15:00:30 <slaweq> #startmeeting neutron_qos
15:00:31 <openstack> Meeting started Tue Apr 24 15:00:30 2018 UTC and is due to finish in 60 minutes.  The chair is slaweq. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:32 <slaweq> hi
15:00:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:34 <openstack> The meeting name has been set to 'neutron_qos'
15:00:48 <mlavalle> o/
15:01:22 <slaweq> morning with many meetings for You, right mlavalle? :)
15:01:32 <rubasov> hi again
15:01:39 <slaweq> hi rubasov
15:01:44 <njohnston> o/
15:01:52 <slaweq> hi njohnston
15:01:57 <njohnston> hi slaweq!
15:02:01 <slaweq> so let's start
15:02:02 <slaweq> #topic RFEs
15:02:03 <mlavalle> slaweq: yeah, Tuesday's meetingitis
15:02:49 <slaweq> there wasn't any new RFE related to QoS reported in last two weeks
15:03:00 <slaweq> so let's go through existing one
15:03:04 <slaweq> first is #link https://bugs.launchpad.net/neutron/+bug/1727578
15:03:05 <openstack> Launchpad bug 1727578 in neutron "[RFE]Support apply qos policy in VPN service" [Wishlist,Triaged]
15:03:14 <slaweq> Specs from zhaobo6 #link https://review.openstack.org/#/c/531074/
15:03:43 <slaweq> For me it's fine but some VPN/L3 experts should take a look on it still - mlavalle, I think You are right person to check it :)
15:04:08 <mlavalle> slaweq: I'll also have yamamoto to take a look
15:04:44 <mlavalle> but yeah, I will take a look
15:04:53 <slaweq> yes, he was commenting this RFE on launchpad a lot AFAIR, would be great if he can check it also
15:05:32 <slaweq> next one is #link https://bugs.launchpad.net/neutron/+bug/1560963
15:05:33 <openstack> Launchpad bug 1560963 in neutron "[RFE] Minimum bandwidth support (egress)" [Wishlist,In progress]
15:05:44 <slaweq> and here there isn't too much to say :/
15:06:16 <slaweq> as we talked on last meeting I sent emails to ovs-discuss and lartc MLs: https://mail.openvswitch.org/pipermail/ovs-discuss/2018-April/046549.html and http://www.spinics.net/lists/lartc/msg23542.html
15:06:33 <slaweq> but there wasn't any response there
15:07:01 <slaweq> so I still don't know how to do this min egress bw for VM
15:07:29 <rubasov> slaweq: sorry to hear that you did not get a response
15:07:30 <mlavalle> should we just start experimenting?
15:07:39 <mlavalle> with TC?
15:08:03 <slaweq> mlavalle: I would like to but for now I have still other things to do and I just don't have enough time
15:08:08 <mlavalle> slaweq: did you ever read an academic paper that I piiunted you too about two weeks ago?
15:08:21 <mlavalle> pointed^^^^
15:08:26 <slaweq> no, I didn't
15:08:39 <slaweq> You pointed to it on meeting two weeks ago?
15:08:44 <mlavalle> I think they had some good ideas. I'll dig it for you again
15:08:59 <slaweq> thx - I probably missed it somehow
15:09:01 <mlavalle> I don't remeber if it was in the meeting or in the Neutron channel
15:09:08 <mlavalle> But I'll find it for you
15:09:10 <mlavalle> again
15:09:13 <slaweq> thx a lot
15:09:58 * mlavalle feels relieved that slawek wasn't ignoring him on the QoS paper
15:10:15 * mlavalle just joking ;-)
15:10:18 <slaweq> mlavalle: :)
15:10:42 <slaweq> ok, next one - #link https://bugs.launchpad.net/neutron/+bug/1578989
15:10:42 <openstack> Launchpad bug 1578989 in neutron "[RFE] Strict minimum bandwidth support (egress)" [Wishlist,In progress] - Assigned to Slawek Kaplonski (slaweq)
15:10:59 <rubasov> I have some news for you
15:11:02 <slaweq> I just wanted to mention two patches made by lajoskatona waiting for second +2:
15:11:06 <slaweq> #link https://review.openstack.org/#/c/554532/
15:11:12 <slaweq> #link https://review.openstack.org/#/c/552938/
15:11:16 <slaweq> please review it
15:11:34 <rubasov> we had a productive hangouts discussion with jaypipes
15:11:39 <slaweq> rubasov, mlavalle: any updates about specs?
15:11:57 <rubasov> due to that gibi is preparing a new patch set for the nova spec
15:12:08 <mlavalle> rubasov: when was that?
15:12:21 <rubasov> together with that I'm preparing a new patch set for the neutron spec
15:12:29 <rubasov> mlavalle: yesterday afternoon
15:12:41 <mlavalle> afternoon your time?
