14:00:00 <jlibosva> #startmeeting networking
14:00:01 <openstack> Meeting started Tue Apr 25 14:00:00 2017 UTC and is due to finish in 60 minutes.  The chair is jlibosva. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:04 <openstack> The meeting name has been set to 'networking'
14:00:07 <jlibosva> Hello everyone
14:00:10 <mlavalle> o/
14:00:23 <dasm> o/
14:00:24 <bcafarel> hello
14:00:26 <haleyb> hi
14:00:40 <jlibosva> #topic Announcements
14:00:44 <hichihara> hi
14:00:46 <john-davidge> o/
14:00:58 <ralonsoh> hi
14:01:10 <jlibosva> It was an announcement that Mitaka is about to be tagged as EOL
14:01:18 <jlibosva> I haven't seen Neutron being mentioned there yet
14:01:29 <jlibosva> #link http://lists.openstack.org/pipermail/openstack-dev/2017-April/115308.html
14:02:09 <jlibosva> dasm: you're no longer release liaison right? Do you have any clue when Neutron is gonna be cut?
14:03:10 <dasm> jlibosva: no one took responsibilities from me, yet. so technically i'm still liaison.
14:03:26 <dasm> jlibosva: but afair, we need to give a green light to mitaka eol
14:03:41 <jlibosva> dasm: but that hasn't been proposed yet, right?
14:03:42 <dasm> jlibosva: i need to sync that with release team and kevinbenton
14:03:46 <jlibosva> ack
14:03:48 <dasm> i don't think so
14:03:50 <jlibosva> dasm: thanks
14:03:53 <jlibosva> next one
14:03:59 <andreas_s> hi
14:04:15 <kevinbenton> I think we only need to opt out
14:04:26 <kevinbenton> by default we will EOL IIUC
14:05:11 <jlibosva> but the question is when :) There is a bug that has some security impact in Mitaka (I want to talk aobut that later in bugs topic)
14:05:35 <dasm> jlibosva: technically, it should already happen last week, iirc
14:06:19 <bcafarel> the original mail mentioned 2017-04-17, so "soon"?
14:06:21 <jlibosva> so if there are any backport patches pending or bugs scheduled to mitaka, we might want to do the final release before the cut
14:06:32 <dasm> " Please look over both lists by 2017-04-17 00:00 UTC"
14:06:43 <dasm> jlibosva: would be good
14:06:59 <jlibosva> ok, we can talk about specifics in bugs section
14:07:44 <jlibosva> next announcement is that our ptl wants to pay better attention to review dashboards to not let patches fall thru cracks
14:07:46 <jlibosva> #link https://docs.openstack.org/developer/neutron/dashboards/index.html
14:08:37 <jlibosva> Next announcement: There is no Neutron CI meeting today
14:09:01 <jlibosva> Last announcement from me is that Boston summit Forum schedule is now known: http://lists.openstack.org/pipermail/openstack-dev/2017-April/115174.html
14:09:03 <jlibosva> #link http://lists.openstack.org/pipermail/openstack-dev/2017-April/115174.html
14:09:12 <jlibosva> there is a pdf document attached
14:09:18 <jlibosva> anybody else has anything to announce?
14:09:56 <kevinbenton> if you're attending the summit
14:10:06 <kevinbenton> put your name on this
14:10:08 <kevinbenton> #link https://etherpad.openstack.org/p/neutron-boston-summit-attendees
14:10:22 <kevinbenton> so we can coordinate and plan a social
14:10:28 <kevinbenton> #link http://lists.openstack.org/pipermail/openstack-dev/2017-April/115766.html
14:11:48 <jlibosva> oh, I also noticed this email: http://lists.openstack.org/pipermail/openstack-dev/2017-April/115723.html
14:12:04 <jlibosva> we may need an api liaison as Henry no longer works on Neutron
14:12:27 <jlibosva> maybe a ML would be better to get one :)
14:12:41 <jlibosva> so if nobody wants to announce something else we can move on
14:12:52 <jlibosva> #topic Blueprints
14:13:40 <jlibosva> there was this spec brought up several times https://review.openstack.org/#/c/203509/
14:14:49 <jlibosva> kevinbenton: do you remember if that bp was discussed by drivers?
