15:01:42 <yamahata> #startmeeting neutron_northbound
15:01:42 <odl_meetbot> Meeting started Mon Jul 31 15:01:42 2017 UTC.  The chair is yamahata. Information about MeetBot at http://ci.openstack.org/meetbot.html.
15:01:42 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:42 <odl_meetbot> The meeting name has been set to 'neutron_northbound'
15:02:00 <yamahata> #topic agenda bashing and roll call
15:02:12 <yamahata> #link https://wiki.opendaylight.org/view/NeutronNorthbound:Meetings
15:02:14 <yamahata> #info yamahata
15:02:16 <rajivk_> #info rajivk
15:02:28 <mkolesni> #info mkolesni
15:02:37 <manjeets_> #info manjeets
15:02:47 <mkolesni> we have a new comer to the project, mpeterson
15:02:48 <yamahata> #chair rajivk mkolesni manjeets
15:02:48 <odl_meetbot> Current chairs: manjeets mkolesni rajivk yamahata
15:03:06 <yamahata> mpeterson: welcome!
15:03:06 <manjeets_> cool welcome to the gang mpeterson
15:03:25 <mkolesni> mpeterson, would you like to introduce yourself?
15:03:39 <mpeterson> thanks :) glad to be here
15:03:53 <mpeterson> mkolesni: sure
15:04:46 <yamahata> mpeterson: do you have any specific interest or area?
15:04:48 <mpeterson> I'm from Argentina and moved to Israel around 2 years ago, 31 years old, being developing since I was around 12 or so...
15:05:13 <mpeterson> yamahata: I love python particularly and mostly backend :)
15:05:33 <yamahata> Cool. pythonists are wanted.
15:05:50 <mkolesni> :)
15:05:55 <rajivk_> mpeterson, From 12.
15:06:20 <mpeterson> rajivk_: yeah, a nerd... started because I wanted to be in the informatics olympiads
15:06:37 <rajivk_> Good, so must be codechef, topcoders right??
15:08:07 <mpeterson> rajivk_: did some of them... but never catched my eye, I really liked competing for my high school and also I remember there were was a competition once a year that basically tested your ability to code outside the box and performance minded, I can't really remember the name though
15:09:18 <rajivk_> Good, let's continue.
15:10:08 <yamahata> okay, any topics to discuss today?
15:10:17 <mkolesni> were at feature freeze
15:10:30 <mkolesni> so what do we do with non bug fixes?
15:10:30 <yamahata> pike-3 was released.
15:10:49 <yamahata> Basically we'll postpone to queens.
15:10:59 <yamahata> Do you have any specific patches?
15:11:06 <yamahata> or features?
15:11:26 <mkolesni> i thought stable branch was being cut or is it happening on R+0?
15:11:53 <yamahata> The branch isn't cut yet.
15:12:02 <mkolesni> well i guess this hints at it "The requirements remain frozen until the stable branches are created, with the release candidates"
15:12:31 <mkolesni> ok so mpeterson started working on the ability to arbitrarily run full sync
15:12:40 <yamahata> The branch will be cut with release candidate
15:13:07 <yamahata> https://releases.openstack.org/pike/schedule.html#requirements-freeze
15:13:23 <mkolesni> i think it will be ok to push it to queens though
15:13:40 <mkolesni> also re CI i had no luck yet to figure out the problems of ODL on the gate
15:13:52 <yamahata> If we have important feature, it's okay to include feature. we can see feature exception procedure in openstack-dev mailing list.
15:14:16 <yamahata> #topic Announcements
15:14:19 <yamahata> pike-3 was released
15:14:26 <yamahata> #topic action items from last meeting
15:14:27 <yamahata> none
15:14:33 <yamahata> #topic Pike/Nitrogen planning
15:14:42 <yamahata> okay, mkolesni please go ahead.
15:15:01 <mkolesni> yamahata, re the CI?
15:15:08 <yamahata> CI and feature.
15:15:13 <mkolesni> ok
15:15:19 <yamahata> do you want to include runnig full-sync to pike?
15:15:35 <mkolesni> just quick question before, did you get confirmation to go to PTG?
