04:01:28 #startmeeting congressteammeeting 04:01:29 Meeting started Fri Jan 25 04:01:28 2019 UTC and is due to finish in 60 minutes. The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot. 04:01:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 04:01:33 The meeting name has been set to 'congressteammeeting' 04:02:06 ekcs: hi 04:02:33 hello akhil_jain ! 04:03:12 happy friday! 04:03:33 Happy friday 04:03:52 topics are kept here as usual #link https://etherpad.openstack.org/p/congress-meeting-topics 04:03:55 let’s take a minute to put down the topics we want to discuss today 04:07:19 ok let’s get started then. 04:07:59 #topic zuul migration & gate 04:09:42 we had the issue that check jobs run from congress-tempest-plugin were not using the specified stable branches. that has been fixed with gmann’s help by adding a specification of branch override in the legacy playbooks. 04:10:09 actually the patch is ready but not merged yet: https://review.openstack.org/#/c/631329/ 04:10:33 * gmann adding that in today review list. 04:10:55 ah thanks gmann haha. 04:11:14 thanks for that migration. that is useful and helpful for infra/QA perspective 04:11:44 along with it, patches are also created to migrate the master branch jobs to non-legacy format so that we should ha;ve fewer problems in the future. https://review.openstack.org/#/c/630414/ 04:11:53 yup and thanks for the help gmann ! 04:12:19 https://review.openstack.org/#/c/631609/ 04:12:24 https://review.openstack.org/#/c/631894/ 04:12:55 in the process we also moved to ubuntu bionic because stein will primarily target bionic rather than xenial. 04:13:29 just wanted to give an update on these changes. any thoughts or comments? 04:14:08 +1 on moving to bionic 04:14:21 ekcs: it seems all running fine on bionic ? 04:14:35 yes looks like it! 04:15:02 cool 04:16:23 ok moving on then. 04:16:29 #topic use case in library for cross project network 04:16:42 akhil_jain: is that your topic? you wanna start us off? 04:17:04 yes. i want to enquire about that use case. 04:17:09 link https://github.com/openstack/congress/blob/master/library/cross_project_network.yaml 04:17:41 we are making differnt groups and check if network from diff group atatched to differnt 04:18:06 but by default according to me demo must not be able to see alt 04:18:14 alt_demo networks. 04:18:34 isnt that it or i am wrong somewhere? 04:19:46 I think I understand each thing you’re saying, but I’m not sure what the problem is =p 04:19:48 one project should not be able to see the network from other project until that is shared. so in which case are using this? 04:20:03 is that is valid use case then? 04:20:58 Ah I see. First I’ll say that the example policy is just an example. 04:21:06 becuse if one project cannot see other project networks. it wont be able to attach to that. so what we are trying to achieve with this use case 04:22:21 oh ok, if that is so. i was checking all of them to list use case we can achieve with congress. 04:23:33 also for cfg_validator and cloudfoundry, do we have usecase? 04:24:05 ekcs: FYI, we are working to list all the use case of congress with all scenario. akhil_jain is preparing the sheet and then policy sample file. once all those policy sample file works fine then he can upstream those and document also 04:24:32 akhil_jain: may be we can share that sheet with ekcs on google sheet etc for his inout 04:24:35 input 04:25:27 ya we can do that. and finally after all use cases are ready we can add to congress lib. that will be helpful for end users. 04:25:55 on this specific case: if there is no way for the violation to happen because of the openstack interface, then it’s not very compelling use case. But I can think of two reasons why we still include the policy example. 04:27:15 1. it illustrates an example of something congress can monitor/enforce even if it’s not necessary in the current openstack context. congress can be used with different virtial networking software and monitor this kind of policy. 04:28:32 i see, make sense. 04:29:25 ekcs: thanks for explaining. got your point 04:30:02 2. in general, the principle for policy is that there can be a separation between the policy author and the system admins. Ideally, the people who write policy do not need to be intimately familiar with what is and isn’t possible. instead, they can simply write how things should be. 04:30:59 3. I guess 3 is that it’s an additional layer of check. suppose a hacker somehow exploited a bug to attach a certain network to a different project. then the policy will find it. 04:31:23 gmann: Ah I see that’s cool. yes I’ll be happy to help with that! 04:32:19 3 is relevant in a multi-tenant situation. 04:32:29 ekcs: ok i will share a google doc with you today 04:32:59 cloudfoundry to be honest I’m not very familiar with it. 04:33:36 cfg validator I can find the emails and specs and send to you. 04:34:08 oh ok. i have found 2 cases as of now. that might not be achieveable by cuurent drivers. but we can with minor additions 04:34:35 very cool! use case work is super valuable. 04:34:57 ok any more on this topic? 04:35:10 yes for making congress user friendly and easy to use. we need to add use cases 04:35:21 +1 04:35:27 thats all for now 04:35:54 #topic congress-neutron gate 04:36:07 ok maybe we can touch base a bit on the neutron test breakage. 04:36:25 great initiative tracking down the problem akhil_jain 04:36:39 on this patch: https://review.openstack.org/#/c/631688/ 04:37:36 I had some problem fully understanding the commit message. is it just the N/A vs OTHER mismatch or you’le also pointing out a different problem? 04:37:53 thanks eric. but according to your comment i think neutron wont be aggreed with changing it. as they are managing floating_ip with differnt api set. and we are using ports api 04:39:00 right. 04:39:25 its about status field only. floating ip unless assigned. wuill show status N/A 04:39:40 that in congress will be evaluated to OTHER 04:39:51 Not sure if thats your question 04:40:10 ok. 04:40:46 ok I understand now. 04:41:02 so whats your opinion on this? change in congress or in tempest-plugin test? 04:41:32 * akhil_jain options 04:42:06 1. skiping in tempest plugin 2. adding NA in congress. 04:42:07 I’d think change in congress. at least as long as neutron team confirms N/A is correct behavior. 04:42:32 I think we can skip the test on floating ip for now, as you have done in the patch. 04:42:36 ok then I will raise bug as you said. and lets see their response 04:42:41 then change congress once neutron team confirms. 04:43:13 does that make sense? 04:43:24 ok sounds good to me 04:43:29 great! 04:43:39 #topic patches 04:43:49 ok do we have any patches to discuss? 04:44:54 not from my side this time. 04:45:23 btw I’ll be pushing the next part to this json polling ingester hopefully today. https://review.openstack.org/#/c/629691/ 04:45:26 ok then! 04:45:56 #topic py36 04:47:38 just a quick mention as people are adding py36 jobs, they run into problems with the current latest version of congress not supporting according to setup.cfg. so I’ll be putting out a new release 04:47:52 i mean congress client. 04:48:21 ok that’s all from me. 04:50:07 #topic open discussion 04:50:07 anything else to discuss for today? 04:51:12 ekcs: not from my side 04:51:33 ok then! 04:52:17 helpful discussions today. let’s end meeting then. 04:52:22 have a great weekend! 04:53:00 Thanks and you too! 04:53:11 #endmeeting