17:00:10 #startmeeting security 17:00:11 Meeting started Thu Sep 10 17:00:10 2015 UTC and is due to finish in 60 minutes. The chair is tmcpeak. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:16 The meeting name has been set to 'security' 17:00:19 #chair hyakuhei 17:00:23 Warning: Nick not in channel: hyakuhei 17:00:24 Current chairs: hyakuhei tmcpeak 17:00:28 o/ 17:00:37 wassup e'rybody 17:00:43 I'll give a few minutes for people to trickle in 17:00:51 * bknudson o/ 17:00:52 o/ 17:00:57 o/ 17:01:09 o/ 17:01:18 good to see you all 17:01:28 hope you've recovered from your travels back from midcycle for those who went 17:01:41 hyakuhei might be out saving lives (literally) 17:01:44 will try to multitask for another meeting. sorry in advance. 17:01:53 hi 17:01:54 so he's asked me to run today 17:02:16 o/ 17:02:18 ahh speak of the devil 17:02:21 want to run it Mr. Rob? 17:02:27 I've added you as a chair already 17:02:32 * hyakuhei just got through the door from a lifeboat shout. 17:02:41 you're efficient 17:02:47 Can you take the first 5 minutes please? I need to cool off hehe 17:02:51 sure, wlll do 17:02:54 Danke 17:02:54 ok 17:02:54 more refugees? 17:03:02 #topic Midcycle Recap 17:03:07 bknudson: drunk students. 17:03:11 we got a lot done last week 17:03:17 o/ 17:03:20 tons of work on Anchor, Bandit, and other projects 17:03:37 I'd like to do a small recap of each activity for those that weren't there and try to assemble next steps to make sure we don't lose momentum 17:03:45 elmiko: care to start with our architecture overview? 17:03:51 sure 17:04:19 we discussed, at some length, the ideas surrounding secure comms between the control plane(also known as under-cloud) and the tenant space 17:04:46 the results of our discussion are an architecture around passing credentials to nodes in the tenent 17:05:01 elmiko: did we end up formalizing and posting that anywhere? 17:05:05 the architecture addresses a privilege dropping scheme 17:05:13 tmcpeak: working on that 17:05:18 awesome 17:05:26 seems like a good devstack post at least, maybe BP coming? 17:05:55 the priv. drop scheme will, hopefully, create a nice way for projects to have their nodes in tenant space gain the creds necessary to perform actions on cloud resources (such as swift( 17:06:11 cool, yeah, seemed promising 17:06:14 this also roles in some interesting ideas about endpoint access as well 17:06:30 i am working on a formalized writeup in a spec-ish format for us to bat around 17:06:31 for those that weren't there and are interested we'll have a formal writeup from this and some diagrams to better illustrate 17:06:42 i'm also gonna do a poc with sahara as the target 17:06:46 perfect 17:06:52 tmcpeak: cool. Thanks. 17:07:03 elmiko: we'll look forward to that 17:07:15 we should probaby talk more at the summit as well, but hopefully we'll have something formal to talk about and a poc to demo 17:07:16 elmiko: thank you so much for looking into this 17:07:19 we'll need a summit session 17:07:20 and doing the documentation 17:07:28 hyakuhei: +1 17:07:37 this seems really useful 17:07:48 hyakuhei: np, thinking about the docs has actually surfaced new questions for me. so it will be nice to rip it apart =) 17:07:59 So I think we came up with some nice ideas. I love the idea of working on a poc 17:08:05 will they have lightning talks? maybe a good place for a demo 17:08:12 bknudson: +1 17:08:14 I’m planning to have a fishbowl on this auth stuff 17:08:27 cool 17:08:39 hyakuhei: you cooled off? want to talk about anchor progress? 