15:12:48 <rubasov> slaweq: so I hope I'll have a new neutron spec patch set this week
15:12:51 <rubasov> mlavalle: yes
15:13:01 <mlavalle> I would liked to attend it
15:13:09 <mlavalle> would have^^^
15:13:17 <rubasov> mlavalle: sorry for missing you out
15:13:31 <rubasov> mlavalle: I think you may have been still sleeping then
15:14:17 <mlavalle> can you give us a summary?
15:14:22 <rubasov> mlavalle: yes
15:14:57 <rubasov> the basic discussion went around the last 30 lines of this etherpad: https://etherpad.openstack.org/p/net-bw-providers
15:15:20 <rubasov> that is: do we need cross-project standardization of the placement db contents
15:15:40 <rubasov> or can we leave it to neutron to define networking resource traits
15:16:06 <rubasov> by that getting nova and placement to treat networking rp info opaque
15:16:48 <rubasov> I think we ended up with a consensus to leave that to neutron
15:17:12 <gibi> Also we agreed to limit the scope of the current effort
15:17:21 <rubasov> gibi is around, will add a bit
15:17:28 <gibi> for example
15:17:57 <gibi> we not try to solve the mapping of the a resourece_request from a given port to a subset of the allocation in placement
15:18:08 <gibi> this could be done in the next iteration
15:18:34 <gibi> until that Neutron can fail at the port binding in complex case when the mapping is ambigouos
15:18:47 <gibi> ambiguous
15:19:21 <gibi> this also means that the Neutron API extension that adds the resource_request to the port
15:19:32 <gibi> will be optional and turned off by default
15:19:54 <gibi> to communicate that this feature is experimental
15:20:16 <mlavalle> so, what is going to change in the next iteration? will placement have more capabilities to support this feature?
15:20:39 <rubasov> we may move the rp tree description to the neutron spec
15:21:05 <rubasov> we'll cut the nova spec significantly shorter (leave the rest for a next spec)
15:21:22 <gibi> mlavalle: extending placement to provide the mapping is one possible solution
15:21:25 <mlavalle> not the next iteration of the spec. I meant the iteration of the feature
15:22:07 <rubasov> ah okay, please see gibi's answer ^^^^
15:23:59 <rubasov> we also would like to start some prototyping in neutron
15:24:11 <rubasov> but with nova mocked away first
15:24:27 <rubasov> I hope we can put together something until the summit
15:24:38 <rubasov> and show it you there
15:24:41 <slaweq> rubasov++
15:25:40 <mlavalle> what is going to happen with Nova in the meantime?
15:26:16 <gibi> mlavalle: we have to reach the point that the nova spec is accepted
15:26:40 <mlavalle> don't we have to reach the point where the neutron spec is accepted?
15:26:42 <rubasov> by the way I wanted to ask if I should add this to the topics etherpad: https://etherpad.openstack.org/p/YVR-neutron-brainstorming
15:26:43 <gibi> mlavalle: then, I will start adding the code to nova that reads the port's resource_request and incorporates it to the GET allocation_candidate request towards placement
15:27:08 <gibi> mlavalle: I don't think that will be ready before the summit
15:27:31 <gibi> mlavalle: the rubasov has a nice plan how to progress with the neutron part without the need of the nova change
15:27:33 <mlavalle> what won't be ready before the ummit?
15:28:11 <gibi> mlavalle: I hope the nova spec can be accepted before teh summit. The nova code that reads the ports' resource_request will not be ready for the summit
15:28:35 <mlavalle> what about the neutron spec?
15:28:54 <rubasov> gibi, mlavalle: I am putting that plan in the next patch set of the neutron spec
15:29:17 <rubasov> mlavalle: I think we may want to get the two specs accepted together
15:29:44 <rubasov> mlavalle: I was hoping the cross-project session on the summit would have been a good candidate for such a thing
15:30:40 <mlavalle> ok, cool, let's work on the specs
15:30:41 <gibi> rubasov: I would be happier if we can do that before the summit, but if more face to face time is needed for that then summit is a good place
15:30:58 <mlavalle> I feel the same
15:31:10 <gibi> btw both rubasov and me will be in Vancouver
15:31:12 <mlavalle> let's try to get the specs mmerged before the Summit
15:31:29 <rubasov> mlavalle: that would be even better
15:31:35 <gibi> and nova folks already proposed a nova-neutron cross project session
15:31:43 * gibi digs the link
15:31:51 <rubasov> http://forumtopics.openstack.org/cfp/details/124
15:31:55 <mlavalle> gibi: noneed to dig it
15:31:58 <gibi> rubasov was faster :)
15:32:00 <mlavalle> I am aware of it
15:32:28 <slaweq> I wasn't aware, thx gibi :)
15:32:45 <slaweq> sorry, thx rubasov
15:32:49 <gibi> :)
15:32:51 <slaweq> :)
15:32:51 <mlavalle> slaweq: it is in the Neutron weekly meeting wiki since last week
15:32:55 <rubasov> :-)
15:33:30 <slaweq> mlavalle: thx, I didn't check it
15:33:40 <mlavalle> I've been mentioning it in the past two meetings
15:33:56 <slaweq> maybe I just forgot, sorry :)
15:33:57 <rubasov> sorry for repeating myself: do we want the min bw work as a topic here: https://etherpad.openstack.org/p/YVR-neutron-brainstorming ?
