00:01:17 <ekcs> #startmeeting congressteammeeting 00:01:18 <openstack> Meeting started Thu Jun 29 00:01:17 2017 UTC and is due to finish in 60 minutes. The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:01:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 00:01:22 <openstack> The meeting name has been set to 'congressteammeeting' 00:02:03 <ekcs> hello all. as usual, topics are here: https://etherpad.openstack.org/p/congress-meeting-topicsfeel free to add & comment! 00:02:08 <ramineni_> ekcs: hi 00:02:08 <ekcs> https://etherpad.openstack.org/p/congress-meeting-topics 00:02:13 <ekcs> hi ramineni_ 00:03:33 <masahito> hi 00:03:40 <ekcs> hi masahito 00:05:42 <ekcs> ramineni_: is there anything to talk about on action execution logging? 00:06:01 <ekcs> anything we can help with? 00:07:58 <ramineni_> ekcs: I couldn't spend much time on it ..I'll check it this week .. I did some initial quick test ..but I couldn't get that info at that place needed 00:08:22 <ramineni_> May be I haven't enabled trace properly 00:09:01 <ramineni_> I'll try out again this week and let you know 00:09:26 <ekcs> ramineni_: ok I’m just checking in to see if anything is needed. Not that it’s urgent for me or anything. 00:09:39 <ekcs> let’s talk about PTG then. 00:09:43 <ekcs> #topic PTG 00:10:23 <ekcs> So where we left off last time was considering whether to do physical or virtual meetup. We have some options to consider for virtual just to get a sense of how workable the time zones are. 00:10:47 <ekcs> another issue to consider is how reliable the connections would be with the meeting software. 00:11:28 <ekcs> the notes are on the etherpad under last week’s meeting. https://etherpad.openstack.org/p/congress-meeting-topics 00:11:31 <ekcs> Any more thoughts? 00:13:58 <masahito> From my experience, some video conference with people from different time zone work for scheduled meetings. 00:14:40 <masahito> like weekly meetings. 00:15:29 <ekcs> so maybe here’s where we should start: let’s see if we can find at least one time slot we are all okay with. 00:16:34 <ramineni_> Opt2 , will be good for me ..8-10 00:16:56 <ramineni_> Office timings 😃 00:18:10 <ekcs> what about you masahito ? 00:18:52 <ekcs> also does it seem reasonably possible to block off that time for up to five days in one week? 00:19:03 <masahito> I prefer opt1 or 2 00:19:43 <ekcs> masahito: got it. 00:20:18 <ekcs> ok assuming opt2 for a moment, do you have a preference between traveling to colorado vs doing virtual? 00:23:06 <ramineni_> If everyone ok with travelling ..I can also make it I guess 00:23:16 <masahito> I expected we have PTG in Denver, so it's ok to travel to there. 00:23:45 <masahito> for me. 00:24:24 <ekcs> ok just to be clear, if it’s up to you to choose would you prefer to travel or do virtual or same or not sure? 00:25:25 <ekcs> or prefer not to say =) 00:27:18 <ramineni_> I feel physical meet-up would be more productive ..never done virtual meet-up though 00:29:00 <ekcs> ramineni_: got it. 00:29:07 <masahito> If everyone has no problem, like schedule, budget or delay of daily work, physical meet-up sounds better. 00:29:40 <masahito> s/delay/absent/ 00:30:18 <masahito> it's more easy to discuss. 00:31:15 <qiangcao> for those who cannot travel to denver, any way for us to somehow participate the discussion? 00:31:45 <ekcs> masahito, okay got it. I think thinrichs likely won’t be able to make it physically, but more likely virtual. But we can always dial him in when he’s available. 00:32:05 <ekcs> hi qiangcao ! I think it depends on the particular project team. 00:32:20 <ekcs> someone would need to setup the conferencing. 00:32:46 <ekcs> I’d like to do it for congress project, but the problem is, you don’t know how well the venue internet connection works until you get there. 00:33:07 <masahito> and in my side, I would be there to join discussion about different project. 00:34:15 <ekcs> masahito: got it. would it conflict with congress PTG sessions? cuz that could be a reason for us to do it virtually at a slightly different week. 00:35:34 <ekcs> qiangcao: I think it should be at least possible to get an audio connection going. but hard to be sure until we get there and see how well internet performs. 00:35:34 <masahito> it wouldn't conflict with congress PTG session. 00:35:55 <ekcs> ok well sounds like we should just do physical meetup then. let’s plan on that! 00:36:16 <qiangcao> hi ekcs! I’m particularly interested in congress. I’m in Durham, NC. 00:37:02 <ekcs> qiangcao: great! we’ll definitely do our best to make it easy for people to join virtually. it’d be great to have you. 00:37:21 <ekcs> qiangcao: would you like to tell us a bit about yourself and your interests? 