15:00:31 <ihrachys> #startmeeting neutron_upgrades 15:00:32 <openstack> Meeting started Mon May 16 15:00:31 2016 UTC and is due to finish in 60 minutes. The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 <openstack> The meeting name has been set to 'neutron_upgrades' 15:00:54 <jlibosva> \o 15:00:59 <johndperkins> o/ 15:01:13 <sc68cal> o/ 15:01:17 <ihrachys> hello everyone! I hope you enjoyed the lack of meetings lately. 15:01:48 <sc68cal> no, I missed you all :( 15:02:05 <ihrachys> I don't see some folks (rossella_s, korzen?) 15:02:25 <korzen> hello 15:02:37 <ihrachys> the previous meeting was supposed to occur, but then I guess some catastrophe happened :) 15:02:39 <pcm_> cd vcode 15:02:47 <ihrachys> so that's the 1st meeting after the summit 15:02:48 * pcm_ wrong window... 15:03:13 <ihrachys> I was off last two weeks, so probably you folks will need to fill us in 15:03:29 <ihrachys> first, actions from the last meeting 15:03:36 <ihrachys> #topic Actions from the last meeting 15:03:49 <ihrachys> "ihrachys to cancel next meetings in openstack-dev@" 15:03:55 <ihrachys> nah, that one is boring. moving on 15:04:03 <ihrachys> "ihrachys to update wiki page with the check list for object adoption (object, database access code, API test coverage, ...)" 15:04:21 <ihrachys> ok, for that one, I started some draft check list of 'things to cover when doing object transition' 15:04:28 <ihrachys> #link https://wiki.openstack.org/wiki/Meetings/Neutron-Upgrades-Subteam#Object_adoption_check_list 15:04:57 <ihrachys> I suggest everyone who did some work for objects to go thru the check list and validate that all of it is covered 15:05:12 <ihrachys> I also hope that people will fill in gaps in the list 15:05:36 <ihrachys> tl;dr it's not enough to land object class to claim a resource done. it's just first step. 15:06:23 * ihrachys gives 1 min to complete reading and complain about the list 15:07:37 <korzen> the wiki steps looks nice 15:07:37 <ihrachys> ok, 1 min ended. I guess everyone is on board with it. 15:08:05 <ihrachys> generally, if folks have cycles, at this point I suggest we spend them on API test coverage since it's going to block a lot of objects. 15:08:42 <ihrachys> so ask me how to help, and I will point you to some tests and patches that we need to push 15:08:59 <ihrachys> that would be something along the lines of https://review.openstack.org/#/c/306272/ 15:09:12 * ihrachys needs to get back to the patch asap 15:09:33 <ihrachys> ok, let's cover usual topics 15:09:36 <ihrachys> #topic Partial Multinode Grenade 15:09:49 <ihrachys> sc68cal: would you mind filling us about where we stand with it? 15:10:06 <sc68cal> Just got +A'd a couple minutes ago, working its way through the gate 15:10:12 <sc68cal> https://review.openstack.org/#/c/299843/ 15:10:34 <ihrachys> oh man. 15:10:36 <korzen> sc68cal, great! 15:10:42 <ihrachys> it's thrilling and scary at the same time 15:10:56 <ihrachys> do we have the very latest result? 15:10:56 <sc68cal> I have to find my notes - I spoke with armax last week and we put some stuff in an etherpad for the next steps 15:11:01 <sc68cal> let me dig up the etherpad link 15:11:04 <ihrachys> I hope it's not breaking :) 15:11:47 <sc68cal> found it 15:11:48 <sc68cal> https://etherpad.openstack.org/p/neutron-multinode-jobs-newton 15:12:06 <ihrachys> fyi non-voting jobs from the very latest patches seem succeeding 15:12:10 <ihrachys> looking at the link 15:12:21 <sc68cal> #link https://etherpad.openstack.org/p/neutron-multinode-jobs-newton Newton multinode grenade next steps 15:13:24 <ihrachys> I think it's a good list of things to follow up on 15:13:43 <ihrachys> for dvr, we have the job in check too 15:15:14 <ihrachys> as per grafana, the failure rate is the same as for non dvr 15:15:21 <ihrachys> (and low) 15:15:41 <rossella_s> hi all, sorry for being late, I am actually off today but I thought I should join this :) 15:15:51 <korzen> one improvement we need is to move dhcp and L3 our of primary node 15:15:55 * ihrachys waves at rossella_s 15:16:02 <korzen> but still using the 2 node setup 15:16:18 <ihrachys> korzen: please add to the etherpad 15:16:48 <korzen> ihrachys, ok 15:17:20 <ihrachys> sc68cal: do you take all of it, or you want someone to help you