21:59:53 <kevinbenton> #startmeeting neutron_drivers
21:59:54 <openstack> Meeting started Thu May 18 21:59:53 2017 UTC and is due to finish in 60 minutes.  The chair is kevinbenton. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:59:55 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:59:57 <openstack> The meeting name has been set to 'neutron_drivers'
22:00:04 <mlavalle> o/
22:00:06 <kevinbenton> hello!
22:00:25 <kevinbenton> armax, amotoki, ihrachys: ping
22:00:28 <ihrachys> o/
22:00:33 <amotoki> hi
22:00:43 <kevinbenton> first time we've had everyone in a long time :)
22:00:48 <armax> wow
22:01:19 <kevinbenton> first i'd like to talk about the community goals really quick
22:01:40 <kevinbenton> ihrachys: we had briefly chatted about switching at least one tempest job to python3
22:01:52 <ihrachys> tempest?
22:02:05 <ihrachys> because I think we also planned to switch functional/fullstack?
22:02:27 <kevinbenton> yeah
22:02:37 <kevinbenton> but one of the tempest as well i think to exercise actual VM boots and all
22:02:41 <ihrachys> sure
22:02:42 <armax> kevinbenton: I thought taht switching tempest needs to be a concerted effort across projects
22:03:03 <ihrachys> I think there was some experimental job somewhere that attempted to switch all things to py3
22:03:04 <armax> otherwise we risk to break because of other projects failing py3
22:03:07 <kevinbenton> ihrachys: well if we switch functional to start, does it pass right now?
22:03:13 <kevinbenton> armax: i think ihrachys had a fix for that
22:03:22 <ihrachys> kevinbenton, I would assume no but I dunno until we check
22:03:22 <armax> kevinbenton: what’s the fix?
22:03:22 <kevinbenton> armax: so projects had to opt into python3
22:03:35 <kevinbenton> i remember that was the case for rally
22:03:57 <ihrachys> kevinbenton, I think dhellmann and sdague were the last ones who were looking at this in devstack
22:04:05 <kevinbenton> ok
22:04:13 <kevinbenton> well lets just do functional/fullstack then
22:04:19 <ihrachys> https://review.openstack.org/#/c/439108/
22:04:20 <kevinbenton> shall we sacrifice the functional job?
22:04:40 <kevinbenton> it's been unusually stable :)
22:05:00 <ihrachys> well some worked on it..
22:05:09 <ihrachys> + for sacrifice
22:05:27 <kevinbenton> ok, we can see where that's at after the meeting to push it forward
22:05:30 <armax> it’s stable because there’s less volume of changes come into the system
22:05:34 <ihrachys> and maybe let's just lag on tempest until community comes up with proper nudge and solution
22:05:37 <armax> :)
22:05:41 <kevinbenton> ihrachys: sounds good
22:05:51 <ihrachys> armax, yeah, 'magic'
22:05:54 <kevinbenton> so the other one is switching to wsgi
22:05:55 <armax> long are the days were I would wake up to 100s of review notifications in my inbox
22:06:09 <armax> ihrachys: +
22:06:16 <kevinbenton> there is a bug tracking this somewhere isn't there?
22:06:18 <armax> let’s switch to something we have full control over
22:06:45 <kevinbenton> i think i'm going to make a blueprint to track the wsgi one because there are several things that need to be done
22:06:56 <ihrachys> I am not aware of a bug. there was a etherpad
22:07:36 <ihrachys> https://etherpad.openstack.org/p/py3-neutron-pike
22:07:45 <ihrachys> well it's not very helpful :p
22:07:58 <ihrachys> I think jlibosva was supposed to come up with a plan but..
22:08:16 <armax> but?
22:08:17 <ihrachys> you can see there is a bug indeed https://bugs.launchpad.net/neutron/+bug/1683301
22:08:18 <openstack> Launchpad bug 1683301 in neutron "python 3.5 support" [Undecided,New]
22:08:30 <kevinbenton> sorry, not a bug for py3
22:08:30 <ihrachys> but something happened, not sure. atm Jakub is on PTO.
