17:33:36 <vilobhmm11> #startmeeting quotas-wg
17:33:37 <openstack> Meeting started Tue Apr 12 17:33:36 2016 UTC and is due to finish in 60 minutes.  The chair is vilobhmm11. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:33:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:33:40 <openstack> The meeting name has been set to 'quotas_wg'
17:33:42 <vilobhmm11> hi guys
17:34:04 <vilobhmm11> sorry i got stuck in traffic today and couldn't make it on time
17:34:49 <vilobhmm11> mc_nair, DuncanT, ninag : ping
17:46:36 <harlowja> vilobhmm11 yo yo
17:46:59 <vilobhmm11> harlowja : ping
17:47:10 <harlowja> let me know if want to chat about the cross-project quota thingy for this library
17:47:15 <harlowja> that i'm somehow the host of, lol
17:47:24 <vilobhmm11> harlowja : sure , wanted to discuss about the need for reservation
17:47:37 <harlowja> aka, https://www.openstack.org/summit/austin-2016/summit-schedule/events/9492?goback=1
17:47:52 <harlowja> i have a feeling that i might not be the best person to talk about that
17:47:56 <harlowja> buttttt vilobhmm11 u might be ;)
17:48:14 <vilobhmm11> harlowja : sure will do
17:48:27 <vilobhmm11> i can talk about it
17:48:57 <vilobhmm11> as i mentioned yesterday too; but it would be nice if you are around :)
17:49:00 <harlowja> :)
17:49:00 <harlowja> ya
17:49:01 <vilobhmm11> harlowja : ^^
17:49:59 <vilobhmm11> harlowja : about the quota engine
17:50:04 <harlowja> ?
17:50:16 <vilobhmm11> as per the spec it will take the project name and list of models as input
17:50:38 <raildo> Qijing, ^
17:50:50 <Qijing> yes
17:50:52 <vilobhmm11> can we just give the projects name and be able to get/construct the models
17:51:10 <vilobhmm11> i don;t think so; hence thought will need to pass in the models list as well
17:51:36 <vilobhmm11> harlowja : do you think passing models info is an unecessary peice of information ?
17:52:17 <vilobhmm11> harlowja : ^^
17:52:27 <harlowja> hmmm
17:52:36 <harlowja> i'd say make it possible, but don't require it
17:52:49 <harlowja> if people want to pass in there own model with X other things, sure, make it possible to do that
17:52:58 <harlowja> but i wouldn't require it
17:53:51 <vilobhmm11> harlowja : also regarding the existing state of transactions (when related to quota ops) in nova
17:53:55 <vilobhmm11> or cinder
17:54:57 <vilobhmm11> harlowja : There often isn't a single DB transaction that creates or
17:54:57 <vilobhmm11> updates or deletes something in Nova. API calls typically traverse 3 or
17:54:57 <vilobhmm11> more separate compute and controller node services, saving state
17:54:57 <vilobhmm11> transitions to a resource record sometimes *dozens* of times over the
17:54:57 <vilobhmm11> course of the API call -- many of which are entirely asynchronous.
17:55:17 <harlowja> yup
17:55:20 <harlowja> its sucky
17:55:29 <harlowja> so u have to have imho a concept of a seq_id then
17:55:41 <harlowja> and this seq_id u pass around
17:55:55 <harlowja> or remove those traversals if u can
17:56:41 <vilobhmm11> but in our case as well the quota engine relies on this infra ; quota engine needs to pass in some additional info to make it better
17:57:01 <vilobhmm11> harlowja : sure
17:58:11 <vilobhmm11> harlowja : but this seq-id needs to be kept track by the library right ?
17:58:50 <vilobhmm11> removing them is not that strightforward
17:59:20 <vilobhmm11> aw we are running out of time…will talk in some other channel
17:59:21 <vilobhmm11> #endmeeting