15:34:32 <mlavalle> rubasov: not necessary. We are already going to discuss it anyway
15:34:41 <rubasov> mlavalle: ack, thanks
15:35:16 <rubasov> I think that's all from me in this topic
15:35:28 <mlavalle> to reiterate, let's strive to have the spces merged before the Summit
15:35:46 <rubasov> mlavalle: working on it
15:35:53 <slaweq> mlavalle++
15:36:13 <slaweq> thx rubasov and gibi for update and for working on this topic
15:36:26 <mlavalle> gibi, rubasov: thanks for your hard work and progress report
15:36:45 * gibi is happy to work on this
15:36:46 <rubasov> slaweq, mlavalle: thank you for the help and the reviews
15:37:20 <mlavalle> rubasov: I don't want to get in the way, but I also want to contribute code to this implementation
15:37:37 <rubasov> mlavalle: you're more than welcome
15:37:44 <mlavalle> cool
15:38:31 <slaweq> I think we can move on to the next rfe
15:38:34 <rubasov> mlavalle: we'll learn from you
15:38:55 * mlavalle will also learn from rubasov :-)
15:39:00 <slaweq> there is only one postponed rfe left on my list: #link https://bugs.launchpad.net/neutron/+bug/1505627
15:39:01 <openstack> Launchpad bug 1505627 in neutron "[RFE] QoS Explicit Congestion Notification (ECN) Support" [Wishlist,Triaged] - Assigned to Reedip (reedip-banerjee)
15:39:17 <slaweq> reedip has this specs still in review: https://review.openstack.org/#/c/445762/
15:39:35 <slaweq> for me it looks good, please add it to Your list of reviews :)
15:40:09 <mlavalle> is he going to implement it?
15:40:35 <slaweq> mlavalle: from what he told me few weeks ago yes, he wants to implement it
15:40:42 <mlavalle> cool
15:40:52 <slaweq> but that's all what I can say :)
15:41:10 <njohnston> I'll definitely take a look
15:41:18 <slaweq> njohnston: thx a lot
15:42:09 <slaweq> ok, let's move on
15:42:09 <slaweq> #topic Bugs
15:42:44 <slaweq> there wasn't any new bug reported for QoS in last two weeks (at least any which I am aware of)
15:42:57 <slaweq> so we have still this one:
15:43:03 <slaweq> #link https://bugs.launchpad.net/neutron/+bug/1758316
15:43:04 <openstack> Launchpad bug 1758316 in neutron "Floating IP QoS don't work in DVR router" [High,In progress] - Assigned to LIU Yulong (dragon889)
15:43:21 <slaweq> and there is already patch for that: https://review.openstack.org/#/c/558724/
15:43:55 <slaweq> I checked that with this patch FIP QoS DVR scenario test is passing, what can be seen in https://review.openstack.org/#/c/558725/5
15:44:07 <slaweq> so looks that it fixes an issue
15:44:35 <slaweq> I think that this patch also need some review from L3 experts - mlavalle :)
15:44:49 <mlavalle> slaweq: I will take a look today
15:44:58 <slaweq> thx mlavalle
15:45:10 <slaweq> and that's all what I have for today
15:45:29 <slaweq> #topic Open Discussion
15:45:40 <slaweq> anyone has something to talk about?
15:45:58 <slaweq> if no, I think we can get 15 minutes back :)
15:46:21 <njohnston> +1
15:46:38 <mlavalle> gibi, rubasov: it will be very nice to see you in Vancouver face to face
15:46:48 <slaweq> +1
15:46:54 <rubasov> mlavalle: looking forward to meeting you
15:46:58 <gibi> +1
15:47:54 <mlavalle> gibi: you are more tnah welcome to attend our team dinner on Tuesday if you want. If you decide to do so, please sign in here: https://etherpad.openstack.org/p/YVR-neutron-atendees
15:48:06 <gibi> mlavalle: thanks for the invitation
15:48:13 <gibi> mlavalle: I have to check the nova schedule
15:48:14 <rubasov> mlavalle: I already advertised it to him :-)
15:48:50 <gibi> mlavalle: but I will try to join
15:48:56 <mlavalle> ++
15:49:14 <mlavalle> slaweq: you can go ahead and close the meeting
15:49:33 <slaweq> ok, thx for attending
15:49:41 <slaweq> and see You in 2 weeks then
15:49:43 <slaweq> bye
15:49:48 <slaweq> #endmeeting