14:01:34 #startmeeting neutron_upgrades 14:01:35 Meeting started Thu Jul 5 14:01:34 2018 UTC and is due to finish in 60 minutes. The chair is njohnston. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:38 The meeting name has been set to 'neutron_upgrades' 14:01:58 Hello all 14:02:03 Hi Nate :) 14:02:23 Unfortunately lujinluo has been called away on urgent family business. We wish her the best! 14:02:44 oh, thanks for the notice, Nate 14:03:08 She asked me to take care of the meeting this week and, if I remember correctly, next as well. 14:04:13 oh, what a bad news 14:04:23 The action items from last week are both ones for her, so let's skip them for now and perhaps we can address them in due course. 14:04:26 hopefully she'll come back soon 14:04:34 #topic OVO 14:04:51 #link https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/adopt-oslo-versioned-objects-for-db 14:05:20 I have a few up this week, starting with 14:05:24 #link https://review.openstack.org/579053 14:06:13 I need to look at the testing more clearly and figure out why tests are failing. 14:06:25 I will have an update on that later today 14:06:43 thanks 14:06:50 #link https://review.openstack.org/579678 14:07:29 I've just left a comment on that one 14:07:57 It'll be merged soon, I think 14:08:12 I saw your comment, TuanVu, and that makes me scratch my head because I put it back. I'll do it again, the right way this time. :-) 14:08:22 thanks for your reviews 14:08:23 awesome :) 14:08:38 next is one of yours 14:08:54 sorry, one of Van's 14:09:01 #link https://review.openstack.org/549168 14:09:54 This one has a long history 14:10:33 but it seems to be failing everything except pep8 and sphinx docs 14:10:53 yes, i've tried to reduce the failed casasse 14:11:08 then you have looked at it more than I have :-) 14:11:13 what did you see? 14:11:52 one off common problem 14:12:24 I encountered for those UTs 14:13:00 is something like " router['router']['id'], KeyError: 'router' 14:13:18 i'm still wrap my head around it 14:13:30 can you give a suggestion 14:14:11 i feel that some OVO code I've converted incorrectly 14:14:28 but haven't figured out it 14:14:48 Hmm, it strikes me as odd that a router would have a field called 'router'. Why would the object be built to nest router information like that? 14:15:57 ok, we can continue the discussion in gerrit 14:16:05 ok cool 14:16:17 thank you nate 14:16:39 #link https://review.openstack.org/579689 14:17:04 This is one of mine; I just asked for a recheck 14:17:24 it looks great 14:17:39 there's only 1 failed fullstack test remaining 14:18:00 I added a note on this to the tracking spreadsheet as well, like 62: This partially covers the needed work. But we need to talk about db/provisioning_blocks.py, because models are being passed to functions there as arguments. 14:18:05 the recheck will solve it, I think 14:18:42 yeah, I see 14:19:21 So the method being called from test_provisioning_blocks.py is add_model_for_resource: https://git.openstack.org/cgit/openstack/neutron/tree/neutron/db/provisioning_blocks.py#n36 14:19:39 that takes a db model and maps it to a callback resource 14:20:20 looks very good 14:21:07 So I think provisioning_blocks.py will be one of the last things converted; it is noted so in the spreadsheet (line 128 in one of lujinluo's sections) 14:21:22 but I wanted to discuss it with you just to be transparent 14:21:52 thanks for this info, Nate 14:22:09 next up is TuanVu with 14:22:15 #link https://review.openstack.org/507772 14:22:33 I’ll need to spend more time on it to figure out how to build the compatibility layer 14:22:33 I’ll upload a new patch set soon 14:23:28 Do you know what it is that vmware-nsx is using? 14:25:35 other than what have been mentioned by boden, I haven't covered all of them yet 14:27:37 I don't have anything else to add on this topic, moving on 14:27:55 #link https://review.openstack.org/579682 14:28:37 Thanks manjeets for the review; he suggested that I incorporate some additional model transitions I mad missed. That is on my todo list for this week. 14:29:13 #link https://review.openstack.org/562489 14:29:21 TuanVu, tell us about this 14:29:40 yeah 14:30:04 As the discussion in last IRC meeting, in order to avoid blocking the delivery of OVO, I've moved joins to object and add corresponding TODO for it. 14:30:47 If you have any other suggestions, please don't hesitate to let me know. 14:31:24 I just +1ed it, it looks good to me 14:31:41 thank you, Nate 14:32:00 so I'll apply the same solution for https://review.openstack.org/#/c/561834/ 14:32:07 very good 14:33:17 #link https://review.openstack.org/565773 14:33:30 that says updated 6 days ago, but I don't see what updated 14:33:52 hopefully lajoskatona has completed the timeboxed assignment and can return to this 14:34:01 I will ask for a status update in gerrit 14:34:34 ok, that's great 14:35:04 #link https://review.openstack.org/561834 14:35:17 TuanVu, tell us about this 14:35:43 it's the same with Service Type patch which has just got +1 from you 14:35:55 about doing "join queries" 14:36:47 indeed, I look forward to the new PS 14:37:09 yeah, I will upload it tomorrow morning 14:38:06 The next few changes are from lujinluo, and are in merge conflict 14:38:24 so I think we should leave them until she can return 14:38:34 ok 14:38:41 I agree 14:39:56 I did want to circle back to one change from earlier 14:40:05 because it has an issue that I am not sure how to resolve 14:40:21 #link https://review.openstack.org/#/c/579053/ revisited 14:41:30 that code is failing the unit test neutron.tests.unit.ipam.test_subnet_alloc.TestSubnetAllocation.test_subnetpool_concurrent_allocation_exception 14:42:18 I see 14:42:25 The code for that test is here: https://git.openstack.org/cgit/openstack/neutron/tree/neutron/tests/unit/ipam/test_subnet_alloc.py#n190 14:42:52 You can see that the test has the expectation that the db model will reject the addition with an ORM error 14:43:03 I am not sure how to best implement that in an OVO way 14:43:11 Does anyone have any thoughts? 14:45:30 hmm, I have no experience with this so I have no idea at this moment. I'll put it on my check list tomorrow and send you feedback if there's any 14:45:33 is that ok? 14:45:49 That would be wonderful 14:46:19 #topic Open Discussion 14:46:31 Does anyone have anything else they would like to discuss? 14:46:42 one question about Rolling Upgrade 14:47:01 for example from Queens to Rocky 14:47:32 is that possible without current state of OVO? 14:47:46 sorry 14:48:04 I mean *is that possible with current state of OVO?* 14:49:27 I would not consider myself a deep expert on that yet, but I feel pretty safe in saying that it is probably not safe yet. 14:50:05 There are a few central changes that are waiting on everything to be converted to OVO, and then they can be done 14:50:11 thank you Nate 14:50:27 I ask this because we're about to do some tests for this 14:50:49 like line 68 on the spreadsheet in neutron/db/db_base_plugin_common.py 14:51:29 and the add_model_to_resource code in neutron/db/provisioning_blocks.py we discussed earlier 14:52:17 thanks for giving your opinion on this matter 14:52:24 I really appreciate it :) 14:52:54 i think it will be less bumpy than before, but the spreadsheet says we are only 26% done so I am afraid I can't give better news :-) 14:53:55 anything else? 14:53:56 yeah, thank you Nate 14:54:01 that's all from me 14:54:24 Thanks for everything TuanVu and hungpv_, look forward to chatting with you next week. 14:54:56 thank you nate, sweee u 14:55:00 see you soon Nate 14:55:04 :) 14:55:39 #endmeeting