14:15:03 <kevinbenton> jlibosva: it wasn't
14:15:37 <hichihara> I think that the spec needs feedbacks or +1 by OVS folks.
14:15:49 <jlibosva> kevinbenton: is it on the radar?
14:15:51 <kevinbenton> hichihara: you mean reference implmentation people?
14:16:33 <hichihara> kevinbenton: and OVN
14:16:40 <yushiro> jlibosva, kevinbenton Hi.  I and annp works on this logging BP.
14:16:54 <kevinbenton> hichihara: ok
14:17:11 <jlibosva> I also promised to have a look but I didn't yet
14:17:27 <jlibosva> yushiro: hi :)
14:17:43 <kevinbenton> yeah, i don't think anything is changing from a drivers perspective for it
14:17:45 <yushiro> In my understanding, we  decided to support OVS native firewall driver first after PTG.
14:18:06 <kevinbenton> we just need an approver
14:18:53 <kevinbenton> we can discuss it in the drivers meeting on thursday
14:19:24 <yushiro> kevinbenton, aha, yes.  Rossella is now approver.  She wants to know some decision in PTG.
14:19:25 <jlibosva> that would make sense, it seems we have devs and they are blocked by processes
14:19:43 <yushiro> kevinbenton, this Thursday, right?
14:19:56 <kevinbenton> yes
14:20:11 <mlavalle> yushiro: 2200 UTC
14:20:35 <mlavalle> so maybe Friday for you
14:21:06 <yushiro> mlavalle, OK, thanks.  I'll join this mtg.
14:21:07 <hichihara> yushiro: If you can, you should get Russell Bryant as OVN folks.
14:21:37 <yushiro> hichihara, OK.  I'll reach out him later.
14:21:50 <jlibosva> I see he's already on the reviewers' list
14:22:14 <hichihara> jlibosva: yes but he left -1
14:22:50 <jlibosva> does anybody have any other bp that lacks attention or is blocked in any way?
14:22:52 <yushiro> jlibosva, Thanks for your attention :)
14:22:53 <jlibosva> hichihara: ah, I see
14:24:02 <jlibosva> if there is no other bp to be discussed, let's move on
14:24:10 <jlibosva> #topic Bugs and gate failures
14:24:20 <jlibosva> I was bug deputy for the last week
14:24:46 <jlibosva> we have a critical bug that has no assignee https://bugs.launchpad.net/neutron/+bug/1683090
14:24:47 <openstack> Launchpad bug 1683090 in neutron "test_dhcpv6_64_subnets fails with IpAddressAlreadyAllocated" [Critical,Confirmed]
14:25:36 <jlibosva> it fails a tempest job on the gate
14:25:45 <jlibosva> does anybody want to look at it?
14:27:08 <mlavalle> jlibosva: I am bugs deputy this week, so I'll take a look
14:27:29 <jlibosva> mlavalle: cool, thanks. ping me if you want to collaborate
14:27:48 <jlibosva> next bug I wanted to highlight is https://bugs.launchpad.net/neutron/+bug/1685237
14:27:49 <openstack> Launchpad bug 1685237 in neutron "port security does not block router advertisements for instances" [Undecided,New]
14:27:50 <mlavalle> will do :-)
14:28:28 <jlibosva> as mentioned before, Mitaka life is going to the end and I wanted to make a decision if we want to backport fix and do release before EOL
14:28:56 <jlibosva> if the backport wouldn't be too invasive
14:29:07 <dasm> if bug is security issue, we should probably try
14:29:19 <haleyb> jlibosva: i wonder if the RA issue is still broken in master, guessing not
14:29:59 <kevinbenton> yeah, we should find out the fix
14:30:07 <jlibosva> haleyb: I did a quick look and didn't find any specific patch but could be part of other fix
14:31:01 <haleyb> we changed the icmpv6 code recently, but on mitaka it could just be a single rule was missed
14:31:17 * haleyb doesn't have a mitaka environment around
14:32:10 <haleyb> jlibosva: i can ask some more questions to narrow it down a bit
14:32:38 <jlibosva> haleyb: thanks, maybe I could spin-up a mitaka vm
14:33:34 <jlibosva> we can talk about that on openstack-neutron after mtg
14:33:40 <haleyb> seeing ip6tables-save output would be a start, i'll ask
14:33:56 <haleyb> jlibosva: i know there's devstack instructions on older releases, been a while...
