21:04:31 <alaski> #startmeeting nova_cells
21:04:31 <openstack> Meeting started Wed Dec  9 21:04:31 2015 UTC and is due to finish in 60 minutes.  The chair is alaski. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:04:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:04:33 <itxaka> bye bye o/
21:04:35 <openstack> The meeting name has been set to 'nova_cells'
21:05:01 <alaski> anyone around today?
21:05:05 <melwitt> o/
21:05:07 <belmoreira> o/
21:05:08 <dansmith> o-
21:05:22 <alaski> #topic Testing
21:05:26 <alaski> melwitt: any word?
21:05:43 <alaski> btw, I don't have much today, hoping to be quick
21:06:02 <melwitt> alaski: not aware of anything like failures but I found a gap in our tempest testing for cells that I'm trying to address
21:06:36 <alaski> okay
21:06:51 <melwitt> I think we don't have coverage of testing network connectivity for instances booted during the job. that used to be covered by the devstack exercise job
21:07:12 <dansmith> heh
21:07:16 <dansmith> seems unimportant
21:07:23 <alaski> heh
21:07:36 <melwitt> but the tempest scenarios all use floating ip and/or security groups during the ssh test, so I'm looking at providing an option to specify security group name instead of creating on the fly
21:07:49 <melwitt> we're blacklisting all the tests that do ssh I believe
21:08:03 <alaski> gotcha
21:08:27 <alaski> the default security group should work as well I think
21:08:50 <melwitt> yeah, that's what I'm aiming for ultimately
21:09:27 <alaski> cool
21:10:02 <alaski> if I can help let me know
21:10:16 <alaski> #topic Open Reviews
21:10:29 <alaski> just want to call out https://etherpad.openstack.org/p/mitaka-nova-priorities-tracking here
21:10:39 <belmoreira> anything blocking https://review.openstack.org/#/c/201606/ and https://review.openstack.org/#/c/213041/ ?
21:10:43 <alaski> please add any open patches under the cellsv2 section
21:11:34 <alaski> belmoreira: I think it's just lack of reviews
21:11:47 <alaski> I will go through them again
21:12:17 <belmoreira> alaski: thanks
21:12:54 <alaski> #topic Open Discussion
21:13:53 <alaski> I'm starting to go through and clean up the boot process in preparation for actually persisting the request spec there
21:14:05 <alaski> and then calling the scheduler from nova-api/conductor from nova-api
21:14:43 <alaski> that will be the first work that starts to require the api db, so I'm hoping to get some attention on it once it's up
21:15:32 <alaski> anyone have a topic for open discussion?
21:15:54 <melwitt> I was just thinking I need to test out the db connection switch patch more
21:16:31 <melwitt> currently there's just the functional test, which I improved by not using the in-memory sqlite for the cell db
21:16:48 <alaski> nice
21:17:44 <alaski> added that to the reviews etherpad
21:17:57 <alaski> that's another one I want to go through again
21:18:28 <melwitt> thanks. the main change is I changed it not to keep connections open to every cell, or not try to
21:19:03 <melwitt> I didn't know if that's okay, i.e. is it very costly to do a new connect each time and teardown. I thought maybe with pooling underneath that's in oslo.db it would be okay
21:19:23 <melwitt> so input on that would be appreciated in the review
21:19:32 <dansmith> jaypipes always said that a new connection to mysql is very fast
21:19:48 <dansmith> like, it was designed that way and thus connecting and disconnecting _should_ be done aggressively
21:20:09 <dansmith> now, that doesn't mean we should rely on that, given that we support other databases, but ...
21:20:16 <melwitt> I see
21:20:19 <alaski> that's good info
21:20:29 <alaski> this is probably something we'll need to offer some tuning knobs for
21:21:30 <alaski> and it would be good to get some input from db experts, like jaypipes, on the review
21:21:40 <melwitt> I was thinking not to prematurely optimize by trying to maintain an open connection cache with expiry when I don't yet know if it's necessary
21:21:53 <melwitt> yeah, agree
21:22:35 <alaski> yeah, start simple
21:23:01 <alaski> anything else?
21:23:56 <alaski> cool.  back to work everyone!  or relax, depending on timezone
21:24:01 <alaski> thanks
21:24:05 <alaski> #endmeeting