <stevebaker> jeblair: ok, I shall use that for conventions for the heat-slow replacement 21:15:00 <mestery> stevebaker: marun is our functional testing champion 21:15:02 <eglynn> markwash: IIUC the focus initially will be on moving stuff out of Tempest (as opposed to, out of the existing in-tree tests) 21:15:10 * mikal drops out of this meeting to go back to his meetup 21:15:31 <stevebaker> mestery: ok, thanks 21:15:37 <markwash> okay cool, /me reads up on the email again 21:15:41 <jeblair> stevebaker: we're still working through some infra issues on it (i need to help marun fix the sudo stuff) 21:16:13 <markwash> fwiw I think this effort, esp how it revolves around devstack, would be really fantastic for us 21:16:14 <clarkb> jeblair: its fixed 21:16:22 <clarkb> jeblair: I figured it out yesterday. sorry if that wasn't advertised 21:16:29 <markwash> since we could add functional tests for uncommon but supported configurations 21:16:31 <dhellmann> jeblair: while we're on testing, is the cross-project unit stuff still on your radar? 21:16:38 <jeblair> clarkb: ha! 21:17:05 <clarkb> jeblair: we weren't running devsdtack with neutron enabled so the sudo rules for neutron rootwrap were not installed 21:17:07 <markwash> without adding unnecessary gating load 21:18:23 <jeblair> dhellmann: i'm just really behind on reviews because of all the traveling 21:18:33 <stevebaker> markwash: heat has similar issues, I wonder if we could have a bunch of experimental jobs which run the functional tests against unusual configurations 21:18:44 <fungi> dhellmann: i have that one up in a tab in my browser and keep not making progress on it. hopefully rsn 21:18:57 <dhellmann> jeblair: understood, but I thought there was some second-thinking going on, too, so I wanted to double check 21:19:24 <jeblair> dhellmann: oh, hrm, i thought we were just executing the plan from the summit 21:19:34 <jeblair> (and the spec was just documenting that with complete sentences) 21:19:42 <dhellmann> jeblair: maybe it's just sdague who has second thoughts, then 21:20:11 <ttx> OK, I think there is lots of excitement around this, but I think the discussion can continue on the mailing-list, unless someone has the killer question 21:20:20 <jeblair> hrm, the spec for it has a +2 from me and no -1s; lifeless had some comments 21:20:31 <ttx> the one that shall not be answered on the mailing-list 21:20:55 <dhellmann> jeblair: ok, this was on irc, so maybe it wasn't strong enough to warrant a change in vote -- I assumed you 2 had talked at the sprint 21:21:55 <dhellmann> jeblair: just saw lifeless comments, so I'll get back to that next week when I return from vacation 21:22:18 <jeblair> dhellmann: i don't recall discussing this with sdague; so either we have not, or my brain has overflowed :) 21:22:37 <dhellmann> jeblair: :-) 21:22:49 <jeblair> (i'd give that even money) 21:22:50 <ttx> ok, anything more on that topic ? 21:22:56 <sdague> this didn't come up there, it was a busy week with other things. And honestly right now it's the meetup here, so need to pay attention in this room. Maybe an ML thread would be better 21:23:11 <ttx> ack 21:23:20 <dhellmann> sdague: if you have comments, please leave them on the review in progress 21:23:31 <jeblair> 95885 21:24:17 <ttx> #topic Nova/Neutron migration and other Neutron gaps 21:24:28 <ttx> So at the TC meeting just before we looked at progress on the Neutron gap coverage plan: 21:24:34 <ttx> #link https://wiki.openstack.org/wiki/Governance/TechnicalCommittee/Neutron_Gap_Coverage 21:24:41 <ttx> There are concerns on gap 5, gap 6 and gap 7 21:24:45 <mestery> ttx: For some of this discussion, markmcclain and I have been talking, and it may be best to postpone until after this week's nova mid-cycle 21:24:51 <mestery> ttx: A lot is going on right now in that room in fact :) 21:24:58 <ttx> ok, that makes sense 21:25:00 <russellb> fine with me. 21:25:10 <mestery> russellb: I still want to sync with you later this week around this though. 