22:08:33 <kevinbenton> one for wsgi
22:08:50 <kevinbenton> i just filed this to track efforst
22:08:50 <kevinbenton> https://blueprints.launchpad.net/neutron/+spec/run-in-wsgi-server
22:08:54 <armax> oh right
22:09:13 <ihrachys> bug #1666779
22:09:14 <openstack> bug 1666779 in neutron "Expose neutron API via a WSGI script" [Wishlist,In progress] https://launchpad.net/bugs/1666779 - Assigned to Ihar Hrachyshka (ihar-hrachyshka)
22:09:14 <kevinbenton> i know there was a devstack patch floating around to actually run
22:09:43 <ihrachys> the devstack piece: https://review.openstack.org/439191
22:09:49 <ihrachys> the author left the team
22:10:11 <mlavalle> yeah electrocucaracha
22:11:00 <kevinbenton> ok
22:11:16 <kevinbenton> #info armax is going to be release person for cycle
22:11:30 <kevinbenton> ok, i have the devstack piece recorded in the blueprint
22:11:39 <kevinbenton> the neutron-rpc-server binary needs to be fixed so it works
22:11:48 <kevinbenton> which is what yamamoto was getting at in the review
22:12:03 <ihrachys> yeah, there is some neutron side work
22:12:05 <ihrachys> "some"
22:12:13 <kevinbenton> https://bugs.launchpad.net/neutron/+bug/1687896
22:12:14 <openstack> Launchpad bug 1687896 in neutron "neutron-rpc-server fails to start on configuration that works under neutron-server" [Undecided,New]
22:12:24 <kevinbenton> looks like someone tried it already and opened a bug
22:12:43 <kevinbenton> ok
22:12:53 <kevinbenton> i will look into these things for that goal
22:13:13 <kevinbenton> are there any other blueprints that are now at risk due to OSIC cuts, etc?
22:13:34 <mlavalle> the whole OVO effort, maybe?
22:13:52 <kevinbenton> ihrachys: is it basically just down to you?
22:14:01 <ihrachys> ehm it's at risk but it's not like I bail out
22:14:13 <ihrachys> yes it's on me basically. also manjeets helps with some bits.
22:14:27 <ihrachys> I am not bailing out though.
22:14:37 <armax> ihrachys: not now anyway? :
22:14:40 <armax> :P
22:14:48 <ihrachys> not today
22:14:50 <kevinbenton> ihrachys: ok, are you using that google docs spreadsheet to track which ones need to be done?
22:14:54 <armax> it’s becoming a race to who’s the last one standing
22:15:21 <ihrachys> kevinbenton, I should make a proper assessment of work items, haven't looked this way for quite a while
22:15:31 <ihrachys> kevinbenton, lemme come up with smth the next week.
22:15:49 <ihrachys> one spec that is not affected by OSIC but by other changes in the team is, I believe, flow manager is going to die: https://review.openstack.org/320439
22:16:03 <kevinbenton> ihrachys: ok, if you get the blueprint whiteboard updated with a todo i might be able to take some on in my spare time
22:16:16 <kevinbenton> ihrachys: yeah
22:16:23 <ihrachys> just talked to David, and he seems like he won't work on this, and we neither have interested parties.
22:16:24 <kevinbenton> ihrachys: i'm getting that suspicion as well
22:16:48 <ihrachys> I am fine to kill it. it's a nice idea but won't deliver. tough life.
22:17:15 <ihrachys> (I actually wanted to kill it in the past, but then David and Thomas asked to keep it on life support)
22:17:22 <armax> yet another victim of ovn defection
22:17:33 <kevinbenton> ihrachys: is it blocking any features on their side?
22:17:34 <ihrachys> armax, basically yes
22:17:34 <armax> natural selection
22:17:34 <amotoki> who needs this spec? I think fwaas or sfc folks need it. is it worth asking them?
