15:01:22 <ihrachys> #startmeeting neutron_upgrades 15:01:25 <openstack> Meeting started Mon May 22 15:01:22 2017 UTC and is due to finish in 60 minutes. The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:28 <openstack> The meeting name has been set to 'neutron_upgrades' 15:01:29 <ihrachys> o/ 15:01:54 <lujinluo> o/ 15:02:31 <ihrachys> hi lujinluo 15:02:34 <ihrachys> anyone else? 15:02:56 <lujinluo> hi ihrachys , maybe we can wait for a while 15:04:33 <manjeets> hi 15:04:39 <manjeets> sorry for being late 15:04:47 <ihrachys> hi manjeets 15:05:06 <lujinluo> hi manjeets 15:05:46 <ihrachys> let's have a look at the action items from the last time we met 15:05:52 <ihrachys> #topic Action items from prev meeting 15:05:56 <ihrachys> first was "ihrachys to respin metering OVO patch: https://review.openstack.org/#/c/433125/" 15:05:59 <ihrachys> the patch is merged, good 15:06:06 <ihrachys> next was "tonytan4ever to respin LIKE patches" 15:06:55 <ihrachys> the patch went through several iterations since then: https://review.openstack.org/#/c/419152/ 15:09:06 <ihrachys> we can discuss it later in OVO section 15:09:19 <ihrachys> next was "manjeets to take over neutron-db-manage CLI command for data migration" 15:10:04 <manjeets> I did few changes but i had one question how are we gonna populate migration scripts 15:10:11 <ihrachys> I saw manjeets respinned the patch, thanks for that. 15:10:21 <manjeets> as victor had only one example script and using that 15:10:46 <manjeets> are we gonna have a separate folder for all those scripts and call from there ? 15:11:32 <ihrachys> manjeets, well I thought that we will have a stevedore namespace where we would expose those 'scripts' 15:11:44 <ihrachys> and then it's a matter of taste where we keep the code in-tree 15:12:01 <ihrachys> a separate dir probably makes sense 15:12:16 <manjeets> ohk, i'll work on that this week 15:12:17 <ihrachys> but there should still be some level of indirection 15:12:31 <ihrachys> for external projects to be able to provide their own scripts 15:12:46 <ihrachys> ok cool 15:12:48 <ihrachys> next was "manjeets to try craft out cycles for SG adoption patch" 15:13:03 <ihrachys> it's this patch: 15:13:03 <ihrachys> https://review.openstack.org/#/c/284738 15:13:21 <ihrachys> oh I see it has two +2s 15:13:32 <ihrachys> I gotta review and merge it after the meeting then ;) 15:13:40 <ihrachys> thanks for respin 15:13:45 <ihrachys> and that's all we had 15:14:04 <ihrachys> for work items I mean 15:14:12 <ihrachys> #topic OVO 15:14:55 <ihrachys> https://review.openstack.org/#/q/topic:bp/adopt-oslo-versioned-objects-for-db+status:open 15:15:46 <manjeets> may be we should start taking ownership of these patches and get it done 15:15:54 <manjeets> as most of them have changed their jobs 15:15:58 <ihrachys> yeah 15:16:04 <lujinluo> agree. now I am not sure which patches are unmaintained 15:16:12 <ihrachys> let's walk through those that fell through cracks 15:16:29 <ihrachys> Router: https://review.openstack.org/#/c/307964/ 15:16:39 <ihrachys> I think lujinluo started looking at it correct? 15:16:45 <lujinluo> yes 15:17:12 <ihrachys> ok. what's missing besides the eager load thing? https://review.openstack.org/#/c/466237/ 15:17:30 <lujinluo> l3 agents as a synthetic field 15:18:06 <ihrachys> is it something router-l3-agent-binding patch helps with? https://review.openstack.org/#/c/377074/ 15:18:25 <lujinluo> this is may due to my lack of knowledge, but i'm not sure how to load only l3 agents but not all agents.. 