15:15:42 <mkolesni> (before i forget)
15:16:08 <yamahata> oh yeah, I can attend PTG. it took a while to get approval due to vacation season.
15:16:44 <mkolesni> ok so youre definitely coming then?
15:16:48 <yamahata> yes.
15:17:21 <mkolesni> great, ill order the travel stuff then
15:17:32 <mkolesni> so for feature you guys can discuss if its important
15:17:54 <mkolesni> for us there's no rush to get it to Pike and we can wait for Queens with this full sync on demand
15:18:32 <rajivk_> mkolesni, can you please brief about full sync things, we are discussing?
15:18:34 <yamahata> I see.
15:20:19 * yamahata waiting for mkolesni
15:21:07 <mkolesni> just the ability to execute it on demand
15:21:19 <mkolesni> mpeterson is currently working to allow it
15:21:39 <mkolesni> the plan is to use SIGHUP signal to trigger a run of the full sync
15:21:54 <rajivk_> yamahata, put a patch something similar. if i remember correctly.
15:22:07 <mkolesni> its still being developed so i can't say for sure it will work but it looks promising
15:22:24 <yamahata> I uploaded it, but it turned out to be too easy.
15:22:37 <rajivk_> this one https://review.openstack.org/#/c/452023/
15:22:38 <mkolesni> yes that patch though is problematic because it doesnt take into account the situation where the full sync is already running
15:22:39 <yamahata> the patch needs more enhancement.
15:22:59 <mkolesni> also if it dies in the middle then what would you do
15:23:37 <rajivk_> okay
15:23:48 <mkolesni> we thought of alternatives and SIGHUP is easiest
15:24:09 <mkolesni> also its a quite common practice to utilize this signal in *nix systems
15:25:47 <mkolesni> mpeterson can elaborate
15:27:40 <yamahata> can we move on to CI?
15:27:42 <mpeterson> mkolesni: I'm not sure what to say in addition to that, as it is comprehensive. It does seem to be the most clean way to do it, it's what is expected from a *nix process and is the direction towards where all openstack processes are heading
15:28:23 <mkolesni> sure, for CI i have no update yet :/
15:28:44 <mkolesni> i cant reproduce the prominent problem from the gate locally because im hitting other problems
15:28:48 <yamahata> Hmm. anyway we need to make CI healthy until Pike release for neutron stadium requirement
15:28:51 <mkolesni> which might or might not be related
15:29:04 <yamahata> At least with carbon+tempest.
15:29:28 <mkolesni> the problem is that the failures are due to odl bugs
15:30:16 <yamahata> Wherever issues exists in openstack or ODL, we should chase them down.
15:30:29 <yamahata> by ourselves or with the help with netvirt folks.
15:30:44 <mkolesni> yes unfortunately i lack the expertise to debug ODL
15:30:56 <mkolesni> im trying to rope in resources but they seem to be busy
15:31:10 <mkolesni> though i believe theyre fixing bugs on the go as well
15:31:30 <yamahata> Yeah, they are working quite hard.
15:32:14 <yamahata> so far we have too many timeout, if timeout occurs, we don't get meaningful log.
15:32:17 <mkolesni> there are plenty of problems there even carbon sr1 is not very stable
15:32:33 <mkolesni> we could try to reduce the timeout
15:32:49 <yamahata> it's the first step to address log issue in order to obtain logs.
15:32:56 <mkolesni> i have this patch https://review.openstack.org/#/c/482546/
15:33:05 <mkolesni> it reduces test timeouts to 60 secs
15:33:22 <mkolesni> i didnt find a proper way to do this so i did it hackily for now
15:34:24 <manjeets_> may be disable some of the tempest tests related to volume which we don't care about ?
15:34:31 <yamahata> I suppose it's only short term issue. Once we have fixed real issues, we can remove timeout stuff.
15:34:43 <yamahata> So it's okay that the patch is hacky.
15:35:03 <mkolesni> ok i can fix it up a bit and add TODO and stuff
15:35:40 <mkolesni> manjeets_, we could definitely though FIPs fail generally and thats just causing them to fail
15:36:12 <mkolesni> but we could do that to reduce the execution time + the size of logs themselves
15:36:20 <manjeets_> FIP's are killing multiple jobs
15:36:34 <mkolesni> today i had real trouble opening karaf log since it's ~300MB on a failed job
15:37:05 <manjeets_> oh wow just karaf lob ~ 300 MB ?