00:37:29 <qiangcao> awesome! 00:37:30 <masahito> qiangcao: welcome! 00:38:44 <qiangcao> yeah, sure. I’m Qiang Cao. I’m based in the computer science dept at Duke University, as a postdoc researcher. I’ve been working on policies in the cloud for a good amount of time. 00:39:27 <qiangcao> as i said, I’m particularly interested in congress. 00:40:02 <ekcs> qiangcao: oh thats great. definitely interested to hear/read more about your research sometime. 00:41:31 <ekcs> qiangcao: look forward to more discussions in future about what you hope to do with congress. 00:41:53 <qiangcao> Sure. I Just noticed the project recently and so now I’m trying to follow the project progress. 00:42:47 <ekcs> qiangcao: Great. definitely feel free to ping me on IRC or email (ekcs.openstack@gmail.com) if you have any questions. as well a stop by the meeting when you like. 00:42:57 <ekcs> ok let’s move on the policy library then haha. 00:43:07 <ekcs> #topic policy library discussion 00:43:43 <ekcs> I think policy library is coming along okay. I’m almost done with the patch that allows activation of policy from library. just adding tests. 00:44:07 <ekcs> still more client and GUI to deal with after all the APIs are in. 00:44:22 <ramineni_> ekcs: great :) 00:45:13 <ekcs> ramineni_: would you like to discuss this issue about storing paths now? or just leave to gerrit later? 00:45:15 <ekcs> https://review.openstack.org/#/c/475605/8/congress/library_service/library_service.py@226 00:45:22 <ramineni_> Are you planning to add tempest tests also 00:46:21 <ramineni_> ekcs: right ..I saw your comment on having unique constraint .. I'm thinking if we have better way 00:47:15 <ramineni_> ekcs: but while loading policies you are checking f already in db or not .. so multiple instances won't load it right 00:47:31 <ramineni_> Do you still need unique constraint 00:47:52 <ekcs> ramineni_: not sure yet how much tempest testing we want/need. But definetly as much as needed. 00:48:38 <ekcs> ramineni_: if we don’t use unique constrant, database will allow insertion of two identical rows each representing a policy. 00:49:29 <ekcs> ramineni_: and if we don’t lock the table, then even if we try checking, node 2 may insert just after node 1 checks existence and just before node 1 finishes inserting. 00:49:35 <ekcs> and then we have duplicate. 00:49:44 <ramineni_> I'm thinking imposing name should be unique would be good idea 00:50:15 <ramineni_> Having policies with same name ifwe support it will be more confusing IMO 00:51:34 <ekcs> that’s how I had it originally. But I do like the use case thinrichs brought up. we could have one rule that says p(x):- security:secure_vm(x) 00:51:59 <ekcs> and in the library there may be different versions of the “security” policy each offering a different definition of `secure_vm` 00:53:07 <ekcs> that way, the policy that contains the rule p(x):- security:secure_vm(x) works generically regardless of which security policy one activates. 00:53:46 <ekcs> if the different versions of the `security` policy are named differently, then the rule p(x):- security:secure_vm(x) needs to be changed. 00:54:07 <ekcs> of course, this is not an ESSENTIAL use case. so we may still decide it’s not worth the trouble of supporting it. 00:54:20 <ekcs> masahito: any thoughts? 00:55:42 <ramineni_> Ok ..I'll check on this usecase .. I need to leave now ..I'll comment on gerrit 00:55:50 <ekcs> maybe it’s best if we just use unqiue name for now and it’s always easy to relax the uniqueness restriction if we decide to. 00:55:50 <ramineni_> Thanks ekcs 00:55:56 <ekcs> ok later ramineni_ 00:56:13 <ekcs> ok anything else for the last few minutes? 00:57:04 <masahito> For the first release, there's no problem for library rules to refer rules or policies that have same name. 00:57:38 <ekcs> masahito: sorry I don’t totally understand. 00:57:56 <ekcs> you’re saying we should allow policies with the same name in policy library? 00:58:05 <ekcs> or that we should NOT allow policies with the same name? 00:59:03 <masahito> sorry, I didn't check the patch yet, so I could miss the problem you discuss above. 00:59:30 <ekcs> ok. well feel free to comment on gerrit. or talk further on #congress. 00:59:35 <ekcs> let’s wrap up then. 00:59:39 <ekcs> thanks everyone! 00:59:43 <ekcs> #endmeeting