15:18:54 <lujinluo> i have not checked that patch. will do after the meeting 15:19:26 <ihrachys> and why is it needed that we expose agents through synthetic field exactly? 15:20:15 <lujinluo> https://github.com/openstack/neutron/blob/master/neutron/db/models/l3.py#L62-#L64 here and https://github.com/openstack/neutron/blob/master/neutron/db/l3_agentschedulers_db.py#L560 here 15:20:53 <lujinluo> in my current integration patch, I can call through router.db_obj.l3_agents but i am not sure if it is a good idea to keep it calling from db 15:22:14 <ihrachys> if it's exposed on db_obj.l3_agents you should be able to extract it from there on from_db_object. 15:23:19 <lujinluo> then we do not need to expose l3 agents as synthetic field? 15:25:05 <ihrachys> I think your links suggest that there is a use case to extract the list of agents from the router. so I now merely say that probably you should have all needed to implement such a field 15:26:55 <ihrachys> let's continue in the review, I will have a look. I also take over the binding patch since Anindita left. 15:27:04 <lujinluo> ok, thanks 15:27:22 <ihrachys> manjeets, is Toni still with us to follow up on LIKE patch? https://review.openstack.org/419152 15:27:48 <manjeets> ihrachys, I am not sure about he was with rackspace 15:28:07 <ihrachys> I don't see him in this channel right now. 15:28:25 <ihrachys> let me ask him on gerrit. if not, I will take over, the patch is very close. 15:28:36 <manjeets> ok sounds good 15:30:02 <ihrachys> the dvr mac address is also on Tony, I will try to understand if he is still with us, and we can discuss that next time. 15:30:18 <ihrachys> next is allocation/endpoints from Anindita: https://review.openstack.org/#/c/367810/ 15:30:29 <ihrachys> it's on WIP 15:30:42 <manjeets> ihrachys, I may take over one or two, I have to attend a conference this week 15:30:48 <ihrachys> oh that's because of update_objects missing. 15:30:52 <manjeets> i'll comment on gerrit after reviewing 15:31:29 <manjeets> for taking over 15:31:43 <ihrachys> manjeets, which one do you want to take? 15:31:48 <ihrachys> allocations? 15:32:03 <ihrachys> it depends on https://review.openstack.org/#/c/452266/ that is on me right now. 15:32:31 <ihrachys> I will bubble up that one in my todo list 15:32:52 <manjeets> ohk I'll take allocations 15:32:53 <ihrachys> as for endpoints, it may need to wait until the update_objects completion. 15:33:16 <ihrachys> manjeets, note it's not clear if update_objects is ok in the shape it is right now. 15:33:53 <manjeets> I'll spare some time to review these 15:34:00 <ihrachys> ok good 15:34:15 <ihrachys> next in the queue is dhcp agent network binding: https://review.openstack.org/#/c/370452/ 15:34:21 <ihrachys> I believe Nakul is no longer 15:34:31 <ihrachys> Kevin tried to respin it in the past 15:34:33 <manjeets> yes he left 15:35:15 <manjeets> seems like patch is close just need to address few comments from kevin 15:35:27 <ihrachys> the patch would really benefit from someone taking over it. 15:35:34 <ihrachys> yeah and those comments seem to be nit 15:35:43 <ihrachys> manjeets, so maybe you can look at it instead of endpoints? 15:35:49 <manjeets> sure 15:35:55 <ihrachys> nice 15:36:50 <ihrachys> next is ip allocation pool: https://review.openstack.org/#/c/396711 15:36:59 <ihrachys> this one seems to be a no-brainer really 15:37:02 <ihrachys> very small, very simple 15:37:08 <ihrachys> shame it's not in yet 15:37:19 <ihrachys> anyone want to take it? maybe lujinluo ? 