14:01:11 <ihrachys> #startmeeting neutron_upgrades
14:01:11 <openstack> Meeting started Thu Mar 22 14:01:11 2018 UTC and is due to finish in 60 minutes.  The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:15 <openstack> The meeting name has been set to 'neutron_upgrades'
14:01:19 <mlavalle> o/
14:01:22 <lujinluo> o/
14:01:38 <ihrachys> hey mlavalle lujinluo thanks for joining
14:01:38 <mlavalle> hi lujinluo. Nice to see you!
14:01:42 <ihrachys> TuanVu, o/
14:01:50 <TuanVu> Hi Ihar :)
14:01:53 <lujinluo> mlavalle: hi! nice to see you here!
14:02:02 <TuanVu> Hi Luo
14:02:04 <mlavalle> hi TuanVu
14:02:08 <TuanVu> Hi mlavalle
14:02:41 <hungpv> hello
14:03:16 <mlavalle> hi hungpv
14:03:21 <ihrachys> we have the usual technical stuff to discuss today + one organizational topic. I suggest we don't hold the latter and discuss it straight away
14:03:37 <ihrachys> #topic Chair
14:04:49 <ihrachys> I may no longer be able to spend much time on Neutron in the near future (think month or so)
14:05:03 <ihrachys> because I am moving to another project (out of realm of OpenStack)
14:05:17 <lujinluo> :( soooo sad to hear that
14:05:29 <ihrachys> that's not going to happen overnight or tomorrow, but that's going to happen
14:05:32 * mlavalle said the same. still in shock
14:06:06 <ihrachys> I am sorry for letting you down
14:06:10 <TuanVu> well, not happy to hear that
14:06:21 <hungpv> :( so sad
14:06:24 <ihrachys> anyway, we need to set everything correctly so that there is no disruption in what we are working on
14:06:50 <TuanVu> I really appreciate your great help so far, Ihar. Not only for me but also for the whole project and other members
14:06:52 <mlavalle> ihrachys: you never let us down. it's just that we love you in Neutron, so we don't want to see you go. But we undrstand it is good for you
14:07:04 <lujinluo> ++
14:07:24 <mlavalle> Make us proud in your next project!
14:07:29 <ihrachys> and first obvious question is, assuming I no longer am able to pay much time to the project, we need to pass some batons I carried till now to responsible folks
14:07:42 <ihrachys> I already discussed a bunch of 'batons' with mlavalle several days ago
14:08:15 <ihrachys> one baton we don't have anyone committed to yet is leading this group / chairing the meeting
14:09:29 <ihrachys> and if it fits your plans lujinluo I would like to ask you to take over the role since you are probably the most involved and experienced person in this group. what do you think?
14:09:43 <mlavalle> ++
14:10:11 <TuanVu> ++
14:10:21 <ihrachys> again, I am not going away over night, and I am here to help you with whatever you need in next month or so.
14:10:50 <mlavalle> and I will lend a hand whenever you think it is necssary
14:11:18 <lujinluo> I see. I'd love to step up and it would be really appreciated if I could get some necessary help from mlavalle or ihrachys when needed
14:11:19 <mlavalle> This team is very importnt to Neutron so we want to keep it in good hands
14:11:52 <lujinluo> thanks for trusting me. i am surprised (in a good way)
14:11:57 <ihrachys> the good news is that I am a terrible leader anyhow, so it should be easy peasy for you to do x10 better. :)
14:12:28 <lujinluo> LOL, no way. You are a wonderful leader and helpful reviewer ihrachys
14:12:36 <mlavalle> ++
14:13:02 <ihrachys> lujinluo, great! I will propose the necessary changes for irc meetings repo to have you a co-chair, and we can discuss when you feel comfortable to take it over.
14:13:21 <lujinluo> understood.
14:14:23 <ihrachys> btw with me eventually gone, it could also make sense to move the meeting to some better time for Asia since everyone else is there. that will be your decision to make of course.
14:15:00 <mlavalle> Absolutely
14:15:03 <ihrachys> lujinluo, thanks for stepping in, I really appreciate (I am sure mlavalle does too)
14:15:23 <mlavalle> of course I am delighted that you are stepping up to the plate
14:15:39 <TuanVu> I believe Luo will be a great leader :)
14:16:07 <lujinluo> again, I am really surprised that you choose to believe in me. I feel like more motivated for work tmr, LOL
14:16:31 <mlavalle> :-)
14:17:04 <ihrachys> yes. another thing related to me stepping down at some point is that the group will loose the only focused core developer with ability to merge. mlavalle will need to look closer at what can be done with it. it may be giving coreship to someone in this group, or assigning someone from outside the group to support the effort.