14:34:36 <jlibosva> that's all bugs I wanted to share with the team
14:35:17 <jlibosva> mlavalle: you're the bug deputy for the next week, right?
14:35:26 <mlavalle> this week
14:35:31 <jlibosva> oh, right :)
14:35:38 <mlavalle> LOL
14:35:54 <mlavalle> unless you want to do double duty :-)
14:36:03 <jlibosva> so we'll need a volunteer for the week of May 1
14:36:05 <kevinbenton> haleyb: https://review.openstack.org/#/q/Ice1c9dd349864da28806c5053e38ef86f43b7771
14:36:51 <haleyb> kevinbenton: thanks, i'll see if htey have that
14:37:10 <kevinbenton> haleyb: if they do, i'm guessing it's related to them having security groups feature disabled
14:37:31 <kevinbenton> haleyb: not something we test
14:38:09 <jlibosva> mlavalle: doubleduty can be also possible if you're willing to :)
14:39:13 <kevinbenton> jlibosva: we can also find someone during next meeting
14:39:24 <jlibosva> kevinbenton: seems like we have to
14:39:42 <jlibosva> let's move on from bug deputies then
14:40:07 <jlibosva> one thing regarding gate, we agreed on the last Neutron CI meeting that we're gonna propose functional job to the gate queue
14:40:23 <jlibosva> to start broader discussion
14:40:31 <jlibosva> as it was quite stable lately, mostly copying the curve of api job
14:41:06 <jlibosva> that's about the gate
14:41:08 <jlibosva> #topic Docs
14:41:12 <jlibosva> john-davidge: hi, are you around ?
14:41:19 <john-davidge> jlibosva: Hi!
14:41:56 <jlibosva> john-davidge: do you want to give us updates regarding docs?
14:42:07 <john-davidge> jlibosva: So, my update this week is to say that with the closing down of OSIC I will likely have to step down from my docs liason role very soon.
14:42:18 <jlibosva> oh :( sad to hear that
14:42:53 <john-davidge> I should know for sure within the next few days. If anyone is interested in taking over then please get in touch with me and/or kevinbenton/asettle
14:43:09 <john-davidge> Docs are fun! I promise :)
14:43:51 <jlibosva> it seems john-davidge knows what he's talking about :)
14:44:05 <jlibosva> volunteers are welcome
14:44:24 <jlibosva> john-davidge: thanks a lot for all the great work!
14:44:35 <john-davidge> jlibosva: Sorry for the potentially bad news, I'll hopefully know more by next week's meeting
14:45:13 <jlibosva> john-davidge: ok, I assume you're gonna talk with kevinbenton :)
14:45:22 <jlibosva> next one
14:45:24 <jlibosva> #topic Transition to OSC
14:45:27 <jlibosva> amotoki: hi, are you around?
14:45:31 <john-davidge> jlibosva: Yes, we've already discussed it and will continue to do so
14:45:39 <amotoki> jlibosva: hi
14:45:54 <jlibosva> amotoki: do you have any updates from osc world?
14:46:15 <amotoki> but there is no update from the past report. L3 agent and tag supports are the last place
14:46:24 <amotoki> s/place/pieces/
14:47:18 <amotoki> L3 agent stuff is good progress. Tag support might need more time but I think we can have it in Pike.
14:47:24 <amotoki> that's all
14:47:28 <jlibosva> oh, that's good
14:47:31 <jlibosva> amotoki: thanks
14:47:44 <jlibosva> #topic Neutron-lib
14:47:58 <jlibosva> boden: hi, you here already?