21:25:21 <russellb> though we do need to add more detail to 6 and 7 to make sure expectations are the same all around 21:25:27 <russellb> and 5, really 21:25:30 <mestery> russellb: +1, that's what you and I can take a cut at. 21:25:30 <markmcclain> russellb: ++ 21:25:38 <mestery> And then we can share next week 21:25:46 <russellb> mestery: yep, sounds good. i'm out friday, part of thursday 21:25:59 <mestery> russellb: Sounds like we're on for tomorrow :) 21:26:04 <russellb> mestery: yep 21:26:04 <ttx> OK, and if we need to talk about it at the next meeting, just add it to the agenda 21:26:10 <mestery> ttx: ack 21:26:13 <ttx> #topic Open discussion 21:27:01 <markwash> quick say something so ttx can't go to sleep 21:27:13 <ttx> All: the TC will have to pick a new nominee for the User committee to replace Ryan Lane. If you have great users that would make awesome dev-friendly reps for the user committee, please communicate names to your nearest TC member 21:27:54 <ttx> Anything else, anyone ? 21:27:59 <ttx> markwash: FAIL 21:28:08 <markwash> lol 21:28:08 <ttx> time to get creative 21:28:15 <russellb> how do you define locality to TC members 21:28:26 <markwash> stare directly at the sun 21:28:28 <ttx> russellb: alphabetically 21:28:41 <ttx> though some hash would make it more DHT 21:29:40 <dolphm> in other news, keystone might revert it's default back to UUID tokens https://bugs.launchpad.net/keystone/+bug/1350000 21:29:41 <jeblair> russellb: kevin bacons 21:29:41 <uvirtbot> Launchpad bug 1350000 in keystone "UUID is a more friendly default token provider than PKI" [Wishlist,Triaged] 21:29:44 * SergeyLukjanov likes using hash 21:30:26 <ttx> dolphm: i would only consider it because you snatched a round bug number 21:30:30 <markwash> the proof of work on that bug number is impressive. 4 zeroes! 21:30:35 <dolphm> ttx: oh wow. lol 21:30:36 <russellb> yeah nice bug number 21:30:59 <jeblair> i too have no substantive comments other than feeling the bug number is friendly 21:31:21 <ttx> dolphm: I don't have enough energy left to bitch about going back and forth and how lame that makes us look 21:31:24 <russellb> i'd be interested in the performance differences between the two, but trust you and the team's judgement 21:31:24 <markwash> dolphm: how about when we figure out how to make keystone use kerberos to authenticate token creation, we teach other services to use kerberos too and forgo tokens altogether? 21:31:44 <russellb> markwash: zing 21:31:47 <jeblair> markwash: ++ 21:31:55 <dolphm> ttx: that's why i've tried to hold on to pki - i didn't want to switch back. i'm ready to give up though 21:32:11 <stevebaker> is hashed PKI an option? 21:32:12 <jeblair> morganfainberg: ^ fyi 21:32:17 <dolphm> markwash: we're always discussing alternatives to bearer tokens 21:32:41 <morganfainberg> jeblair, i'm watching 21:32:43 <dolphm> stevebaker: sure, but that negates the purpose of PKI and you might as well use UUID for less effort 21:32:52 <stevebaker> true 21:32:55 <ttx> dolphm: do we ahve any idea what users use there ? 21:33:01 <russellb> ah, so token revocation events deals with the original perf problem with polling when using UUID tokens ... 21:33:04 <morganfainberg> ttx, mostly UUID 21:33:15 <dolphm> ^^ 21:33:20 <ttx> The change is backward-compatible though so I guess that's OK -- youwould still support PKI tokens 21:34:00 <dolphm> ttx: and continue to improve support for PKI tokens, which as it turns out, has improved our support for UUID at the same time 21:34:12 <dolphm> in terms of UUID performance, etc 21:34:29 <russellb> yeah 21:34:32 <markwash> dolphm: sorry for any snarkiness 21:34:34 <ttx> ISTR PKI was originally introduced to reduce load on keystone server 21:34:52 <ttx> by letting service autovalidate 21:35:00 <ttx> is that no longer true / no longer an issue? 21:35:00 <markwash> but seriously, the time when we drop the keystone IdP (or make it its own thing) might be an opportunity to rethink the authn and high level authz story 21:35:17 <morganfainberg> markwash, that is an active discussion even now. 21:35:23 <dolphm> markwash: ++ 21:35:30 <markwash> how can I join such discussion? 21:35:39 <markwash> I've been slogging through some of this stuff privately 21:35:44 <markwash> sorry glance 21:35:45 <morganfainberg> markwash, hang out in #openstack-keystone, join our meetings. 21:35:51 <markwash> kk 21:35:53 <morganfainberg> markwash, but mostly #openstack-keystone and ML 21:36:28 <dolphm> dreamy stuff like that has mostly been in IRC and out-of-session summit get-togethers 21:37:31 <ttx> well, that kept me awake for sure 21:37:37 <ttx> Now I won't be able to sleep ;) 21:37:52 <russellb> nightmares? 21:38:07 <dolphm> PKIZ_<base64-encoded-nightmare> 21:38:15 <markwash> haha 21:38:44 <morganfainberg> dolphm, PKIZ_<base64-encoded-5k+-encoded-nightmare 21:39:21 <dolphm> > 21:39:34 <markwash> phew, closure 21:39:36 <dolphm> #endnightmare 21:39:40 <ttx> #endmeeting