with it? 15:18:00 <sc68cal> I'm happy to have people join 15:18:28 <ihrachys> ack 15:18:55 <ihrachys> #topic Object implementation 15:19:21 <ihrachys> generally, the topic of object adoption was raised on a design summit session 15:19:35 <ihrachys> and the decision was made that new features are required to provide objects with them. 15:19:51 <ihrachys> it's up to us folks to identify those patches without objects in flight though 15:20:19 <ihrachys> note that we still need to do preparation work for features that are e.g. extending a resource 15:20:33 <korzen> it is good we have new OVO for vlan aware VMs: https://review.openstack.org/#/c/310410 15:20:37 <ihrachys> we don't expect other developers to implement objects just to add a small attribute somewhere 15:21:07 <ihrachys> korzen: that's cool 15:21:48 <ihrachys> apart from general notes above, I don't have specifics 15:21:56 <ihrachys> korzen: rossella_s: anything to mention? 15:22:07 <rossella_s> ihrachys, I don't have anything 15:22:18 <korzen> rossella_s, what is the status of port OVO? 15:22:27 <korzen> is it close to merge? 15:23:00 <rossella_s> korzen, nope it's not, unfortunately I didn't have time to work on it lately 15:23:19 <ihrachys> rossella_s: I guess it's also blocked by API tests for sorting/pagination? 15:23:26 <korzen> we lost Martin, is there anyone to help rossella_s? 15:23:30 <ihrachys> rossella_s: or we talk about object itself, without adoption? 15:23:30 <rossella_s> ihrachys, that too 15:23:50 <rossella_s> ihrachys, the object without adoption 15:23:56 <ihrachys> ok, from my side, I will concentrate on getting API layer ready for objects. 15:24:03 <rossella_s> ihrachys, great! 15:24:33 <ihrachys> I hope in parallel we'll get the object itself in shape so that we can meet each other at the adoption cross road quicker 15:25:04 <korzen> I would love to see any review on Subnet OVO: https://review.openstack.org/#/c/264273 15:25:32 <rossella_s> korzen, will do...I was partly off and partly focusing on other stuff after the summit, sorry for that 15:25:34 <korzen> and Network OVO as well ;) https://review.openstack.org/#/c/269658 15:25:39 <johndperkins> I've got a couple ovo patches out that just need reviews as well 15:25:46 <johndperkins> https://review.openstack.org/#/c/304862/ and https://review.openstack.org/#/c/307969/ are network/router 15:26:17 <ihrachys> johndperkins: ok 15:26:35 <ihrachys> korzen: we don't have adoption patches for those, right? 15:26:45 <korzen> ihrachys, nope, not yet 15:26:50 <ihrachys> korzen: would be cool to try that and see what's missing 15:27:05 <korzen> ihrachys, yes, I will do this this/next week 15:28:05 <ihrachys> johndperkins: would be cool to switch focus of the team to bits missing for existing objects. 15:28:32 <korzen> johndperkins, could you or some from OSIC help with port object? 15:28:44 <johndperkins> korzen: absolutely 15:29:00 <dasm> ihrachys: korzen: which objects require attention? 15:29:19 <korzen> dasm, port OVO 15:29:40 <dasm> korzen: just this? 15:30:01 <ihrachys> dasm: I think at this point it's more about getting API layer to the point where we can start switching first class resources to objects. API layer work is needed to support sorting/pagination. that's one thing. I bet when we start adopting more in the db code, we will hit more missing bits. 15:30:20 <dasm> ihrachys: mhm. fair enough 15:30:25 <ihrachys> dasm: we have patches in flight, but they will probably need some love. 15:30:47 <dasm> ihrachys: do we have one place where links are stored? 15:30:49 <ihrachys> dasm: fyi I plan to tackle some of it this week, but since I am dug under emails from 3weeks, some help would always be welcome. 15:30:55 <ihrachys> dasm: ovo topic 15:31:00 <dasm> ihrachys: gotcha 15:31:05 <ihrachys> https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:ovo 15:31:25 <dasm> ihrachys: i think *a lot* of patches require "some love" :D 15:31:43 <ihrachys> yes, and we better not increase their number until they are in better shape 15:32:27 <dasm> understandable. OSIC people are involved, like johndperkins, so we think we can try to help decrease the number. 15:32:42 <ihrachys> ok. I guess interested folks will sync after the meeting if needed on specific pieces to share. 