14:48:05 <boden> jlibosva hi yes
14:48:09 <jlibosva> good
14:48:14 <jlibosva> boden: stage is yours :)
14:48:24 <boden> just a few quick FYIs
14:49:20 <boden> the transition to neutron-lib’s callbacks is underway.. rather than digress here, I’m going to be sending a note to dev ML about it soon, so if interested please watch for the email
14:50:49 <boden> the other thing is there’s 2 WIP patches out for review that (a) try to solidify the notion of a service definition and (b) provide the plugin loading functionality of NeutronManager.. if anyone is interested in this topic please feel free to watch/comment on those patches (thanks kevinbenton for initial feedback)
14:51:01 <boden> #link https://review.openstack.org/#/c/455451/
14:51:10 <boden> #link https://review.openstack.org/#/c/456343/
14:51:12 <boden> respectively
14:51:24 * mlavalle adding them to my starred reviews
14:51:33 <boden> that’s really all I have, unless others have questions/comments
14:52:04 <jlibosva> boden: thanks for updates, no questions from me :)
14:52:08 <boden> and of course the usual: please have a peek at neutron-lib reviews when you get time :)
14:53:01 <jlibosva> so unless there are no questions about neutron-lib we can go to open discussion
14:53:14 <jlibosva> #topic Open discussion
14:53:41 <jlibosva> the release liaison was discussed on prev meeting so there is no item in On demand agenda
14:53:51 <jlibosva> does anybody have a subject to discuss?
14:53:58 <haleyb> jlibosva: i added an item this morning
14:54:04 <jlibosva> oh I didn't see
14:54:10 <haleyb> https://bugs.launchpad.net/neutron/+bug/1668145
14:54:11 <openstack> Launchpad bug 1668145 in neutron "[RFE] Allow operator control of "on-link" routes for subnets in the same Neutron network" [Wishlist,In progress] - Assigned to Brian Haley (brian-haley)
14:54:17 <jlibosva> aah
14:54:19 <jlibosva> haleyb: thanks :)
14:54:26 <jlibosva> #topic Control sending of on-link routes
14:54:48 <jlibosva> #link  https://bugs.launchpad.net/neutron/+bug/1668145
14:55:16 <haleyb> kevinbenton: really a question for you - i proposed changing the Network model and API instead of a config value, can we still get that into Pike if it's acceptable?
14:56:14 <haleyb> https://review.openstack.org/#/c/459492/ is the change i just sent, but just realized the RFE isn't approved, but this looks more like a bug
14:56:17 <kevinbenton> what i don't understand is why this isn't expressed using routed networks
14:56:42 <kevinbenton> if two subnets aren't on the same l2 domain, that should be reflected in the api
14:57:02 <kevinbenton> and once that's there we can put the logic on the dhcp agent to only add onlink routes for things in the same l2 segment
14:57:32 <haleyb> i think these are in the same L2 domain, that's why that code was added
14:57:53 <kevinbenton> then why would you want to disable onlink routes?
14:58:04 <kevinbenton> as far as i understand the RFE, the segments weren't directly attached
14:58:53 <kevinbenton> haleyb: to answer your question, this feature does look simple enough to land in Pike
14:58:54 <haleyb> the original code assumed they were
14:59:05 <kevinbenton> but i do want to make sure there is a use case
14:59:08 <kevinbenton> right
14:59:24 <kevinbenton> the correct fix sounds to me like just ensuring that subnets are on the same segment
14:59:31 <kevinbenton> inside of the dhcp agent
14:59:38 <kevinbenton> before setting up onlink routes
15:00:02 <jlibosva> we're on top of the hour, can we continue the discussion on #openstack-neutron ?
15:00:06 <kevinbenton> sure
15:00:09 <haleyb> yes
15:00:10 <mlavalle> o/
15:00:12 <jlibosva> cool
15:00:15 <jlibosva> thanks everyone for joining
15:00:18 <jlibosva> #endmeeting