17:08:44 maybe combine with a(nother) dynamic policies 17:08:53 Yeah happy to 17:08:54 dynamic policies discussion 17:09:04 bknudson: +1 17:09:04 although that might be too much for 1 room. 17:09:11 cool off also equals do all the things I was supposed to do but couldn’t because I was on a lifeboat rescuing drunk people :P 17:09:23 #topic Anchor 17:09:35 Finally tagged with a version pre-api breaking changes being accepted 17:09:48 I also created some docker containers to make spinning up test instances a bit easier 17:09:53 next time bring a drone and post on youtube. 17:10:01 As this is what all the cool kids are doing 17:10:17 hyakuhei: is the container working well? 17:10:24 Seems to :) 17:10:31 https://review.openstack.org/#/c/221285/ 17:10:38 I’d appreciate people kicking the tyres on that 17:10:54 if, unlike tmcpeak you could look the Readme first thatd be great ;) 17:11:13 haha 17:11:19 hey, sorry Im late 17:12:03 In other news I created a docker org for openstack security 17:12:11 So if you want to be added ping me your username 17:12:50 nice 17:13:05 RTFM is actually useful sometimes ;) 17:13:19 It’s nice because it makes things look legit :P 17:13:34 #link https://review.openstack.org/#/c/221285/5/Dockerfile.anchorbase - topline 17:13:46 So I think that’s all I had on anchor - tkelsey ? 17:14:17 I guess that’s it then 17:14:22 #topic Bandit 17:14:33 tons of work done last week 17:15:01 largely around unit tests, tkelsey refactored result store, I wrote a little tool to automated upstream project checks 17:15:11 chair6 has added multi-threading 17:15:14 (in flight) 17:15:17 nice 17:15:21 ccneill has an output formatter done 17:15:26 well in review 17:15:31 I'm sure there's lots of stuff that I"m fogetting 17:15:32 hey, sorry stepped away for a sec 17:15:41 elmiko, browne, chair6, tkelsey, bknudson, what else? 17:15:48 so i posted to the ML about Bandit next steps etc 17:16:06 fyi I created a Bandit team under the “openstacksecurity” org if you guys want to consume some of the coolaid. 17:16:07 pleas read and voice any thoughts there :) thanks 17:16:08 fixed a bunch of bugs 17:16:37 i had a question on bandit 17:16:45 browne: what's up? 17:16:49 browne: ? 17:16:50 tox.ini shows it tests with pypy, but do we care about it? 17:16:58 and what version if we do? 17:17:36 the Pypy tests come from copy n past of boilerplate tox code 17:17:38 yea, the pypy stuff always causes me trouble 17:18:00 ok, so are we cool if i remove it? 17:18:16 I am ok with it 17:18:20 sounds good to me 17:18:22 with removing it, that is 17:18:28 ok, cool 17:18:47 what about the crypto tracking? how's that coming along? 17:19:00 FYI: Bandit thread: #link http://lists.openstack.org/pipermail/openstack-dev/2015-September/073723.html 17:19:37 browne: so the crypto stuff needs a report formatter, I'll start working on that soon, but have had no time 17:19:39 browne: good point, where did we get with that? 17:19:48 i got distracted by concurrency :) 17:20:07 yeah chair6 :) I spotted you did somthing for that, ill have to check it ou 17:20:10 *out 17:20:22 ok, np. just interested because i think it'll be very useful 17:20:33 browne: +1 17:20:50 im going to find time to work on it as well, just haven't managed yet 17:21:14 tkelsey: this post looks good 17:21:40 tmcpeak: thanks, I invite people to follow up, so we have a more active showing on the list etc 17:21:46 Yeah good email 17:21:52 also it looks like severity filtering might be broken, git blame tkelsey? :) 17:22:11 heh, probably my bad 17:22:20 bug it and i'll go fix it 17:22:25 cool, will do 17:23:12 chair6: do you want to talk about your concurrency stuff ? 