18:02:15 #startmeeting networking_policy 18:02:16 Meeting started Thu Sep 28 18:02:15 2017 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:02:18 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:02:21 The meeting name has been set to 'networking_policy' 18:02:22 #info agenda https://wiki.openstack.org/wiki/Meetings/GroupBasedPolicy#Sept_28th_2017 18:02:29 sorry, I wan't able to progress on this this week 18:02:48 annakk: no problem, i just wanted to ask if you want to split the patch 18:02:59 so I don't know yet if remaining SubnetInUse exceptions are connected or different cause 18:03:01 and merge what you have as is 18:03:15 sure, that's an option 18:03:22 if there is more work, we can do a follow up patch 18:03:27 annakk: okay your call 18:03:52 #chairs tbachman rkukura annakk 18:03:59 :) 18:04:02 #chair tbachman rkukura annakk 18:04:03 Current chairs: SumitNaiksatam annakk rkukura tbachman 18:04:29 always thought it was a funny verb 18:04:34 in case i goof up again :-) 18:04:38 heh 18:05:03 we have this minor patch #link https://review.openstack.org/#/c/507765/ 18:05:08 Update range() to six.moves.range() for both python2 and python3 18:05:30 i havent checked if this has happened in neutron as well, if so we can merge this 18:05:52 #link https://review.openstack.org/#/c/507303/ (Include dns domain to gbp_details) 18:05:53 i think its a pep8 that we probably have disabled 18:06:06 annakk: aha, good catch annakk 18:06:09 SumitNaiksatam: that one failed CI 18:06:27 tbachman: okay :-) 18:06:42 tbachman: and so you will be following up with Ivar right? 18:06:51 SumitNaiksatam: already done, off list 18:06:55 tbachman: thanks 18:07:02 SumitNaiksatam: np! 18:07:21 this one is getting update often #link https://review.openstack.org/#/c/501847/ (Security Group support) 18:07:26 on the dns patch, looked to me like UTs failing 18:07:39 rkukura: right, thats what tbachman was pointing to as well 18:07:52 i noticed as well, though i didnt check which ones 18:07:52 oh, I thought the pep8 was for that 18:08:01 it was the RPC UTs 18:08:20 i think the pep8 one is fine 18:08:41 sorry about the confusion 18:08:41 pep8 -> six.range 18:08:46 rkukura: np 18:08:59 regarding the security groups 18:09:12 that is an apic/aim_mapping specific patch 18:09:27 and it needs a rebase 18:09:38 will talk to kent 18:09:43 I’m in the middle of reviewing the current version 18:09:50 rkukura: okay thanks 18:09:53 so far, just a nit to file 18:09:59 rkukura: sure 18:10:05 now the UI patches 18:10:26 annakk: you had a comment on this patch #link https://review.openstack.org/#/c/497235/ 18:10:32 Replaces multi select combos with transfer tables 18:10:54 “I think in multiselect tables, when a user presses the plus button, the new value is supposed to be added to the selected list (currently it replaces all values in list)” 18:10:56 and 18:11:10 “I'm looking at external connectivity/external routes selection, and there the old behavior appends new value to the selected list. But I'm ok with fixing this in a separate patchset.” 18:11:24 i was trying to understand your comment 18:11:35 (i couldnt get the patch to work in my env, will try again) 18:11:39 yes, second opinion here would be good 18:11:48 are you saying that the older behavior is preserved or is not preserved? 18:12:01 no, the old behavior is changed 18:12:18 ah, i suspected you were saying that 18:12:25 that is a bit of a problem 18:12:48 but I suspect this specific case is not a standard use-case 18:13:20 did you see this changed behavior across the UI or in one particular place? 18:13:31 perhaps its not a transfer table use-case, since there is no "list" to preserve other than selected values 18:14:17 I assume the standard case is like in action select, where you can add an action, but then not choose it 18:14:51 with external segments, AFAIR, there is no need to keep values that are not selected 18:15:05 is that correct? 18:15:14 annakk: yes, though not as intuitive (it relies on the user knowing that only one selection is relevant in that case), it was the older behavior 18:16:17 so maybe the tables are not a good match in this case 18:16:18 annakk: good question, in the case where multiple items can be selected, and a ‘+’ option is provided, i think the users expectation would be that the new items are added to the list 18:16:31 yes, I agree 18:16:34 and the whole list is being taken as input once added 18:17:01 in the older case, was the whole list being taken as input only if you selected some items, or regardless? 