21:00:01 <dansmith> #startmeeting nova_cells
21:00:01 <openstack> Meeting started Wed Apr  4 21:00:01 2018 UTC and is due to finish in 60 minutes.  The chair is dansmith. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:06 <openstack> The meeting name has been set to 'nova_cells'
21:00:10 <dansmith> hoo dis
21:00:18 <tssurya> o/
21:00:52 <dansmith> hmm, lonely
21:01:48 <melwitt> o/
21:01:50 <melwitt> sorry yall
21:02:09 <dansmith> aight well, maybe mriedem is busy with other stuff, which is fine
21:02:12 <dansmith> #topic bugs
21:02:19 <dansmith> new bug opened today: https://bugs.launchpad.net/nova/+bug/1761197
21:02:21 <openstack> Launchpad bug 1761197 in OpenStack Compute (nova) "Not defined keypairs in instance_extra cellsV1 DBs" [Undecided,New] - Assigned to Surya Seetharaman (tssurya)
21:02:31 <dansmith> I have looked at it a tiny bit, but not in detail
21:02:50 <dansmith> the need for another type of migration from api_db for cellsv1 people is probably a thing we need, which sucks, but.. whatever
21:02:53 <dansmith> the more concerning thing to me,
21:03:19 <dansmith> is the secondary issue that instances aren't being created with instance_extra.keypairs in the child cells
21:03:37 <tssurya> we are still investigating it as well, we might just do a downstream tool to recreate the keypairs
21:03:37 <mriedem> o/
21:03:41 <melwitt> hmm
21:03:52 <dansmith> I haven't spent much time on that, and getting tangled up in the cells code made me want to do something else
21:04:09 <dansmith> tssurya: okay well, I think we probably need to upstream something for the migration bit because you're not the only one that will hit that
21:04:15 <melwitt> guess that implies we're missing a target_cell somewhere
21:04:22 <dansmith> tssurya: but yeah, if you can investigate the creation issue too that would be good
21:04:26 <dansmith> melwitt: I don't think so
21:04:37 <dansmith> melwitt: we create that stuff when we create the instance, and flavor is getting set
21:04:42 <tssurya> dansmith: yes , it is of priority to us, so will work on this and find the source
21:04:57 <dansmith> melwitt: I think it's more likely a filterage or recreating-the-object-in-cells-rpc-land sort of thing
21:05:01 <dansmith> tssurya: cool
21:05:21 <melwitt> okay
21:05:30 <dansmith> so anyway, yeah, I don't have much else to say about that bug
21:05:39 <dansmith> any other new bugs we need to discuss?
21:05:43 <tssurya> https://bugs.launchpad.net/nova/+bug/1761198
21:05:44 <openstack> Launchpad bug 1761198 in OpenStack Compute (nova) ""Orphan" request_specs and instance_mappings" [Undecided,New] - Assigned to Surya Seetharaman (tssurya)
21:05:59 <tssurya> we might need something to clean the remaining req_specs
21:06:04 <tssurya> if an archival doesn't work
21:06:15 <tssurya> orphaned ones from before too
21:06:19 <dansmith> or for things from before
21:06:22 <tssurya> yes
21:06:38 <dansmith> tssurya: I think that adding yet another cleanup step is probably less desirable,
21:06:42 <mriedem> this is a pre-req for using request specs in the API if a cell is down?
21:06:53 <dansmith> so it might be good to roll that into something else like archive itself or something
21:07:00 <dansmith> or online_data_migrations
21:07:06 <tssurya> dansmith: cool
21:07:28 <tssurya> which is preferable ?
21:07:32 <dansmith> mriedem: acutely I guess, but it's also a hygiene thing
21:07:53 <dansmith> tssurya: I dunno, I'd have to think and look.. it's going to expensive
21:07:55 <mriedem> once you get to queens, and archive does this for you,
21:07:58 <mriedem> then it's no longer an issue right?
21:08:09 <dansmith> mriedem: shouldn't be, except for the older ones
21:08:15 <mriedem> in that case,
21:08:19 <mriedem> i'd roll it into online_data_migrations
21:08:19 <melwitt> yeah cause you have to verify no instance exists per instance_mapping and request_spec before hard-deleting rihgt
21:08:26 <tssurya> mriedem: yes, except for the ones we are already carrying
21:08:29 <dansmith> mriedem: yeah that makes it easily removable
21:08:54 <mriedem> lee was going to add a similar cruft cleanup thing for bdms
21:08:57 <mriedem> per the ptg talk
21:09:05 <mriedem> before we can add a unique constraint on that able
21:09:08 <mriedem> *table
21:09:24 <dansmith> yeah, reasonable I think
21:09:26 <tssurya> okay I will head in that direction then
21:09:32 <dansmith> okay, any other bugs?
21:09:35 * dansmith glares at tssurya
21:09:37 <tssurya> nope
21:09:39 <tssurya> :D
21:09:41 <dansmith> good. :P
21:09:57 <dansmith> #topic open reviews
21:10:04 <dansmith> so, tssurya has a billion of these open
21:10:11 <dansmith> and I've only reviewed like half a million of them
21:10:16 <dansmith> which has me feeling inadequate
21:10:30 <dansmith> but, I get the sense they're seeing decent review from people
21:10:37 <tssurya> haha, no cores a doing a great job in this team
21:10:49 <tssurya> or this sub-team at least :)
21:10:55 <dansmith> my placement-req-filter stuff is all mergificated, which is cells-related,
21:11:04 <melwitt> I have this one that needs a +W https://review.openstack.org/#/c/540258/ to make the AffinityFilters work with multiple cells
21:11:28 <dansmith> and will be pushing on jay bird to get the mirroring spec done and maybe help him with implementation, which will be the last nail for cern to be able to try this reasonably
21:11:32 <mriedem> i looked at that when it was being written, i can go back to it
21:11:46 <mriedem> i think jay is waiting on me for that spec
21:12:03 <mriedem> and i am,
21:12:03 <dansmith> melwitt: also, Einst Crazy is good with that one, which is a major mark of quality
21:12:06 <mriedem> dispassionate at best
21:12:23 <dansmith> mriedem: you don't have to if you don't want to, but ... :)
21:12:30 <melwitt> heh
21:12:36 <mriedem> but smiley face?
21:12:59 <dansmith> mriedem: but it'd help :)
21:13:26 <dansmith> that was me trailing off into just a guilt-laden grin
21:13:33 <mriedem> i just want to be able to say i told you so generically in a few years when the mirroring bites us in the ass somehow
21:14:03 <dansmith> well, if you really feel like that, then voice it on the spec
21:14:09 <dansmith> anyway,
21:14:09 <mriedem> i think i did
21:14:12 <mriedem> yes anyway
21:14:17 <dansmith> well, then -something it then
21:14:21 <dansmith> maybe MINUS FIIIVE
21:14:25 <tssurya> :D
21:14:26 <mriedem> heh
21:14:30 <dansmith> other open reviews?
21:14:43 <tssurya> no
21:14:57 <dansmith> tssurya: jeez tssurya, you're not the only one with reviews.. let others speak
21:15:07 * dansmith looks around the room
21:15:14 <dansmith> okay maybe you are :D
21:15:21 <melwitt> I already said mine
21:15:34 <dansmith> heh
21:15:38 <dansmith> #topic open discussion
21:16:44 <mriedem> nada
21:16:45 <dansmith> crickets?
21:17:14 <melwitt> crickets
21:17:26 <dansmith> okay that means I can put on my bikini and go wash the car
21:17:41 <melwitt> \:|
21:17:42 <dansmith> #endmeeting