14:00:21 <lujinluo> #startmeeting neutron_upgrades 14:00:22 <openstack> Meeting started Thu Oct 11 14:00:21 2018 UTC and is due to finish in 60 minutes. The chair is lujinluo. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:26 <openstack> The meeting name has been set to 'neutron_upgrades' 14:00:31 <lujinluo> o/ 14:00:35 <ralonsoh> hi 14:00:57 <lujinluo> hi ralonsoh! 14:01:13 <njohnston> o/ 14:01:21 <lujinluo> o/ njohnston 14:01:35 <lujinluo> ok, let's get started 14:01:40 <lujinluo> #topic OVO 14:01:48 <lujinluo> #link https://review.openstack.org/#/q/topic:bp/adopt-oslo-versioned-objects-for-db+status:open 14:02:22 <lujinluo> #link https://review.openstack.org/#/c/608198/ NetworkRBAC in neutron/db/db_base_plugin_v2.py 14:03:00 <ralonsoh> I replied to njohnston 14:03:12 <ralonsoh> IMO, is almost almost ready 14:03:28 <ralonsoh> It will depends on https://review.openstack.org/#/c/609691/ 14:03:33 <ralonsoh> the functional tests 14:04:13 <njohnston> just added +1 14:04:21 <ralonsoh> njohnston thanks!! 14:04:31 <lujinluo> ack. i will review it today 14:05:31 <lujinluo> #link https://review.openstack.org/#/c/608628/ Use NetworkRBAC OVO in neutron/db/external_net_db.py 14:05:59 <ralonsoh> functional tests in https://review.openstack.org/#/c/609692/ 14:06:18 <ralonsoh> addressed all comments in last commit 14:07:58 <njohnston> Sorry I have not been active on that review; I'll try to take a pass later 14:08:05 <lujinluo> awesome. by taking a quick look at the codes, i think they are generally fine. i will review it today as well. 14:08:29 <lujinluo> (me neither. will try to get back on track this week, sorry :( 14:09:15 <lujinluo> next two are functional tests of the previous two patches 14:09:36 <lujinluo> #link https://review.openstack.org/#/c/609691/ functional tests of #link https://review.openstack.org/#/c/608198/ 14:10:00 <lujinluo> #link https://review.openstack.org/#/c/609692/ functional tests of #link https://review.openstack.org/#/c/608628/ 14:10:32 <lujinluo> I will review the four of them within this week! 14:10:38 <ralonsoh> thanks! no rush 14:10:55 <lujinluo> thanks for the great work ralonsoh! 14:11:14 <ralonsoh> lujinluo: no problem, that was a suggestion made by njohnston 14:11:45 <lujinluo> thank you! njohnston 14:11:47 * njohnston is very appreciative of ralonsoh's great work 14:11:59 <lujinluo> next is port binding #link https://review.openstack.org/#/c/544206/ 14:12:38 <lujinluo> i addressed most of the comments from hongbin and manjeet, except for 2 left, which i do not think i will revise the codes, but i need to run some local tests to justify the changes 14:13:07 <lujinluo> i will finish it this week too. then reviews are welcome :) 14:13:56 <njohnston> I do have one question 14:14:11 <njohnston> You turn on new engine facade in this change 14:14:16 <lujinluo> yes 14:15:09 <njohnston> This is pre-empting the engine-facade section in the agenda I realize, but I wonder if that should be a separate change? Or does it positively affect the functioning of what you're changing here? 14:16:29 <lujinluo> yes, the changes made in that patch requires using new engine facade. let me find an example here 14:18:02 <lujinluo> https://review.openstack.org/#/c/544206/24/neutron/plugins/ml2/plugin.py@399 14:19:34 <njohnston> ok 14:19:40 <lujinluo> this update() inside _process_port_binding() then __create_port_db() is inside a writer transaction 14:20:01 <njohnston> yep, that makes sense 14:20:16 <njohnston> ok, cool, I am good to move on :-) 14:20:16 <lujinluo> good! 14:20:46 <lujinluo> let's move to the next one 14:21:31 <lujinluo> #link https://review.openstack.org/#/c/608638/ 14:21:35 <lujinluo> Use NetworkRBAC OVO in tests/unit/db/test_db_base_plugin_v2.py 14:21:58 <ralonsoh> last comment already addressed 14:23:58 <lujinluo> good job! and i will review it right after the meeting 14:24:09 <lujinluo> thank you so much for your work ralonsoh!! 