22:17:35 <armax> that’s good
22:17:43 <ihrachys> armax, both resource wise and philosophically
22:18:12 <armax> well, I think we need to seriously consider the impact on people who geniunely are still interested in writing this stuff in python
22:18:16 <ihrachys> amotoki, I am beyond the point where I will ask if they need. if they would be interested, I would see more activity on the spec.
22:18:22 <armax> because we’re not trading apples with apples here
22:19:01 <kevinbenton> right, OVN definitely isn't a replacement for a thing we can actually develop new features on
22:19:13 <kevinbenton> so we'll probably be stuck with OVS ref for quite a while
22:19:52 <armax> we should try and not to kill the spec itself though
22:19:52 <armax> https://review.openstack.org/#/c/320439/
22:19:57 <ihrachys> kevinbenton, well it depends on how you look at it. even for ovs, you need to lay ground work before you consume its features in neutron. it's just more stuff pushed into platform.
22:19:59 <armax> it’d be nice to park it somewhere
22:20:03 <armax> for future reference
22:20:21 <ihrachys> armax, like devref?
22:20:26 <armax> ihrachys: if we merged it as it
22:20:28 <armax> is*
22:20:34 <armax> but not work on it
22:20:37 <armax> it’ll go in the backlog
22:20:39 <armax> and stay there
22:20:48 <kevinbenton> ihrachys: only for a limited set of things. anything 'steering traffic' using openflow has been suppored by OVS since openflow became a thing
22:20:57 <armax> and eventually archive
22:21:02 <armax> https://github.com/openstack/neutron-specs/tree/master/specs
22:21:09 <kevinbenton> which is what things like SFC and BGPVPN need
22:21:17 <kevinbenton> OVN just gives much higher level primitives
22:21:45 <ihrachys> armax, I am fine either way. maybe just move it to backlog right away and merge?
22:21:46 <armax> which actually made me notice that there’s still an ocata backlog that hasn’t been dealt with
22:21:47 <armax> https://github.com/openstack/neutron-specs/tree/master/specs/backlog/ocata
22:21:54 <ihrachys> with a NOTE at the top
22:22:06 <armax> ihrachys: I am fine either waty
22:22:09 <kevinbenton> armax: what does "dealt with"
22:22:11 <kevinbenton> mean?
22:22:11 <armax> way*
22:22:47 <armax> those blueprints should either be retargeted to pike
22:22:50 <armax> or moved to archive
22:22:52 <ihrachys> mmm, thanks for the notice, I need to move the online-upgrades one, it's still relevant for Pike, and we have something to land for it.
22:23:07 <kevinbenton> ack
22:23:39 <kevinbenton> agentless-driver has a blueprint for pike
22:23:40 <kevinbenton> https://blueprints.launchpad.net/neutron/+spec/agentless-driver
22:23:43 <kevinbenton> but no patches it looks like
22:24:03 <armax> ihrachys: ack
22:24:15 <armax> kevinbenton: I thought garyk was interested in taht one
22:25:12 <amotoki> looking at the bug report, i found https://review.openstack.org/#/c/349921/ but there has been no progress from January
22:26:11 <kevinbenton> ok, let's look through some of the RFEs
22:26:28 <kevinbenton> i think some are as a result of the summit
22:26:30 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bugs?field.status%3Alist=Triaged&field.tag=rfe
22:26:54 <mlavalle> kevinbenton: https://bugs.launchpad.net/neutron/+bug/1686035
22:26:56 <openstack> Launchpad bug 1686035 in neutron "[RFE] More detailed reporting of available QoS rules" [Wishlist,Triaged] - Assigned to Slawek Kaplonski (slaweq)
22:27:03 <mlavalle> was at the top of the list earlier
22:27:19 <mlavalle> but I re-classified at rfe and went to the bottom
22:27:28 <mlavalle> I saw slaweq_ signing in
22:27:46 <mlavalle> maybe he wants to discuss
22:27:51 <kevinbenton> ok
22:28:21 <mlavalle> slaweq_: are you there?