15:37:07 <mkolesni> and the log collector does it in HTML making it necessary to open with some browser
15:37:12 <mkolesni> yes
15:37:38 <mkolesni> sorry 250MB
15:37:51 <mkolesni> but in the 100s of MB it doesnt matter its way too big
15:38:56 <yamahata> We can control log level more precisely with changing configure_opendaylight
15:39:40 <yamahata> around line 100
15:40:11 <yamahata> the log level can be revised if necessary.
15:40:13 <mkolesni> im not sure if all the trace logs are necessary for debug or not though
15:40:21 <mkolesni> we can try
15:40:25 <manjeets_> ++
15:40:56 <yamahata> it's also defined for legacy netvirt, so probably we'd like to update them for new netvirt.
15:41:04 <mkolesni> regarding test load i have a 'do not merge' patch to reduce it
15:41:07 <mkolesni> https://review.openstack.org/#/c/484638/3/devstack/devstackgaterc
15:41:19 <mkolesni> but its very agressive.. :)
15:41:30 <mkolesni> but we can take maybe a few parts from it..
15:41:42 <mkolesni> yamahata, can you look at the log thing then? adapt it to new netvirt?
15:42:11 <yamahata> Okay.
15:42:27 <yamahata> anyway it's default log level for new netvirt bundles.
15:43:01 <yamahata> we can just increase it to log more. = bigger log size.
15:43:29 <mkolesni> ok
15:44:54 <yamahata> anything else for CI.
15:45:04 <yamahata> how about grenade? manjeets ?
15:45:41 <yamahata> Regarding to grenade, we don't have to run all the tempest tests, right?
15:46:25 <yamahata> Oops, we lost manjeets.
15:46:36 <yamahata> we have 15mins left.
15:47:03 <yamahata> manjeets: how about grenade? we don't have to run all the tempest tests, right?
15:47:57 <yamahata> seems manjeets_ connection issue.
15:48:04 <yamahata> let's move on to patches.
15:48:06 <yamahata> #topic patches/bugs
15:48:32 <yamahata> so far we need to complete doc-migration. we have mostly done it.
15:48:47 <yamahata> I suppose pramod and manjeets_ will follow up.
15:49:17 <yamahata> aldo dhcp port creation, Karthik would follow it up to clean up nitpicks and tests.
15:49:24 <yamahata> any other patches/bugs?
15:49:34 <yamahata> #link https://review.openstack.org/#/q/project:openstack/networking-odl+status:open
15:50:16 <yamahata> there are low hanging fruits for bug fix patches. For Pike release, we should clear them up.
15:50:50 <mkolesni> yamahata, do you have a list?
15:51:01 <yamahata> I have only the linke above.
15:51:24 <yamahata> some are for bug fix, some would be for queens.
15:51:31 <mkolesni> oh thats a general one
15:51:41 <mkolesni> i think we need to do some bug scrub
15:51:53 <yamahata> Yeah, right.
15:51:56 <mkolesni> sory mean gerrit scrub
15:52:18 <mkolesni> also patches older than 1 month can be auto-abandoned like in neutron
15:53:25 <yamahata> With kevin, it's not. and it looks like some old patches seem meaningful.
15:53:56 <mkolesni> they can always be restored though
15:54:57 <yamahata> Right.
15:55:11 <yamahata> we have 5mins left.
15:55:19 <yamahata> anything else?
15:55:25 <yamahata> #topic open mike
15:57:06 <yamahata> nothing?
15:57:08 <yamahata> going one
15:57:10 <yamahata> going once
15:57:23 <mkolesni> thanks Isaku :)
15:57:25 <yamahata> going twice
15:57:30 <yamahata> thank you everyone.
15:57:33 <mkolesni> have a nice day guys
15:57:36 <mkolesni> bye
15:57:40 <yamahata> #topic cookies
15:57:43 <yamahata> #endmeeting