17:23:45 tmcpeak: I'll add in unit tests while fixing stuff as well 17:23:52 tkelsey: you the man 17:24:01 not if i broke it :P 17:24:11 but you're fixing it and adding unit tests :) 17:24:27 https://review.openstack.org/221861 is where the initial stab at multiprocessing is 17:24:39 we could really use unit tests of the manager.py 17:24:42 take a look at the code and at the discussion, add your own thoughts :) 17:24:54 browne: +1 and lots of other places :) 17:25:12 manager seems to be one of the big remaining areas 17:25:12 chair6: will do! interesting stuff 17:25:20 chair6: nice simple approach 17:25:21 I like it 17:25:24 yeah, but manager.py is the big one. 17:26:09 that might be good on Bandit for now 17:26:11 ? 17:26:12 ok, we should focus on that next then 17:26:45 I want to squawk about recruiting ;) 17:26:56 im good for bandit stuff, chair6 looks like a tidy patch :) 17:27:07 browne: I'll look at manager testing as well 17:27:26 thanks tkelsey 17:27:28 tkelsey: thanks. 17:27:34 since hyakuhei isn't stopping me, I'll squawk about recruiting 17:27:43 so we'd like to get some new membership for OSSG 17:27:44 tmcpeak: +1 17:27:49 we have a lot of great projects in flight 17:27:57 but seems like we'll become spread thinly if we aren't already 17:28:08 Yeah - very thin 17:28:10 for example some of the agenda items at the midcycle we couldn't address and nobody was available to lead 17:28:26 I started OSSG with a talk bdpayne and nkinder were doing in the Bay 17:28:30 I'd like to get similar talks going again 17:28:43 +2 17:28:48 so we worked on a deck last week 17:28:50 think it's in fairly good shape 17:28:57 talks about OpenStack Security and all the related efforts 17:28:59 Yeah, needs a few tweaks still 17:29:07 Probably runs 10-15 minutes 17:29:31 it looks good now 17:29:32 is anybody who hasn't said so already interested in presenting locally? 17:29:55 so far we have bknudson will do one in wherever it is he lives ;) 17:30:04 hyakuhei and dg_ are going to do one in London'ish area 17:30:06 dg___: got any info on the london one? 17:30:11 browne and I will do the bay 17:30:23 I will do in San Antonio and Austin 17:30:24 michaelxin is going to do something in Texas somewheres 17:30:39 there's a minnesota meetup every once in a while 17:30:41 I was considering trying to piggyback some on local OWASP meetups etc 17:30:44 kyle mestery chairs it 17:31:01 also I should mention that we'll take a three pronged approach where possible 1) openstack meetup, 2) security focused meetup (like OWASP), 3) perhaps something for a security program in a university 17:31:06 bknudson: Ack, would love to get folks there! 17:31:36 that's kind of where we got to 17:31:44 I want to make sure this does't get dropped though 17:32:00 some effort on our part here could boost our numbers meaningfully going forward 17:32:01 I will use OWASP meet up first. 17:32:15 michaelxin: please let us know how it goes 17:32:16 mestery: at the last security group meetup we discussed doing an intro to openstack security group at meetups. 17:32:20 Excellent, thank you tmcpeak 17:32:23 +1 17:32:31 Should we track our presentations ? 17:32:40 michaelxin: yeah, definitely 17:32:41 bknudson: That would be pretty epic! Would love to get that in at our Minnesota meetup 17:32:42 wikipage maybe? 17:33:08 mestery: great, we can discuss offline. 17:33:09 hyakuhei: sure or etherpad 17:33:13 I had one but I lost it 17:33:15 bknudson: ack 17:33:17 should probably start a new one 17:33:17 hehe 17:33:45 let's use this: 17:33:48 #link https://etherpad.openstack.org/p/security-project-recruiting 17:33:54 so who will be first guinea pig for the presos? 17:34:24 browne: we'll have to do ours later 17:34:32 I'll be out of the bay until mid December 17:34:45 tmcpeak: np 17:35:04 oh forgot to mention chair6 and sicarie are doing Seattle :) 17:35:06 I will try to do one for OWASP San Antonio in October since September is already booked. 