18:03:44 #startmeeting networking_policy 18:03:44 Meeting started Thu Jun 1 18:03:44 2017 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:03:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:03:47 The meeting name has been set to 'networking_policy' 18:03:59 #info agenda https://wiki.openstack.org/wiki/Meetings/GroupBasedPolicy#June_1st_2017 18:04:05 #topic Ocata sync 18:04:23 annak: starting looking at your patches, and they look good 18:04:29 *started 18:04:55 annak: last week you mentioned there was something you wanted to discuss in this context 18:05:06 annak: you want to bring that up now? 18:05:31 SumitNaiksatam: I think i've just messed up the latest patch.. gerrit didn't understand I was rebasing :) will fix 18:05:54 annak: oh okay 18:06:15 note that there were some gate issues late last night 18:06:17 SumitNaiksatam: I figured it out (the thing from last week) 18:06:23 annak: ah nice 18:06:34 the gate was breaking with a “import six” error 18:06:47 that i believe has been fixed this morning and patches are again passing the gate 18:07:21 annak: anything you wanted to discuss about the ocata sync today? 18:07:52 SumitNaiksatam: just about your suggestion - where to place the PLURALS workaround 18:08:03 i know that rkukura and tbachman have been busy with the ipv6 and dual stack stuff so they might have not had a chance to review yet 18:08:06 annak: yeah 18:08:15 annak: it was more of a question 18:08:51 SumitNaiksatam: I'm not sure if there's a convention to avoid putting utils stuff in init 18:08:53 annak: is it placed there to ensure that it always gets safely initialized? 18:09:17 annak: i dont know of a written convention but sometimes it can have unintended effects 18:09:45 SumitNaiksatam: no, just because I didn't find another good place 18:09:57 i know rkukura has some thoughts on placing things in __init__ :-) 18:10:40 I just never think to look in __init__.py 18:11:04 rkukura: ah :-) 18:11:09 ah, another thing I wanted to verify about PLURALS is that only GBP resources need to be there 18:11:10 rkukura: we are discussing this one: https://review.openstack.org/#/c/465755/2/gbpservice/neutron/extensions/__init__.py 18:11:39 * tbachman stumbles into the room 18:11:58 SumitNaiksatam, annak: I don’t have any real issue with that. 18:12:00 annak: yeah, whatever is defined in the GBP extensions 18:12:05 tbachman: lol 18:12:16 SumitNaiksatam: sorry I’m late 18:12:21 tbachman: i was in that kind of mode in the meeting today morning 18:12:27 lol 18:12:34 (stumblig and panting) 18:13:05 annak: so i am good with defining it __init__ as well 18:13:16 SumitNaiksatam: ok :) 18:13:29 just FYI, tbachman’s dual stack patch was merged last week 18:13:49 #link https://review.openstack.org/#/c/459823/ 18:13:55 tbachman: thanks for your work on this 18:14:02 SumitNaiksatam: thanks for the help and reviews! 18:14:15 tbachman: credit goes to you and rkukura 18:14:20 * tbachman notes that annak has already been busy at moving this to the RMD 18:14:34 annak: yes, thats awesome 18:15:06 annak’s patch; #link https://review.openstack.org/#/c/469681/ 18:15:15 its WIP but if you guys could have a quick look to verify the direction is ok 18:15:22 annak: yes sure 18:15:41 tbachman: and rkukura perhaps you can take a quick look too ^^^ 18:15:47 SumitNaiksatam: ack 18:15:58 had a quick look, but need to spend some time reviewin 18:16:02 reviewing 18:16:40 tbachman: sure, you have your hands full with the dual stack work, but that also makes you probably the best person to take a quick look at this since its fresh in your mind :-) 18:16:48 SumitNaiksatam: ack 18:17:39 annak: thanks again for all the progress! 18:17:47 #topic Pending NFP patches 18:17:50 np 18:17:54 it seems songole is not here 18:18:07 i had a look at the patches, and i think they are close to getting merged 18:18:12 i will follow up with him 18:18:17 #topic Open Discussion 18:18:49 SumitNaiksatam: you pinged about qos last week or so right? 18:19:00 igordcard: yes 18:19:15 igordcard: i backported your patch to stable/newton 18:19:25 igordcard: however we were running into more problems 18:19:52 igordcard: so we decided we will keep it in master and have it available starting Ocata 18:20:12 igordcard: also avoids merging a feature into a “stable” branch 18:20:28 SumitNaiksatam: okay! 18:20:29 igordcard: hope you are fine with that arrangement 18:20:42 all good 18:20:52 I have a q about tempest - is there a common suite of tests for all drivers? I need to start putting together tests for the NSX driver 18:21:00 of course if there is a strong customer/user need to have it in newton, we will revive the backport patch and get it merged 18:21:17 (sorry for interrupting.. ) 18:21:19 annak: ah good question, but tough one :-) 18:21:47 annak: one thing we definitely did not do a good job as far as GBP is concerned is we didnt add temptest tests 18:22:27 that meant we have been only relying shell scripts (exercise scripts to be more precise) in the devstack jobs to do the integration testing 18:23:07 oh I see.. I never looked how these were executed, but assumed they were tempest.. 18:23:31 annak: unfortunately not, they are shell scripts 18:24:14 annak: so short answer to your question, we dont have a tempest framework taht you can use for GBP 18:24:37 annak: for the APIC/ACI drivers there are integration tests which are being used and which run in the gate as well 18:24:51 annak: but these are more adhoc 18:24:52 does it make sense to borrow some of the scripts for vmware job or these are too specific? 18:25:12 annak: yes sure, we can definitely explore that option 18:25:40 annak: if you have a set of tempest tests, that you understand well and want to reuse here, we can definitely do that 18:25:51 SumitNaiksatam: ok thanks, I'll look closer at that 18:26:09 SumitNaiksatam: I don't know much about tempest, but I'll ask around 18:26:41 annak: you can start by using those tests for, say the VMware driver, and over time we can generalize those 18:27:03 SumitNaiksatam: yes, sounds reasonable 18:27:06 annak: we are also following a practice of accepting driver or technology-specific stuff in the GBP contrib directory 18:27:51 annak: #link https://github.com/openstack/group-based-policy/tree/master/gbpservice/contrib 18:28:04 thanks, will look at that too! 18:28:09 annak: so if its very specific to vmware you can put it there 18:28:33 I need to go offline for 5 min, will rejoin if the meeting is still on 18:28:39 the above location is for putting main code 18:28:45 if not - thanks for all the info and bye :) 18:28:56 there is one more location: #link https://github.com/openstack/group-based-policy/tree/master/gbpservice/tests/contrib 18:28:59 for the tests 18:29:02 annak: thanks 18:29:06 sorry for starting the topic and running away :( 18:29:11 annak: np 18:29:18 i was planning to wrap up anyway :-) 18:29:32 rkukura: tbachman igordcard: anything more for today? 18:29:39 SumitNaiksatam: nothing from me 18:29:50 nothing from me 18:29:53 alrighty, thanks all for joining! 18:29:59 see ya next week, bye! 18:30:03 #endmeeting