18:17:04 but there should be an option to cancel an item that was added 18:17:31 annakk: good point, there isnt, in which case you can argue that the new behavior is better 18:17:45 in the older version, the + button opened a dialog to add the new item, and it was added to selected list, but there was no "discarded" list 18:18:29 right, but was the whole list being taken as input (in the case of the external routers), or only the selected routes in that list? 18:18:51 looks like the new behavior + change to add new values to selected list is more user friendly than the old way 18:19:03 annakk: after this discussion, i agree 18:19:17 you brought up a very good point about discard 18:19:54 okay so i am good with this, other folks can also chime in if they have objections 18:20:01 else we can merge this 18:20:12 i will check one and accordingly put my vote on the review 18:20:19 no objection from me, but I’m not very familiar with the UT 18:20:21 UI 18:20:30 rkukura: okay thanks 18:20:51 @SumitNaiksatam: I'm not sure regarding selection.. I think the whole list was taken, but need to verify 18:20:52 to summarize, i think this is a god enhancement (not necessarily same behavior as before) 18:21:01 annakk: thats okay 18:21:14 i think change avoids all the confusion any way :-) 18:21:20 god enhancemet 18:21:25 that’s impressive ;) 18:21:49 tbachman: agree and thanks for following the discussion, it was kind of difficult to explain 18:22:05 /ME was having a go at a typo, and made one himself 18:22:05 https://review.openstack.org/#/c/497124/ ( 18:22:05 Unifies the compute and GBP member launch modals) 18:22:28 i have to try that since i couldnt get that to work before 18:22:33 and also this one: 18:22:52 #link https://review.openstack.org/#/c/497197/ (Adds policy checks to UI elements) 18:23:03 annakk: per your suggestion i am going to restack for every change 18:23:33 later we would have to check with the distro install (once everything merges) that all changes do actually take effect 18:23:49 SumitNaiksatam: yes its a pain.. and might be an upgrade problem as you've pointed out 18:23:56 yeah, just saying 18:24:07 distro install -> distro upgrade 18:24:08 i assume horizon handles this somehow 18:24:40 annakk: the troubling thing for me was taht there were some pyc files which were getting cached somewhere 18:24:48 :( 18:24:52 and i couldnt fix it unless i completely restacked 18:24:57 tbachman: yeah, kind of 18:25:03 restack is probably an overkill, they should be an easy way to do it.. 18:25:10 but it could be a devstack install consequence 18:25:13 not sure why cleaning the browser doesn't help 18:25:22 yeah it didnt work for me 18:25:29 i tried a different browser just to be sure 18:25:44 anyway 18:25:54 SumitNaiksatam: thanks for going through all of those! 18:26:03 tbachman: np 18:26:04 one more 18:26:09 and annakk! 18:26:14 #link https://review.openstack.org/488596 (https://review.openstack.org/#/c/488596/) 18:26:21 yes great work annakk! 18:26:29 (use neutron-lib address scope apidef) 18:26:44 this one ^^^ i suspect we will have to be in pike or master? 18:26:50 looks like this one failed CI 18:26:52 yes, exactly 18:27:10 so we will have to wait a little more on that 18:27:16 but not to be forgotten 18:27:20 I think its importing from exceptions when it should be importing from extensions 18:27:30 I wanted to bring up pike sync but should probably hide first :) 18:27:38 heh 18:27:49 rkukura: yeah, I saw that same thing 18:27:52 was confused 18:27:54 annakk: lol 18:28:16 yeah we need to get to it 18:28:37 perhaps in a week or so 18:28:54 i mean we can start that work in a week or so 18:29:28 i haven’t estimated the scope 18:29:59 pike was a shorted release, so i am hoping less painful to catch up 18:30:08 *shorter 18:30:08 probably a bit painful, more ovo changes 18:30:38 annakk: oh thats right 18:30:49 i think we dodged that bullet in ocata since those were being phased in 18:31:00 * tbachman has to cut out in 15 minutes 18:31:06 tbachman: sure 18:31:13 if nothing else we can wrap up 18:31:27 +1 18:31:40 okay, i will take the cue :-) 18:31:40 tbachman: are you going to sydney eventually? 18:31:46 looks like i'll be there 18:31:50 heh 18:31:55 annakk: awesome 18:32:00 I’ll be going in my own rowboat if I want to be there 18:32:03 annakk: congrats! 18:32:07 tbachman: lol 18:32:09 :) 18:32:18 annakk: keep the GBP flag flying! 18:32:34 ok :) 18:32:37 all right thanks all for joining 18:32:41 bye! 18:32:43 SumitNaiksatam: thanks! 18:32:44 #endmeeting