22:29:23 <mlavalle> mhhh, maybe not
22:29:25 <kevinbenton> ok
22:29:31 <kevinbenton> let's go from top down then for now
22:29:35 <mlavalle> yeap
22:29:49 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1505627
22:29:50 <openstack> Launchpad bug 1505627 in neutron "[RFE] QoS Explicit Congestion Notification (ECN) Support" [Wishlist,Triaged] - Assigned to Reedip (reedip-banerjee)
22:29:58 <kevinbenton> ihrachys: who is this qos team you speak of? :)
22:30:05 <kevinbenton> ihrachys: what are their feelings on it
22:30:25 <ihrachys> well there is Rodolfo and Slawek at least
22:30:37 <kevinbenton> ihrachys: shall we wait for them to reply?
22:30:49 <mlavalle> and slawek also shows up to their bi-weekly meeting on Tuesdays
22:31:09 <ihrachys> I dunno what their feelings are, but I see a lot of stuff they seem to care, and I wonder if that other thing is on their radar for Pike. I am not sure.
22:31:17 <armax> we should can this one
22:31:32 <armax> it’s been sitting in the pipe dream stage for too long
22:31:39 <kevinbenton> armax: that's not a reason
22:31:53 <armax> OK, then approve all the things
22:31:53 <kevinbenton> let's wait for feedback from them
22:32:08 <mlavalle> kevinbenton: I attend their meeting regularly. I can bring it up next time
22:32:12 <kevinbenton> mlavalle: thx
22:32:14 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1558812
22:32:16 <openstack> Launchpad bug 1558812 in neutron "[RFE] Enable adoption of an existing subnet into a subnetpool" [Wishlist,Triaged] - Assigned to Reedip (reedip-banerjee)
22:32:16 <armax> the thread on the report goes back and forth for months and months
22:32:21 <kevinbenton> this is an old one as well
22:32:35 <mlavalle> with a good summary from Carl
22:32:36 <armax> kevinbenton: what confidence do you have this will suddenly becomes high priority?
22:32:44 <armax> anyhoo
22:32:46 <armax> let’s wait
22:33:52 <kevinbenton> I think we should be able to adopt subnets now if each adoption bumps the revision of the subnet pool
22:34:01 <kevinbenton> that will protect us from the race condtion that plagued this before
22:34:32 <kevinbenton> or bumps the revision of the address scope
22:35:40 <kevinbenton> what do we think about just putting this to approved and then decide based on teh complexity of the patch
22:35:49 <armax> the latest patch I can recall was https://review.openstack.org/#/c/348080/
22:35:49 <kevinbenton> it looks relatively short and isolated as it is
22:36:17 <kevinbenton> there is also this one https://review.openstack.org/#/c/407697/2
22:36:30 <armax> that was was a clone of ryan's
22:36:40 <armax> and I blocked it to avoid losing the history of the initial patch
22:37:02 <armax> if we do have someone willing to work on the patch
22:37:09 <armax> we should just restore the blueprint and RFE
22:37:15 <armax> it was approved at one point
22:37:43 <kevinbenton> ok, reedip said he was interested
22:37:46 <armax> we can provisionally put it back on the radar and see what thappens
22:37:48 <armax> happens
22:37:51 <kevinbenton> that's why i brought it back from postponed
22:38:09 <kevinbenton> armax: ok, set to rfe-approved and lets see what happens
22:38:12 <mlavalle> ++
22:38:46 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1561824
22:38:47 <openstack> Launchpad bug 1561824 in neutron "[RFE] Enhance BGP Dynamic Routing driver with Quagga suppport" [Wishlist,Triaged] - Assigned to Steve Ruan (ruansx)
22:38:55 <kevinbenton> this one seems like it should maybe be rfe-postponed for now
22:39:35 <kevinbenton> nobody restored the spec
22:40:21 <mlavalle> Steve used to show up to the L3 subteam regularly
22:40:39 <kevinbenton> mlavalle: ok, if he shows up again just have him restore the spec
22:40:44 <mlavalle> But he hasn't shown up since maybe october - novemeber
22:40:51 <mlavalle> will do
22:40:54 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1596611
22:40:55 <openstack> Launchpad bug 1596611 in neutron "[RFE] Create L3 IPs with qos (rate limit)" [Wishlist,Triaged] - Assigned to LIU Yulong (dragon889)
22:41:02 <kevinbenton> ihrachys: this one needs buy-in from qos folks as well?
