15:09:42 <rossella_s> #startmeeting neutron_upgrades
15:09:42 <sayalilunkad> hello!
15:09:42 <korzen> hello
15:09:42 <rossella_s> the bot is not working
15:09:42 <openstack> Meeting started Mon Jul  4 15:09:42 2016 UTC and is due to finish in 60 minutes.  The chair is rossella_s. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:09:43 <korzen> let's way a sec longer
15:09:43 <korzen> last time it had lags
15:09:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:09:44 <rossella_s> ok let's wait few more minutes then we will start
15:09:46 <openstack> The meeting name has been set to 'neutron_upgrades'
15:09:46 <rossella_s> maybe the bot is on holidays
15:09:46 <korzen> it has PTO
15:09:46 <korzen> or maybe you have to be Ihar to start the meeting?
15:09:47 <openstack> rossella_s: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
15:09:47 <rossella_s> it worked last time
15:09:47 <rossella_s> well let's start
15:09:48 <rossella_s> action from last time, "korzen to spin up a devref patch "
15:09:48 <rossella_s> korzen, did you start it?
15:09:48 <korzen> rossella_s,
15:09:48 <korzen> rossella_s, yes
15:09:49 <rossella_s> korzen, great! link?
15:09:49 <korzen> #link https://review.openstack.org/336518
15:09:49 <rossella_s> korzen, I will try to review it in the next day
15:09:50 <korzen> rossella_s, you cna wait until PS2
15:09:50 <rossella_s> korzen, ok
15:09:51 <rossella_s> korzen, do you know anything about the dvr granede job?
15:10:03 <rossella_s> the bot is back
15:10:07 <korzen> oh, the bot is alive
15:10:33 <rossella_s> #link https://review.openstack.org/336518 korzen's patch , devref for ovo
15:10:58 <rossella_s> Ihar is not here so we don't have any news regarding the grenade job
15:11:06 <rossella_s> #topic Object implementation
15:11:14 <korzen> I've some patch from Ihar
15:11:21 <korzen> let me check
15:11:56 <korzen> #link https://review.openstack.org/336793 Add an experimental grenade multinode job for Neutron Linuxbridge
15:12:07 <korzen> not a DVR
15:12:48 <korzen> or let's continue on OVO
15:12:51 <rossella_s> thanks korzen I missed that
15:13:02 <rossella_s> on my side no news for ovo, korzen sayalilunkad how is it going?
15:13:31 <sayalilunkad> I have updated the sec group patch https://review.openstack.org/#/c/284738/ again
15:14:07 <korzen> sayalilunkad, yes so I have one question, how does the is_default handled in neutron code base?
15:14:27 <korzen> sayalilunkad, would we need is_default when doing the get_objects? filtering?
15:14:43 <sayalilunkad> korzen: so basically we set is_default only while creating the object
15:15:06 <sayalilunkad> and we put it as non updatable because we have no way to update it using the api
15:15:23 <sayalilunkad> and when this is set as true the default group entry should be created
15:15:50 <sayalilunkad> we probably won't need is_default while doing get_objects
15:16:25 <sayalilunkad> it's just a way to add the default group for now as far as I understand
15:17:21 <korzen> sayalilunkad, I did not have time to look at the usage patch for SG, but from my understanding current implementation for get_objects will return all SG as is-default False
15:17:47 <rossella_s> korzen, why? is_default is part of the synthetic fields isn't it?
15:17:57 <rossella_s> I should also look at the patch again
15:18:29 <korzen> rossella_s, ok, when the field is in synthetic field, then we do not fill it
15:18:43 <korzen> ok, that make sense
15:19:00 <sayalilunkad> yup
15:19:17 <korzen> ok thanks, rossella_s and sayalilunkad
15:19:30 <sayalilunkad> there are still some failures that need to be fixed mainly regarding forign key contraints
15:20:07 <sayalilunkad> but another concern I had was if we need to create a security group before we run the DefaultSecurityGroupDbTestCase stuff
15:21:18 <sayalilunkad> and ihar suggested using utils.AuthenticIPNetwork here https://review.openstack.org/#/c/284738/41/neutron/objects/port/extensions/securitygroups.py
15:21:29 <sayalilunkad> butI am not sure how that would help really
15:21:32 <rossella_s> sayalilunkad, to solve the foreign key constraint you need to create all object in the db that are foreign keys
15:22:44 <sayalilunkad> rossella_s: I see..so using mock?