15:37:26 <lujinluo> sure :) 15:38:33 <ihrachys> thanks l 15:38:48 <ihrachys> meh :) that was awkward. thanks lujinluo 15:39:20 <ihrachys> the next was ip allocation: https://review.openstack.org/#/c/396718/ 15:39:50 <ihrachys> that one seemed fine, but has some functional test issue 15:40:34 <ihrachys> should be an easy thing to land if we tackle the test 15:40:46 <lujinluo> i can try to take care of it too 15:40:54 <ihrachys> thanks!! 15:41:09 <lujinluo> no problem! 15:41:45 <ihrachys> do we want to go forward in the list or we have enough on our plates already? 15:42:50 <ihrachys> manjeets, lujinluo ? 15:42:53 <manjeets> I'd say lets get these done first, if anyone wants to take ove, just comment over gerrit and go ahead with re spinning 15:42:58 <lujinluo> i think i have enough 15:43:16 <ihrachys> ok, let's focus on those, and we will revisit the list next time. 15:43:49 <ihrachys> there are no new UpgradeImpact tagged patches. 15:44:02 <ihrachys> I am also not aware of any ongoing work in gate for upgrade use cases. 15:44:10 <ihrachys> with that... 15:44:13 <ihrachys> #topic Open discussion 15:44:19 <ihrachys> ...do we have anything else to discuss? 15:44:24 <lujinluo> i have 1 question 15:44:30 <ihrachys> shoot 15:44:46 <lujinluo> my employer is pushing about this for a long time. are we still target OVO for Pike? 15:46:04 <ihrachys> I think we have a good chance to complete most of the OVO work this cycle. I should have some more time, again, to work on it. 15:46:25 <ihrachys> lujinluo, that being said, who's the employer, and which use cases do they have in mind? 15:47:04 <ihrachys> are we talking more about online-upgrades as the real thing they are interested in? because OVO is just code restructuring/framework thing. 15:47:23 <ihrachys> they are interrelated but not the same 15:47:26 <lujinluo> i am employed by fujitsu. i do not think my manager has any specific use cases for OVO yet, but he purely wants to see it land asap 15:47:54 <lujinluo> i also explained that online-upgrades do not really require OVO 15:48:26 <ihrachys> I see. this work is definitely on the high priority radar for Pike. PTL is supportive and even said he will have some cycles to push it forward which is promising. 15:49:08 <ihrachys> does it answer the question? 15:49:08 <lujinluo> ok, i will try to explain this to him too, thanks 15:49:13 <lujinluo> sure, it does 15:49:14 <ihrachys> ok cool 15:49:26 <lujinluo> sorry, i have one more to manjeets 15:49:33 <manjeets> sure 15:49:37 <ihrachys> shoot again ;) 15:50:14 <lujinluo> regarding online data migration, i have a patch under reviewing that may be used as another script migration example. https://review.openstack.org/#/c/288271/ i hope so 15:51:48 <ihrachys> mmm, thanks for pointing it out. I need to have a look at how it will play with rolling upgrades. it seems very similar to multiple port bindings case in terms of database evolution. 15:52:20 <lujinluo> ihrachys, yes, please, this is why I link it here 15:52:21 <manjeets> i'll review that 15:53:18 <lujinluo> and that's all from me ;) 15:54:13 <ihrachys> great. you seem busy with different things. :) 15:54:30 <ihrachys> I bet you are also good at juggling :) 15:54:34 <ihrachys> ok 15:54:41 <lujinluo> ;) 15:54:53 <ihrachys> if nothing else, I end the meeting in 3 15:54:57 <ihrachys> 2 15:55:01 <manjeets> 1 15:55:01 <ihrachys> 1 15:55:04 <ihrachys> boom! 15:55:06 <lujinluo> 0 15:55:09 <ihrachys> thanks everyone 15:55:10 <manjeets> thanks !! 15:55:11 <ihrachys> -1 15:55:13 <ihrachys> #endmeeting