14:17:22 <ihrachys> lujinluo, I am surprised you are surprised but ok!
14:17:57 <mlavalle> I will keep a close eye on reviews and I will help to develop cores in this area
14:18:15 <lujinluo> yes. and we should probably shout out to slawek more often
14:18:20 <lujinluo> ;)
14:19:07 <ihrachys> yeah that could work
14:19:30 <ihrachys> ok enough of this topic, let's talk code now :)
14:19:40 <mlavalle> yeah!
14:19:59 <ihrachys> #topic OVO
14:20:34 <ihrachys> I guess the first thing to discuss should be https://review.openstack.org/#/c/553617/ since it was blocking a bunch of patches
14:20:44 <ihrachys> it's "objects: automatically detect whether engine facade is used"
14:20:57 <TuanVu> yes
14:21:02 <ihrachys> that would allow objects to use the correct way to nest subtransactions depending on runtime context
14:21:27 <lujinluo> we got new comments from Michael
14:22:12 <ihrachys> we may want to have it moved to autonested_transaction at some point to broaden the scope to outside OVO since it seems like a good approach to stop carrying about the new and old ways of nesting, not just for OVO.
14:23:12 <mlavalle> that's a good point
14:23:27 <lujinluo> yeah
14:23:47 <ihrachys> yeah I see Michael's comments, posted just now
14:25:02 <ihrachys> I will need to dive in but looks like he is not happy about it in general. which means I will need to rework it. One thing he noticed there is that refresh() doesn't really refresh relationships. which is very surprising to me. (I am sure this detach code always assumed it.)
14:25:18 <ihrachys> seems like there is more work to do there.
14:26:35 <mlavalle> it did well with zuul, though
14:27:23 <ihrachys> yeah that's a good part. but maybe we overcomplicate things there with the whole refresh / expunge thingy
14:27:42 <ihrachys> I may need to talk to Michael about why we (think we) need the code
14:28:04 <mlavalle> I am going to rebase https://review.openstack.org/#/c/545501/ on top of it to see what happens
14:29:13 <ihrachys> mlavalle, yeah I guess it's worth experimenting like that. the general idea of the patch will stand, it's just that maybe we will need to rework sqlalchemy specifics.
14:29:25 <mlavalle> yeap
14:29:38 <lujinluo> agree
14:31:06 <ihrachys> ok I will spend some time on understanding what Michael feels bad about there :)
14:31:23 <ihrachys> let's now move to pure OVO patches
14:31:24 <ihrachys> https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/adopt-oslo-versioned-objects-for-db
14:31:52 <ihrachys> https://review.openstack.org/507772 "Use Network OVO in db_base_plugin"
14:32:23 <TuanVu> I'm still working on the answers for your concerns
14:32:31 <ihrachys> I believe TuanVu had some issues with the same detach code there. I tried to rebase the patch on top of mine and it seemed to have passed the previously failing test case.
14:33:04 <TuanVu> wao, really?
14:33:19 <ihrachys> so probably the test case TuanVu was most concerned about would be ok with the fix
14:33:30 <TuanVu> that's awesome!
14:33:38 <TuanVu> thank you so much, Ihar :)
14:33:40 <lujinluo> good to hear that. TuanVu seems to be struggling with the detach codes for a long time
14:33:43 <ihrachys> TuanVu, yeah, I dropped you an email 15 minutes before the meeting
14:34:15 <TuanVu> thank you Ihar, I'm not at work at the moment, so I cannot check that email right now
14:34:36 <ihrachys> np
14:34:53 <TuanVu> however, I have a question
14:34:58 <ihrachys> shoot
14:35:14 <TuanVu> https://review.openstack.org/#/c/507772/39/neutron/objects/base.py@320
14:36:06 <ihrachys> wait. I now realize that I may have tested with the exception handler. sec.
14:37:05 <ihrachys> eh sorry TuanVu I am dumb, I haven't checked that the handler was still there. it still fails.
14:37:19 <TuanVu> ok, and one more note: the test (queries constant) was modified in my last patch set
14:37:21 <ihrachys> now take back all the praise and thanks you gave me
14:37:38 <TuanVu> haha, no problem, Ihar
14:37:57 <ihrachys> right, you are talking about https://review.openstack.org/#/c/507772/39/neutron/tests/unit/db/test_db_base_plugin_v2.py
14:37:59 <TuanVu> no way I'm gonna take them back :))
14:38:06 <TuanVu> yes
14:38:41 <TuanVu> I've already sent you the explanation in my last email
14:39:18 <TuanVu> I've tried to remove the exception, and also tried to make sure new facade is being used
14:39:32 <TuanVu> however, there's still error:
14:39:39 <TuanVu> InvalidRequestError: Instance '<Network at 0x7f9654402f90>' is not persistent within this Session
14:40:08 <ihrachys> yeah I now see my fix doesn't make a dent there :-x
14:41:02 <TuanVu> maybe I'm not doing it correctly, I'm not sure :-?