17:35:12 michaelxin: great 17:35:17 meanwhile I'll continue to refine the deck 17:35:26 is there anybody that's interested that doesn't have access to the deck? 17:35:29 Austin will have LASCON 17:35:32 PM me your google docs email addy 17:35:39 We will try to present there too. 17:35:45 Any resemblence of the current recruitment deck to other HP decks is purely conincidental…. 17:35:52 :P 17:35:58 As they’re almost certainly copyright... 17:36:01 LASCON is next month. 17:36:03 So yeah, shiny and new :) 17:36:27 action items here? 17:36:33 want to make sure this doesn't stall out 17:36:41 Volunteer yourself, update the etherpad 17:36:44 refine the deck 17:36:57 michaelxin, browne, hyakuhei, dg_, chair6 can you each take a few minutes this week and try to refine the deck? 17:36:59 bknudson 17:36:59 Add non-HP content (none of the content is vendor specific but more authors is probably better) 17:37:03 tmcpeak: Sure 17:37:10 tmcpeak sure 17:37:15 tmcpeak: Got it. 17:37:19 awesome, thank you 17:37:20 Well, I’ll try, drunk people are making it hard for me today lol 17:37:34 lol 17:37:35 tmcpeak: i'll try. flying to Purdue tomorrow for a career fair event 17:37:39 I'm definitely going to get hammered and go boating when I visit UK this December 17:37:46 haha 17:37:52 and demand for hyakuhei to personally come save me 17:37:52 lol 17:38:27 cool, so that's probably good for recruiting 17:38:35 michaelxin: summary for Syntribos? 17:38:50 Yeah that’d be cool 17:38:55 I already updated that part 17:39:02 for the slide 17:39:11 no I mean summarize on where we got during midcycle please 17:39:25 ic 17:39:47 I gave a demonstration to the group about its current function including different ways to fuzz 17:39:51 how to run it. 17:40:01 configuration and payloads. 17:40:07 There are some good feedbacks. 17:40:11 I really enjoyed the demo 17:40:16 great work michaelxin and team 17:40:19 Thanks. 17:40:24 did we identify concrete next steps to drive it forward? 17:40:30 The group liked it and wanted me to move it to Stackforge. 17:40:36 I already got the process started. 17:40:44 michaelxin: +1 17:40:47 awesome! 17:40:52 nice 17:40:54 https://review.openstack.org/#/c/220351/ 17:40:55 +1, syntribos demo was nice 17:41:04 elmiko: tkelsey Thanks. 17:41:13 awesome 17:41:17 At the same time, send me more feedbacks. 17:41:34 michaelxin: will do, it's on my queue of things to try to break things with Syntribos 17:41:46 I will meet my team and give them update syntribos. 17:41:52 michaelxin: looking forward to seeing it land :) 17:41:58 tmcpeak: Thanks. 17:42:04 also we pretty much smashed through the OSSN queue :P 17:42:09 Yeah we did 17:42:15 Though there’s plenty waiting for late +2s 17:42:34 cool, anything we need to do there to get those through? 17:43:03 They need iterating through, making sure we’ve got project cores 17:43:03 who are our +2's? hyakuhei, sicarie, elmiko, and nkinder? 17:43:09 Yeah 17:43:47 cool, what's the best way to get those finished? 17:43:53 i'll take another pass through the open ones 17:43:57 better earlier while they're still fresh ;) 17:44:01 I’ll run through them tonight and badger people 17:44:07 perfect 17:44:09 i can't +2 my own though... ;) 17:44:22 well i can, but you know 17:44:24 elmiko: I’ve got you covered bro! 17:44:48 ^5 hyakuhei 17:45:02 awww 17:45:09 hehe 17:45:10 what a loving community 17:45:19 tmcpeak: +1 17:45:51 #topic Any other business 17:46:22 I just want to say a massive thank you to everyone who’s helped and contributed to our work recently 17:46:28 hyakuhei: +1 17:46:32 hyakuhei: +1 17:46:33 hyakuhei: +1 17:46:42 hyakuhei: +1 17:46:51 I'm always amazed how much we get done at these things 17:46:54 The Security project has so many cool activities running now, we’re generating great cotent, grokking the code, providing ways to encrypt many of the things etc. It’s awesome. 