15:32:51 <korzen> rossella_s, do you have any action items that dasm and OSIC can take right away? 15:33:22 <rossella_s> ihrachys, regarding the port ovo? 15:33:32 <korzen> rossella_s, yes 15:34:01 <rossella_s> ihrachys, just update the patch 15:34:44 <korzen> rossella_s, like this one: https://review.openstack.org/#/c/253641/ ? 15:34:46 <ihrachys> rossella_s: I assume you won't have time for it this week and folks are safe to take it? 15:35:10 <rossella_s> ihrachys, yes this week for sure 15:35:21 <ihrachys> ack. great to know. 15:35:24 <dasm> rossella_s: ok. good to know. 15:35:24 <rossella_s> ihrachys, I will try to free up later 15:35:51 <rossella_s> anyone interested please ping me 15:36:10 <ihrachys> ok, I guess we have something to roll with on that topic. 15:36:12 <ihrachys> #topic Open discussion 15:36:16 <korzen> https://review.openstack.org/287756 Integrate the port allowed address pairs VersionedObject in Neutron this also need some love 15:36:42 <ihrachys> I don't have anything more to talk, feel free to raise more stuff now 15:36:47 <korzen> I have one open 15:36:56 <korzen> tenant_id -> project_id renaming 15:37:08 <dasm> korzen: ++ 15:37:11 <korzen> should we handle the project_id right now? 15:37:24 <korzen> or wait until dasm work will be ready? 15:37:34 <dasm> basically, spec is not merged yet: https://review.openstack.org/#/c/257362/ 15:37:51 <korzen> I would prefer to go with project_id from the beggining 15:37:53 <dasm> but most of the things are there. so in some time, we will also need to use project_id 15:38:42 <ihrachys> korzen: for new objects? fine. as long as they write in consistent way (same field) in db 15:39:05 <ihrachys> korzen: note that e.g. qos policy has tenant_id, so we would need a switch anyway. 15:39:16 <korzen> I have updated the Subnet OVO to use the project_id https://review.openstack.org/#/c/264273/24/neutron/objects/subnet.py@151 15:39:57 <ihrachys> korzen: that's nice. if dasm goes with alembic, he will just need to remove the translation 15:40:09 <dasm> ihrachys: yes. 15:40:33 <ihrachys> korzen: thanks for looking into the future 15:41:03 <ihrachys> anything more to discuss? 15:41:18 <ihrachys> or everyone can have 20 mins till next meeting slot? :) 15:41:31 <korzen> I have introduced the null testing in: https://review.openstack.org/#/c/308441 15:41:45 <korzen> still have issue with cmp not present in Python3 15:42:07 <korzen> do you think it is worth continueing? 15:43:03 <ihrachys> I am not sure what's the goal there. I will need some time to actually read it thru. 15:43:26 <korzen> main reason was to have UT covered for getting nulls for generated values 15:43:46 <korzen> it can fast test if null can be set in DB 15:44:00 <korzen> and for example currentlu 15:44:18 <korzen> corrently, when we do not set the property in OVO, the exceptin is raised 15:44:34 <korzen> and we are not able to set null fetched from DB in OVO 15:45:47 <ihrachys> korzen: is "null from db" case different from "no value in db at all"? 15:47:33 <korzen> ihrachys, not sure right now 15:47:59 <ihrachys> ok, I will need to take a closer look. maybe playing with code will help me to understand the issue. 15:48:18 <korzen> we can for example set the gatway to null 15:48:25 <korzen> gateway* 15:48:33 <korzen> when defining the subnet via CLI 15:49:35 <korzen> we can continue in patch review 15:49:43 <ihrachys> aye 15:50:02 <dasm> ihrachys: one more question about: https://review.openstack.org/#/c/306272/9 15:50:12 <dasm> ihrachys: do you need any help with it? 15:50:17 <dasm> it seems to be blocking a lot of stuff. 15:50:42 <ihrachys> dasm: that's the patch I was going to get right after I handle all critical emails I have since 3weeks. 15:51:07 <ihrachys> dasm: but we could start building more resource tests on top of it already, right? 15:51:16 <dasm> ihrachys: True. 15:51:55 <ihrachys> dasm: so let's say I handle the patch 1st priority, and people start building more on top while it's still in review. it should not change much itself, just need to fix gate and other nits. 15:52:41 <dasm> ok. fair enough :) 15:53:52 <ihrachys> ok, I guess that's what we have right now 15:53:59 <ihrachys> let's use those 7 mins for good 15:54:02 <ihrachys> #endmeeting