16:59:59 <dansmith> #startmeeting nova_cells
16:59:59 <openstack> Meeting started Wed Dec  6 16:59:59 2017 UTC and is due to finish in 60 minutes.  The chair is dansmith. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:02 <openstack> The meeting name has been set to 'nova_cells'
17:00:05 <mriedem> o/
17:00:37 <dansmith> well, this is quaint
17:00:43 <mriedem> melwitt: tssurya?
17:00:47 <dansmith> ssshh
17:00:51 <melwitt> gah
17:00:59 <melwitt> I was distracted
17:01:07 * tssurya_ waves
17:01:11 <dansmith> #topic bugs
17:01:14 <dansmith> sooo,
17:01:31 <dansmith> my patch for belmiro's bug isn't passing the cellsv1 gate job because reasons
17:01:49 <dansmith> mriedem figured out why (we think), but making it work is going to take more time
17:02:00 <dansmith> I'm putting it at the bottom of my priority list for the moment, given the other things
17:02:02 <dansmith> but it's there
17:02:07 <dansmith> related to that,
17:02:15 <melwitt> legit fail? that's surprising considering it should be transparent I had thought
17:02:20 <dansmith> legit, yeah
17:02:26 <dansmith> while writing tests, I figured out that one of the existing ones is broken
17:02:43 <dansmith> like, throwing an exception all the time, which is swallowed by the dispatch logic
17:02:50 <dansmith> fix for that is here: https://review.openstack.org/#/c/525304/
17:03:01 <dansmith> it's a simple refactor, worth getting in before we lose track
17:03:27 <melwitt> guh
17:03:55 <dansmith> anyway, any other bugs?
17:03:58 <mriedem> there are a few more bugs in https://wiki.openstack.org/wiki/Meetings/NovaCellsv2#Agenda
17:04:12 <mriedem> looks like melwitt really hates https://review.openstack.org/#/c/524641/
17:04:35 <mriedem> heh, everyone else hates https://review.openstack.org/#/c/524755/
17:04:41 <melwitt> lol yes, that was "hate"
17:04:54 <tssurya_> :D
17:04:55 <dansmith> ah, I hadn't seen that
17:05:23 <dansmith> I'll circle back to that other one of mine
17:05:26 <melwitt> that's a lot of firey red -1s
17:05:39 <mriedem> i haven't dug into why yet
17:05:47 <mriedem> but will hold off on that until hongbin updates it
17:06:05 <tssurya_> for starters it doesn't have a unit test case
17:06:19 <mriedem> yeah that's usually a problem
17:06:46 <mriedem> i think that's it for bugs
17:06:59 <dansmith> #topic open reviews
17:07:08 <dansmith> kinda on this topic anyway
17:07:27 <dansmith> I unpinned the bottom of the alternates thing so we can merge the object and the first few patches there
17:07:45 <mriedem> there are 5 patches in the gate i think :)
17:07:57 <dansmith> not in the gate
17:07:59 <mriedem> but yeah, ed has to update some stuff
17:08:01 <dansmith> but headed there I guess
17:08:11 <mriedem> there are 3 remaining changes,
17:08:20 <mriedem> the bottom one is pretty simple to get fixed and passing functoinal tests
17:08:27 <mriedem> middle patch is +2 - just a refactor,
17:08:35 <mriedem> last one actually starts using alternates for building instances
17:08:39 <mriedem> and i have some issues with that one
17:08:43 <mriedem> not insurmountable though
17:08:55 <mriedem> and he's working on a patch not yet posted for doing alternates with resize
17:09:15 <dansmith> okay
17:09:19 <mriedem> if we can get the build + alternates very close by tomorrow i think we're in good shape
17:09:24 <mriedem> tomorrow is q-2
17:09:36 <dansmith> oh boy
17:10:02 <mriedem> in other review news,
17:10:08 <mriedem> i see the genericify instance list stuff is all merged
17:10:14 <mriedem> i need to get back to yikun's changes that used it
17:10:17 <dansmith> cool
17:10:33 <mriedem> but hopefully his stuff will all be much more sane now
17:10:43 <dansmith> yeah,
17:10:46 <dansmith> I'm pretty happy with how that worked out
17:10:51 <mriedem> me too
17:10:54 <dansmith> I was expecting it to be a lot harder and more ugly
17:10:55 <mriedem> glad you did it
17:11:00 <dansmith> yeah
17:11:05 <dansmith> anything else on reviews?