22:41:41 <ihrachys> I would imagine they may have some input, yes
22:41:46 <ihrachys> the use case seems interesting
22:41:52 <ihrachys> and the WIP patches seem promising
22:42:07 <ihrachys> I reviewed the spec, needs some refinement
22:42:45 <mlavalle> kevinbenton: again, I can bring it up in their next meeting
22:43:02 <kevinbenton> ihrachys: LIU has been a pretty good contributor in the past, shall we just move to rfe-approved and see if we can get agreement on the spec from everyone?
22:43:03 <mlavalle> just keep in mind it is bi-weekly and they had it this week
22:43:17 <ihrachys> kevinbenton, I am fine, he seems active
22:43:48 <ihrachys> my main concern about the other qos rfe was that it's not clear who's going to support it. here, I see how it's useful and hence can help.
22:44:01 <kevinbenton> ack
22:44:04 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1639566
22:44:04 <openstack> Launchpad bug 1639566 in neutron "[RFE] Add support for local SNAT" [Wishlist,Triaged]
22:44:48 <kevinbenton> this is distributed SNAT where different colored lipstick
22:44:54 <mlavalle> lol
22:45:24 <kevinbenton> wearing*
22:45:50 <kevinbenton> based on the meeting at the summit it sounds like we won't have review bandwidth to make that change this cycle
22:46:21 <kevinbenton> and even the person that filed the RFE said they won't be able to write the code
22:46:31 <kevinbenton> so i think rfe-postponed is appropriate for this one
22:46:42 <kevinbenton> agreement?
22:46:53 <mlavalle> yeap
22:47:17 <amotoki> sounds fair
22:48:02 <armax> I thought the feedback fromt he summit was to let users have a go at dragonflow
22:48:28 <kevinbenton> yeah, that's the same thing as not doing it in DVR :)
22:48:31 <armax> and see if l3 flavor enhancements were required in case users wanted to use dragonflow in conjunction with other L3 options
22:48:52 <kevinbenton> yeah, we can definitely make changes to L3 flavors to help if needed to test DVR
22:49:02 <kevinbenton> i already mentioned that in the neutron meeting on monday
22:49:37 <armax> so rather than postpone we should possibly close the RFE altogehter?
22:49:42 <kevinbenton> no
22:49:50 <kevinbenton> in the future someone may actually want to change DVR
22:49:56 <armax> OK
22:49:56 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1658682
22:49:57 <openstack> Launchpad bug 1658682 in neutron "port-security can't be disabled if security groups are not enabled" [Wishlist,Triaged]
22:50:13 <kevinbenton> this now just sounds like a regular bug to me
22:50:32 <kevinbenton> we are requiring security groups to be enabled to touch an API value that is unrelated
22:51:35 <amotoki> agree. in some summit forum session, someone complained the same thing.
22:51:48 <armax> amotoki: I think it was the same person
22:52:16 <amotoki> perhaps
22:53:23 <armax> OK, I can fix it
22:53:38 <amotoki> I think we need to clarify what we expect with port security enabled and SG disabled.