15:23:20 <rossella_s> sayalilunkad, not really if you look at the setup method of the tests we created the object that are needed, for example network if you are testing a port
15:23:23 <korzen> sayalilunkad, the AuthenticIPNetwork is helpful when API tests are expecting the same format on output as on input
15:23:30 <rossella_s> sayalilunkad, no sure what's your specific failure
15:24:05 <korzen> sayalilunkad, the example of failing unit test when format is not the same: http://logs.openstack.org/56/287756/9/check/gate-neutron-dsvm-api/878efe5/testr_results.html.gz
15:24:14 <korzen> sayalilunkad, or API tests
15:25:00 <sayalilunkad> korzen: but using the utils stuff didn't seem to change anything
15:25:14 <sayalilunkad> the format was the same as before
15:25:31 <korzen> sayalilunkad, I guess that there is no unit test to see the difference
15:25:46 <sayalilunkad> rossella_s: thanks, ill take a look
15:26:27 <sayalilunkad> korzen: ok I thought it would be seen when we run the existing tests as the input parameters or something
15:26:29 <rossella_s> korzen, anything else?
15:26:36 <sayalilunkad> but I'll take a look
15:26:38 <sayalilunkad> thanks!
15:27:05 <korzen> ok, from my side I was working on getting the unit test passing for Subnet OVO
15:27:49 <korzen> the interesting thing was filtering by tenant_id when we have the project_it defined in Subnet object
15:28:32 <korzen> I have added the ability to filter both by tenant_id and project_id, but we decided with Ihar that filtering by tenant_id should be dropped
15:28:38 <rossella_s> korzen, yes I was looking briefly into it
15:28:39 <korzen> dropped in the object*
15:28:56 <rossella_s> korzen, yep I recall some irc conversation
15:28:59 <rossella_s> right?
15:29:17 <korzen> the solution should be to change the tenant_id before passing filters to object
15:29:23 <korzen> yes
15:30:22 <korzen> I have to synch with dasm on approach how to convert the filters before passing kwargs to object
15:31:21 <korzen> I have also worked out the way of adding new extension
15:31:49 <korzen> it is not possible to add an arbitrary fields to subnet dict in runtime anymore
15:32:24 <rossella_s> I see
15:32:35 <korzen> so I have changed the unit tests to temporary define new field in subnet object, fetch the data and fill in the output dict
15:32:39 <rossella_s> anything else?
15:33:28 <korzen> this week I will be working on splitting the patches for subnet object adjustment
15:33:41 <korzen> and addressing the comments in my current patches
15:34:55 <rossella_s> #topic Other patches on review
15:36:12 <rossella_s> I don't have anything to highlight here, korzen sayalilunkad ?
15:36:31 <korzen> I guess that there are new objects ported, dns: https://review.openstack.org/#/c/334695
15:36:46 <korzen> the unique keys to be used in get_object by ivc
15:37:21 <korzen> #link https://review.openstack.org/322024
15:37:24 <rossella_s> #link https://review.openstack.org/322024 this one?
15:37:30 <korzen> yes
15:38:26 <korzen> I do not have noting more in mind
15:38:40 <rossella_s> let's move on
15:38:42 <sayalilunkad> me neither
15:38:52 <rossella_s> #topic Open discussion
15:39:11 <rossella_s> Ihar proposed a blueprint regarding online upgrades
15:39:17 <rossella_s> #link https://blueprints.launchpad.net/neutron/+spec/online-upgrades
15:40:05 <korzen> nice
15:40:11 <sayalilunkad> that sounds good
15:40:49 <rossella_s> the submission of the talk for the barcelona summit is open
15:40:53 <rossella_s> #link https://www.openstack.org/summit-login/login?BackURL=%2Fsummit%2Fbarcelona-2016%2Fcall-for-presentations%2F
15:41:33 <korzen> to anyone that is going to midcycle, I have put the upgrades on agenda
15:41:59 <korzen> rossella_s, will you be traveling?
15:42:04 <rossella_s> I guess it will be you and Ihar, korzen ...if Ihar's visa processing is fast enough
15:45:55 <korzen> ok I do not have anything more
15:50:51 <korzen> rossella_s, are we done here? end meeting?
15:51:38 <rossella_s> korzen, yes
15:51:49 <rossella_s> #endmeeting