17:46:56 and it's always nice to meet face to face 17:47:03 +1 17:47:04 ++ 17:47:14 hear hear, +1 17:47:22 true that 17:48:38 hyakuhei: should we start planning for summit things now? 17:48:39 or later? 17:48:47 I won't be there but it sounds like quite a few of you will 17:48:56 tmcpeak: good point 17:49:10 tmcpeak: It’s started 17:49:16 We’ve got two fish bowls and two work rooms 17:49:22 Bandit can have a WR if it wants it 17:49:27 Not sure about the other WR 17:49:37 FB - one will be the multil privilege auth issue 17:49:37 should we start an etherpad? 17:49:38 hyakuhei: great start 17:49:41 yes 17:49:43 might be cool to do some inter-project socializing 17:49:46 of Bandit 17:50:03 wide receiver? fullback? 17:50:03 tmcpeak: interesting 17:50:05 There’s the cross-project design sessions that we should also try to target heavily 17:50:42 maybe invite some project guys to come voice their thoughts on bandit gates 17:50:51 hyakuhei: +1 17:51:24 yeah, sounds good 17:51:24 I think once we get enough projects on bandit there will be enough momentum to get every project on it 17:51:35 bknudson: hope so :) 17:51:43 bknudson: that would be awesome 17:51:48 eventually it'll be required for a tag or badge or whatever they call it 17:51:56 nice 17:52:36 this just popped up on the ml, more food for thought about the privilege cloud user thing. #link http://lists.openstack.org/pipermail/openstack-dev/2015-September/074126.html 17:53:08 elmiko: interesting 17:53:13 good spot 17:53:34 Yeah that is interesting, we actually had that on the mid-cycle etherpad 17:53:44 that spec has been around for awhile now, but looks like its being moved out of nova into openstack-specs 17:53:54 The solution tightly couples to Bandit for identiy which is somewhat scary. 17:54:03 hehe 17:54:10 Honestly I think the solution we came up with is considerably more elegant 17:54:13 s/Bandit/Barbican/, i hope? :) 17:54:19 yeah 17:54:19 hyakuhei: s/Bandit/Barbican/ 17:54:22 heh lag 17:54:24 That would be _really_ scary 17:54:47 tightly couple all the things to Bandit 17:54:50 elmiko: Maybe we can work on our alternative option next week 17:54:54 * tkelsey hides 17:54:57 hyakuhei why scary? 17:55:01 hyakuhei: sounds good 17:55:44 redrobot: because I don’t want identity to pivot back to systems that hang off some other identity provider, it creates wierd cyclic auth paths 17:55:56 However, I haven’t read _all_ the words in that spec yet 17:56:29 redrobot: We love Barbican but sometimes people want to do “interesting” things with it 17:56:47 hyakuhei lol 17:57:02 also, the spec should probably be focusing on castellan instead of barbican, no? 17:57:13 if castellan now does Certificate magics. 17:57:16 elmiko nope, not in this case 17:57:25 ah, ok 17:57:27 Which I think someone told me it does? 17:57:39 hyakuhei nope, no certs in castellan 17:57:47 ok, that’s probably a good thing :) 17:57:49 actually, I take that back, you can store a cert via Castellan 17:57:56 heh 17:58:00 but you can't (and porbably never will) provision certs with Castellan 17:58:24 OpenStack : The cloud controller brought to you from the minds of “Inception” … wheels within wheels…. 17:58:40 lol 17:58:51 lol 17:58:54 Ok, I think that’s time people! 17:59:08 Anyone else want to annoy redrobot before I call it? 17:59:14 :) 17:59:21 #endmeeting