22:53:50 <armax> amotoki: just spoofing rules applied
22:53:51 <amotoki> my understanding is that anti-spoofing is enabled but no filtering like SG
22:53:53 <armax> as kevin said
22:54:01 <kevinbenton> yes
22:54:15 <amotoki> thanks. move on
22:54:21 <armax> I’ll give it a stub tonight
22:54:22 <armax> if I can
22:54:35 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1662650
22:54:36 <openstack> Launchpad bug 1662650 in neutron "[RFE] Advance configuration of SR-IOV ports- api extension" [Wishlist,Triaged] - Assigned to Trevor McCasland (twm2016)
22:54:59 <kevinbenton> trevormc: hey, you were going to break this into smaller RFEs, has that been done yet?
22:55:08 <mlavalle> This is what we discussed with Trevor and team on Thursday
22:55:17 <trevormc> yes I have
22:55:21 <trevormc> three are on the way
22:55:25 <mlavalle> I already saw a couple of them in the ML
22:55:26 <trevormc> two are currently available
22:55:32 <trevormc> https://bugs.launchpad.net/neutron/+bug/1690937
22:55:33 <openstack> Launchpad bug 1690937 in neutron "[RFE] Support allowed address pairs without ip address" [Wishlist,Triaged]
22:55:38 <trevormc> https://bugs.launchpad.net/neutron/+bug/1690921
22:55:39 <openstack> Launchpad bug 1690921 in neutron "[RFE] Manage Broadcast, Unicast, and Multicast traffic" [Wishlist,Triaged]
22:55:49 <kevinbenton> ok, we'll probably get to those next session
22:55:57 <kevinbenton> i'll move to triaged state after reading through
22:56:00 <trevormc> I have a meeting with the research team next week.
22:56:06 <kevinbenton> oh i see
22:56:10 <armax> 4 minutes
22:56:10 <kevinbenton> already triaged
22:56:20 <kevinbenton> trevormc: so can we dump https://bugs.launchpad.net/neutron/+bug/1662650 ?
22:56:21 <openstack> Launchpad bug 1662650 in neutron "[RFE] Advance configuration of SR-IOV ports- api extension" [Wishlist,Triaged] - Assigned to Trevor McCasland (twm2016)
22:56:32 <trevormc> yes
22:56:39 <mlavalle> trevormc: will you be ready by Thursday of next week?
22:56:54 <trevormc> mlavalle: I hope so, you will see the new RFEs if so
22:57:01 <mlavalle> ++ :-)
22:57:38 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1671634
22:57:39 <openstack> Launchpad bug 1671634 in neutron "[RFE] Allow to set MTU for networks" [Wishlist,Triaged] - Assigned to Ihar Hrachyshka (ihar-hrachyshka)
22:57:39 <armax> I thought this was already solved: https://bugs.launchpad.net/neutron/+bug/1682775
22:57:41 <openstack> Launchpad bug 1682775 in neutron "[RFE] Tag mechanism supports all resouces with standard attribute." [Wishlist,Triaged] - Assigned to Hirofumi Ichihara (ichihara-hirofumi)
22:57:46 <armax> sorry
22:57:50 <kevinbenton> #link https://bugs.launchpad.net/neutron/+bug/1671634
22:57:54 <kevinbenton> ihrachys: can you take over that one?
22:58:18 <ihrachys> kevinbenton, which of?
22:58:19 <ihrachys> mtu?
22:58:30 <ihrachys> yes, once it gets approved, I will move to it
22:58:35 <kevinbenton> someone had a very clear use case of connecting a neutron net to another with l2tp so MTU needed to be downgraded to match
22:58:38 <kevinbenton> so i think we can approve this
22:59:10 <amotoki> IIRC armax had opinions on this in a past meeting
22:59:22 <kevinbenton> armax: please release your opinions onto us in 30 seconds
22:59:43 <armax> I am ok with the it
22:59:46 <kevinbenton> yay!
22:59:48 <armax> I only fear the backlash
22:59:50 <ihrachys> \o/
22:59:52 <amotoki> :)
22:59:52 <armax> of bugs that will creep in
23:00:03 <kevinbenton> that's okay
23:00:04 <ihrachys> bugs? I write no bugs
23:00:04 <armax> because of the allow_post=True now
23:00:10 <kevinbenton> #endmeeting