18:02:19 #startmeeting networking_policy 18:02:20 Meeting started Thu Feb 1 18:02:19 2018 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:02:21 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:02:23 The meeting name has been set to 'networking_policy' 18:02:56 #topic Vancouver Summit prep 18:03:03 * tbachman scrambles in 18:03:09 SumitNaiksatam: annakk rkukura hi! 18:03:11 sorry I’m late 18:03:20 hi tbachman 18:03:21 and thanks SumitNaiksatam for the kind reminder ;) 18:03:23 tbachman: not really, just started :-) 18:04:01 annakk: did you have your moment of epiphany with the talk abstract? :-P 18:04:13 :) 18:05:10 I don't seem to have the inspiration. I have some text, but its standard stuff.. 18:05:21 annakk: ok np, we still have time 18:05:49 if anyone has an idea of how to craft it so it catches attention, please share 18:05:49 rkukura: tbachman: any further thoughts at your end on this topic? 18:05:56 SumitNaiksatam: alas no 18:06:02 nope 18:06:05 * tbachman just likes saying alas 18:06:26 tbachman: lol, we can try to be positive than that 18:06:53 annakk: whenever you feel comfortable sharing, please do 18:07:17 and we can provide our input/contribute 18:07:24 I'll share what I have, maybe this will help it evolve 18:07:30 annakk: sure 18:07:37 okay moving on 18:07:42 #topic Pending patches 18:07:52 before that, just a quick update on the UI 18:08:17 while testing something else, i stumbled on some critical issues with the UI 18:08:20 for pike that is 18:08:44 i put the fixes in two patches and they have merged (thanks rkukura and tbachman for reviewing merging) 18:08:55 SumitNaiksatam: np! 18:09:07 hopefully this makes the UI usable again, need to watch out though 18:09:24 on that point, annakk any luck getting your CI working? 18:09:46 yes, the nsx driver works with pike 18:09:55 annakk: ok nice 18:10:04 but i discovered that CI reports success very optimistically :) nee dto fix that 18:10:15 nsx CI, that is 18:10:16 ah okay 18:10:19 but i am guessing you did not exercise the UI? 18:10:24 no 18:10:34 ok good, so i beat you to it :-) 18:10:43 hopefully you will not see issues when you do 18:10:56 regarding the pending patches 18:11:03 :) 18:11:34 tbachman: thanks for adding the UT, didnt mean to create more work for you, i was just curious 18:11:44 with reference to this patch #link https://review.openstack.org/#/c/539609/ 18:11:46 SumitNaiksatam: no, it was a good point 18:11:56 i will vote shortly 18:12:29 most of these patches are apic/aim related, so annakk bear with us 18:12:32 #link https://review.openstack.org/#/c/532717/ 18:12:39 kent seems to have gotten this out of WIP 18:13:20 rkukura: you were looking at this and Ivar’s patch #link https://review.openstack.org/519788 ? 18:13:38 SumitNaiksatam: I will have some issues to file on Kent’s but probably not until Monday (I’m off tomorrow) 18:14:16 rkukura: okay 18:14:48 ivar’s patch is still WIP, and so is yours #link https://review.openstack.org/#/c/529991/ 18:14:59 rkukura: anything do discuss regarding your patch? 18:15:49 Mine could be merged at any point, with followon work to complete it 18:16:12 rkukura: you have workflow -1 on your patch 18:16:18 but it will need some cleanup, so probably not until next week 18:16:24 rkukura: how is this validation tool triggered? 18:16:49 currently, its just python code called from UTs 18:16:56 but there will be a basic CLI 18:17:39 rkukura: you CLI that calls a REST API? 18:17:42 *you mean 18:18:00 no, its more like db_manage - it works against the DB 18:18:10 rkukura: okay 18:19:05 so the mechanism drvier, plugin, etc modules are used as a “library” by this utility 18:19:25 right 18:19:43 okay 18:20:07 there were a couple of other small patches 18:20:09 #link https://review.openstack.org/537265 18:20:15 “Remove SCREEN_LOGDIR from devstack” 18:20:42 so if you do happen to use devstack, starting with Pike, you will notice, that by default it no longer uses screen 18:20:50 it uses systemctl 18:21:09 however in Pike there is still an option to use the older screen based process launching 18:21:28 we can merge the above patch on the master, but not for pike 18:21:39 * tbachman noticed that when trying to run devstack on his ubuntu system 18:21:44 (14.04) 18:21:48 tbachman: right 18:22:07 i am more used the older model, so i was happy that an option existed to use screen 18:22:17 I never used screen 18:22:19 Ah, good 18:22:21 i find it easier to just open the log file in vim 18:22:43 tbachman: there is a setting in the local.conf, ping me if you need it 18:22:45 SumitNaiksatam: I think there’s an issue b/c of systemctl in ubuntu? 18:22:53 but i guess we need to start getting used to the newer model 18:22:54 meaning theer isn’t one in 14.04 18:22:57 there 18:23:15 tbachman: oh could be, i am not aware, i have 16.04 18:23:16 so, screen will be helpful 18:23:18 ah 18:23:19 lol 18:23:29 in any case, I’ll get that from you at some later point 18:23:30 thx 18:23:30 tbachman: that probably explains 18:23:38 #link https://review.openstack.org/#/c/518269/ 18:23:46 “ 18:23:46 Add providing groups attribute of rule set in resource map” 18:23:49 another small patch 18:24:11 i think this is fine, i believe we are already populating providing groups in the dict 18:24:25 i think that covers most of the pending patches 18:24:34 #topic Open Discussion 18:24:34 I was concerned that this defined an attribute but did not implement it 18:24:52 rkukura: its read only, like the other relationships 18:25:13 rkukura: i tested it 18:25:18 but is the value actually returned? 18:25:25 rkukura: i believe so 18:25:31 And is it covered by a UT? 18:25:32 annakk: great, thanks 18:25:48 rkukura: agree, it can be added to the UT coverage 18:26:04 If its been there all along and just wasn’t defined in the extension, then this may be all that is needed 18:26:20 i think i had checked, but we were not validating other such relationships in the UTs as well, so i wasnt sure if we wanted to enforce in this case 18:26:45 ok 18:27:07 ideally we should have been testing all such relationships in the UTs (by UTs I am referring to the API level UTs) 18:27:12 i had one tip to share 18:27:21 perhaps you gusy already know it 18:27:55 in Heat its possible to pass, what they call, a "value_spec” 18:28:32 meaning, if you have extended a resource, like say in neutron, you extend the network with a custom attribute (we have apic_dn with apic) 18:29:14 you dont necessarily need to redfine the resource, you can pass a value_spec, which is essentially a key-value pair for that attribute 18:29:31 (thanks to ifti for discovering this) 18:30:06 interesting 18:30:08 SumitNaiksatam: yeah, good find 18:30:17 of course you cannot do attribute validation in heat for this, but this is expedient in cases when you cant update the upstream heat resource definition 18:30:34 tbachman: credit goes to Ifti for finding it :-) 18:30:44 ack 18:31:01 alrighty, anything else we want to discuss? 18:31:09 else we can wrap up in 30 mins :-) 18:31:24 not from me 18:31:31 me either 18:31:42 annakk: rkukura tbachman: thanks much for joining! 18:31:46 bye 18:31:48 SumitNaiksatam: thx! 18:31:48 thanks! bye 18:31:51 #endmeeting