14:41:23 <TuanVu> but as far as I understand, it's just
14:41:27 <TuanVu> - Network to "new_facade = True"
14:41:28 <TuanVu> - db_api.autonested_transaction(context.session): => cls.db_context_reader(context):
14:41:39 <TuanVu> that's all, am I right?
14:42:36 <ihrachys> yes that would work. with my patch merged it would not require any action on your side at all since OVO object would be aware of execution context (which way to open nested subtransaction was used above it)
14:43:26 <ihrachys> I now understand that you still have this issue even with the patch. I will consider your use case when working on the next patch set.
14:43:27 <TuanVu> however, as you've tested (rebased my patch on your patch), it's still not working, right?
14:43:52 <mlavalle> yeah, it's good input to Ihar's patch
14:44:36 <ihrachys> yeah the patch is clearly not ready and I will work on it today. I am hopeful Michael will have more to say when I explain him how we got there.
14:44:54 <TuanVu> ah, I got it!
14:45:01 <TuanVu> thank you in advance, Ihar :)
14:45:06 <TuanVu> I really appreciate it
14:46:20 <ihrachys> as for number of queries test change, I would still like us to understand better what triggers additional queries.
14:46:53 <TuanVu> yeah, I'll continue to dig deeper
14:48:10 <ihrachys> I am hesitant to touch the test case, it's meant to be a sign of a mistake somewhere. I saw you mentioned external network fetch triggering the additional queries. if so, maybe we don't eagerly load a relationship for external attribute. I would need to understand it better but let's focus on facade / detach failures first.
14:49:07 <ihrachys> ok next patch is https://review.openstack.org/#/c/549168/ "Use Router OVO in l3_db.py"
14:49:10 <ihrachys> it's all red in CI
14:49:11 <TuanVu> yeah, sure
14:49:28 <ihrachys> so I am not sure whether it makes sense to look at it.
14:50:58 <ihrachys> I don't see any other suspects in the review queue worth discussion. (it became rather calm in terms of new patches lately while we are resolving framework issues with facade and detach...)
14:51:07 <hungpv> yes, I have some issue
14:51:11 <ihrachys> does anyone have a patch asking for discussion?
14:51:15 <ihrachys> hungpv, go
14:51:38 <hungpv> maybe i'll send you an email to explain ihrachys?
14:52:17 <hungpv> i'm little tired today
14:52:22 <lujinluo> one update of A.I from last week. i pushed a follow-up of the new_facade UT fix https://review.openstack.org/#/c/553754/ and it is already merged (yay!)
14:52:54 <ihrachys> hungpv, well yeah but maybe better to keep discussions on gerrit (if there is a patch of course)
14:53:42 <TuanVu> congrats, Luo :)
14:53:49 <mlavalle> lujinluo: it went right in :-)
14:53:55 <ihrachys> TuanVu, that's actually true about emails you send me. it makes sense to have everything on gerrit, especially now that I may not have as much time for neutron in the near future.
14:54:20 <ihrachys> lujinluo, oh that's great.
14:54:21 <lujinluo> mlavalle: yeah. slawek has been very responsive
14:54:44 <ihrachys> heh. slaweq is an unsung hero of neutron :)
14:55:23 <TuanVu> @Ihar: yeah, I intended to post on Gerrit at first, however I forgot my password and password recovery was not working
14:55:30 <TuanVu> so ... I emailed to you
14:55:44 <TuanVu> I'll make sure to update on gerrit first next time :)
14:55:55 <ihrachys> TuanVu, no problem, I am just saying it will better to have it in public eye and for everyone to be able to chime in
14:56:08 <ihrachys> ok seems like we don't have anything else to discuss today
14:56:08 <mlavalle> ++
14:56:12 <hungpv_> @ihrachys: sure, I'll keep posting on Gerrit
14:56:22 <TuanVu> thank you Ihar, I got it :)
14:56:39 <mlavalle> TuanVu: you are in Ho Chi Minh City, right?
14:56:59 <ihrachys> thanks lujinluo for stepping in, I really appreciate. I will reach out to you in next week about our plans for passing the role
14:57:07 <TuanVu> yes, I'm in HCM city timezone
14:57:15 <lujinluo> ihrachys: understood!
14:57:36 <ihrachys> thanks folks!
14:57:38 <mlavalle> how about you lujinluo and hungpv_.... I what cities are you located?
14:57:39 <ihrachys> #endmeeting