17:11:14 <tssurya_> I have two not so important reviews
17:11:20 <tssurya_> I will update them in the agenda
17:11:23 <tssurya_> later
17:11:37 <melwitt> what's the agenda? like the wiki?
17:11:37 <mriedem> https://review.openstack.org/#/q/owner:suryaseetharaman.9%2540gmail.com+status:open
17:11:40 <mriedem> those ^ ?
17:11:45 <tssurya_> yes
17:11:54 <melwitt> k cool
17:11:55 <dansmith> #link https://review.openstack.org/#/q/owner:suryaseetharaman.9%2540gmail.com+status:open
17:12:22 <mriedem> melwitt: i've just been posting bugs in the meeting agenda wiki yeah
17:12:26 <mriedem> since we had a few in triage last week
17:12:38 <melwitt> oh, got it
17:12:39 <dansmith> #topic open discussion
17:12:49 <tssurya_> I have one - https://bugs.launchpad.net/nova/+bug/1726301
17:12:49 <openstack> Launchpad bug 1726301 in OpenStack Compute (nova) "Nova should list instances even if it can't connect to a cell DB" [Medium,Confirmed] - Assigned to Surya Seetharaman (tssurya)
17:13:03 <tssurya_> I know dansmith and belmiro had a discussion during the summit
17:13:08 <tssurya_> regarding this
17:13:28 <tssurya_> just wanted to know if dansmith is going to work on this or if I should have a look :)
17:13:58 <dansmith> tssurya_: it's lower on the list for me than other things, but it's also not super trivial
17:14:07 <dansmith> tssurya_: however, if you want to start looking at it, that's cool
17:14:20 <tssurya_> dansmith : okay!
17:14:25 <melwitt> I think it depends on what we want to show for a down cell instance right?
17:14:36 <melwitt> or did we already come up with something
17:14:56 <dansmith> melwitt: that's one of the reasons I haven't done anything, because it's going to involve a bunch of api semantic discussion :)
17:15:24 <mriedem> hmm yeah, either don't list them, or list instance mappings which only has uuid
17:15:25 <mriedem> or other
17:15:25 <dansmith> tssurya_: https://github.com/openstack/nova/blob/master/nova/compute/multi_cell_list.py#L243
17:15:34 <dansmith> not listing them is wrong
17:15:56 <melwitt> and there was talk of the fallback DB repllica too
17:16:04 <melwitt> *replica
17:16:06 <dansmith> we need to list them, but have some way to explain why only the uuid is provided
17:16:11 <dansmith> yep
17:16:25 <mriedem> "SOL" for all response parameters
17:16:45 <dansmith> fine with me
17:16:45 <mriedem> or that 'idk' ascii guy
17:16:46 <dansmith> or just system('/usr/bin/fortune') for each
17:16:59 <melwitt> thumbs up to that
17:17:07 <mriedem> ¯\_(ツ)_/¯
17:17:26 <dansmith> anyway, tssurya_ you can work on catching that scenario, doing something (anything) and then we can use that as the discussion point
17:17:36 <mriedem> https://i.ytimg.com/vi/crPl0ITIkS0/hqdefault.jpg ?
17:17:38 <dansmith> def fill_missing_instance(): # TODO
17:17:45 <tssurya_> hmm, okay I will dig deeper, meanwhile if there are other tasks I can assist with let me know
17:18:09 <mriedem> is this worth starting a thread in the ML now? or premature?
17:18:22 <mriedem> raising the issue now might freak everyone out
17:18:32 <dansmith> mriedem: I've got a week left in the year, so if you start it now, you finish it without me :)
17:18:45 <mriedem> january it is
17:18:55 <dansmith> so on that topic,m
17:18:59 <mriedem> honestly it's probably ptg fodder at this point
17:19:09 <dansmith> I'll be here next weds for this meeting, but will be gone after that for the rest of the year, just FYI
17:19:16 <dansmith> I'd propose we cancel all meetings after next weds
17:19:22 <mriedem> works for me
17:19:25 <tssurya_> cool
17:19:26 <mriedem> i'm out the last week of the year
17:19:27 <dansmith> because I know you can't possibly have this meeting without me
17:19:45 <melwitt> we don't want to, no :)
17:19:48 <dansmith> mriedem: +1 for ptg
17:20:04 <melwitt> agree
17:20:06 <dansmith> aight, anything else for open discussion here/
17:20:11 <tssurya_> nope
17:20:22 <mriedem> nope
17:20:25 <dansmith> #endmeeting