14:24:57 <lujinluo> next is the retry decorator changes in neutron-lib #link https://review.openstack.org/#/c/605365/ 14:25:20 <lujinluo> i replied to michel's concern but he has not replied yet 14:26:01 <njohnston> He's right about it, but it is a transitional accomodation 14:26:09 <njohnston> with a finite lifetime 14:26:16 <lujinluo> yeah 14:27:14 <lujinluo> will wait for a while before approaching him on irc 14:27:47 <lujinluo> #link https://review.openstack.org/#/c/602497/ use ovo for L3HARouterAgentPortBinding in l3_hamode 14:28:13 <lujinluo> for this patch, does zuul hang somewhere? or did i miss anything 14:29:39 <njohnston> hmm 14:30:12 <lujinluo> let me try if recheck would help or not 14:30:24 <njohnston> it's parent has not merged https://review.openstack.org/#/c/607076/ 14:30:41 <njohnston> so that is to be expected 14:31:06 <lujinluo> oh, got it 14:31:26 <lujinluo> thanks for the recheck 14:31:57 <njohnston> :-) 14:32:05 <lujinluo> #link https://review.openstack.org/#/c/507772/ network OVO 14:32:31 <lujinluo> TuanVu: since you are here today, do you happen to have any updates on it? 14:33:38 <njohnston> last time I looked at it I came away with the same questions you had posted lujinluo so I maintained my -1 14:33:55 <TuanVu> Hi Luo and Nate 14:34:01 <njohnston> Hi TuanVu! 14:34:05 <lujinluo> hi TuanVu! 14:34:21 <TuanVu> I intended to update it this week but couldn't manage to have enough time 14:34:26 <TuanVu> I will try to update it soon 14:34:34 <TuanVu> thanks a lot for your kind help 14:35:12 <lujinluo> ack! 14:35:30 <njohnston> thanks 14:35:56 <lujinluo> the rest of the patches are not updated since previous meeting. and let's give them some time 14:36:02 <lujinluo> #topic engine facade 14:36:12 <lujinluo> #link https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/enginefacade-switch 14:36:36 <lujinluo> last week, i said i have a local patch working on, but sadly i have not finished it :( 14:36:49 <lujinluo> i will try to finish it this week and push for review 14:37:24 <lujinluo> #topic open discussion 14:37:31 <njohnston> Are there any other objects that we think we can switch to the new facade, on a per-object basis? 14:38:37 <njohnston> the Agent object, for example? 14:38:48 <lujinluo> i have not gone through the codes so i cannot tell for sure, but theoretically speaking for the objects that we have already converted, should be fine to switch on new facade 14:38:52 <njohnston> according to the tracking spreadsheet, work on that jas completed 14:39:20 <njohnston> I think I'll push up something to enable the new facade on that, and let's see how well it works 14:39:32 <lujinluo> great 14:40:11 <njohnston> if that works OK then we can move on Network and Router as next steps 14:40:36 <lujinluo> yep! 14:40:59 <njohnston> #action njohnston to push change that enables new engine facade for Agent object 14:41:26 <lujinluo> thanks! 14:41:43 <lujinluo> does anyone have anything else? 14:42:31 <lujinluo> if not, i have one small favor i would like to ask 14:42:39 <ralonsoh> sure 14:42:56 <njohnston> ok 14:42:58 <lujinluo> first i am sorry i was not active enough to track the changes last week, and i saw people actually added me as reviewers to those new patches 14:43:13 <lujinluo> but that one goes to my previous company email, which i did not have access to.. 14:43:23 <lujinluo> i mean i dont have access anymore 14:43:40 <lujinluo> so if possible.. please add the account with my gmail address, :) 14:43:44 <njohnston> ok 14:43:47 <ralonsoh> ok 14:44:10 <lujinluo> i will also contact the admin ppl to see if they can remove the old account for good 14:44:17 <lujinluo> thanks! njohnston ralonsoh 14:44:27 <ralonsoh> thanks! 14:44:32 <njohnston> np 14:44:46 <lujinluo> ok, if that's it. let's call it a day :-) 14:44:53 <njohnston> I have one other thing 14:44:59 <lujinluo> sure 14:45:14 <njohnston> I am starting to play around with how best to implement baked qieries 14:45:30 <njohnston> I have a very WIP change here: https://review.openstack.org/609715 14:45:38 <lujinluo> is it the one we may need for all the joins? 14:47:04 <njohnston> it speeds up operations. Basically it short cuts the process of constructing the text of the SQL query. Especially for frequently invoked queries that never change it will be a significant speed boost. 14:47:44 <lujinluo> sounds very impressive 14:48:18 <njohnston> it was already implemented in ODL; check out https://review.openstack.org/#/c/591363/ for comparison 14:49:36 <lujinluo> got it. will check on both patches. this will be a big performance gain in large scale systems 14:50:40 <njohnston> absolutely 14:52:41 <lujinluo> yeah, would be very interested to see how it works! 14:52:49 <njohnston> that is all for me 14:53:04 <lujinluo> ok if nothing else, let's call it a day! 14:53:10 <lujinluo> thanks everyone for joining 14:53:15 <lujinluo> have a good day/night 14:53:22 <ralonsoh> bye 14:53:26 <lujinluo> and hope to see you all next week! 14:53:44 <lujinluo> #endmeeting