Monday, 2018-02-12

*** persia has left #openstack-dragonflow01:56
*** yamamoto has joined #openstack-dragonflow03:54
*** yamamoto has quit IRC06:05
*** yamamoto has joined #openstack-dragonflow06:15
*** Natanbro has joined #openstack-dragonflow06:35
oansonMorning06:53
Natanbromorning06:56
irenabmorning07:01
*** yamamoto has quit IRC07:20
*** yamamoto has joined #openstack-dragonflow08:21
lihiGood morning08:21
*** yamamoto has quit IRC08:29
*** snapiri has quit IRC08:56
*** yamamoto has joined #openstack-dragonflow09:43
*** yamamoto has quit IRC09:44
*** snapiri has joined #openstack-dragonflow09:53
*** yamamoto has joined #openstack-dragonflow10:45
*** yamamoto has quit IRC10:51
*** yamamoto has joined #openstack-dragonflow11:20
*** yamamoto has quit IRC11:32
*** yamamoto has joined #openstack-dragonflow12:02
*** yamamoto has quit IRC14:24
*** irenab has left #openstack-dragonflow14:48
*** yamamoto has joined #openstack-dragonflow14:57
*** Natanbro has quit IRC15:07
*** mlavalle has joined #openstack-dragonflow15:34
mlavalleoanson: are you going to Dublin?15:37
*** yamamoto has quit IRC17:49
*** yamamoto has joined #openstack-dragonflow17:54
*** yamamoto has quit IRC18:00
*** yamamoto has joined #openstack-dragonflow18:11
*** yamamoto has quit IRC18:11
oansonHi. Weekly starting in a minute in #openstack-meeting-418:59
*** yamamoto has joined #openstack-dragonflow19:12
*** yamamoto has quit IRC19:19
mlavalleoanson: hey20:10
oansonHi20:10
mlavalleI was writing you an email20:11
mlavalleif it is too late for you, go to bed and respond to my message tomorrow morning20:11
oansonNo, not too late.20:11
mlavalleok, so let's talk here then20:12
oansonIt's not daylight saving time, so it's not too late here :)20:12
oansonSure20:12
mlavalleI messed up the time for the meeting this week20:12
mlavallewe fell to standard time for the the winter20:12
mlavalleso the meeting is at 1pm my time20:12
mlavallenot 2 pm like during our Summer20:13
oansonI see. Well you didn't miss much, I was alone.20:13
mlavalleI read the log20:13
oansonYes. So you saw we practically messed the time every week this year? :)20:13
mlavallebut I want you to know that this is the correct week for the late night meeting (your time)20:14
mlavalleit was me who missed it by 1 hour20:14
oansonGoogle calendar says it is, so I'll go with that. I'm guessing anyone looking it up will take the ics file in the eavesdrops page as gold standard20:14
mlavalleyes20:14
mlavallein the open agenda topic you mentioned that you are going to be in Dublin Mon - Wed20:15
oansonYes20:15
mlavalleAre you all day long on Wednesday?20:15
oansonYes. We leave Thursday.20:15
mlavallewho is we?20:15
mlavallebesides you20:15
oansonirenab and snapiri20:16
mlavalleNice!20:16
mlavallewill be glad to see them20:16
oansonYeah. I'm guessing irenab you already know. I'll be sure to introduce snapiri20:17
mlavalleso, you added the "port-behind-port" topic to the Neutron PTG etherpad20:17
mlavalleyou still want to discuss it in Dublin?20:18
oansonI still want the RFE. Yes.20:18
mlavalleok, I'll schedule it to be discussed20:18
oansonI think it should make life easier in the long run.20:18
mlavalleHelp us undertand your vision ;-)20:18
oansonAnd I think we can settle most of the big issues live faster than asynchronously on the bug page.20:19
oansonmlavalle, will try :)20:19
mlavallewe are a little hard headed, but have good intentions.... LOL20:19
mlavalleI have another two topics:20:20
oansonNo worries :)20:20
oansonSure20:20
mlavalleNext one is https://bugs.launchpad.net/neutron/+bug/169042520:21
openstackLaunchpad bug 1690425 in neutron "[RFE] neutron cells aware" [Wishlist,Triaged]20:21
mlavallewe've been discussing this RFE on and off in the drivers meeting over the past couple of months20:22
oansonI haven't seen that one. Reading up cells is on my todo list, but I haven't got that far.20:22
oanson(I see armax added a link to an introduction)20:22
mlavallewell, we are worried about not being able to  scale up Neutron when users start deplying big cells 2 in Nova20:22
mlavallewe agreed that we want to pursue a POC using this approach: https://www.youtube.com/watch?v=R0fwHr8XC1I20:23
mlavallebut I wonder is the publisher / subscriber DB approach of Dragonflow would be useful for this20:24
oansonI see. I've been talking with Ken on a Dragonflow/oslo.messaging/AMQP approach20:24
mlavalleGreat!20:25
oansonI think the pub/sub is supposed to overcome the bottleneck by being a 'fire and forget' mechanism. Most chances are the message will make it. But the sender doesn't wait for a reply.20:25
mlavalleah ok20:25
oansonThe broker can also be less concerned with anything more than a 'best effort' mechanism.20:25
mlavalleso you think https://www.youtube.com/watch?v=R0fwHr8XC1I is a good approach?20:26
oansonAnd Dragonflow is supposed to be resilient - if a message is missed, eventually we'll notice by re-syncing with our distributed database20:26
oansonI have to watch the video first :)20:26
oansonBut I can review it tomorrow and give my opinion.20:27
mlavalleso if you find that interesting and I schedule the topic to be discussed in Dublin, would you be interested?20:27
mlavallein participating20:27
oansonSure20:27
mlavallegreat, let mw know tomorrow what you think20:27
oansonSure. On the surface, it looks very interesting.20:28
mlavallenow let me dig a topic you and dimak brought up after the Boston summit during a Neutron meeting20:29
mlavalledo you remember it?20:29
oansonNot of the top of my head :)20:29
mlavallelet me dig then20:29
mlavallehang on20:29
oansonAh, L3 flavour20:29
mlavalleyes20:29
oansonWe kinda let that one slide due to lack of resources.20:30
mlavallehttp://eavesdrop.openstack.org/meetings/networking/2017/networking.2017-05-15-21.01.log.html#l-9920:31
oansonHas someone shown interest in this? Or is this a gentle reminder? :)20:31
mlavalleLOL20:32
mlavallethe ODL guys are adding a feature to theeir plugin where they need to use the L3 favors framework20:32
oansonI'm scared to ask which question you found funny... xD20:32
mlavallethe gentle reminder thing. No, I am not to try to boss you around20:33
mlavalleand they have found gaps in the L3 flavors framework20:33
mlavallethey are the first ones to actually use it20:33
mlavalleI am considering to have a conversation about it in Dublin20:34
mlavalleif you are interested, I can schedule it on Wednesday so you can attend20:34
oansonI think yamamoto had a patch adding some missing events a long while back.20:34
oansonSure. That'd be great.20:34
oansonThanks!20:34
mlavalleyeah but it got a little more complicated than that20:35
oansonDid they say what they are missing?20:35
mlavallethis way, when / if you get back to that topic we discussed back in May, you already know the experience from the ODL guys20:36
oansonThis is yamamoto's patch, btw: https://review.openstack.org/#/c/483173/20:36
oansonYes. That would be very helpful.20:36
mlavallethe real hard work has been done on the ODL side20:38
mlavalleand they discovered that they needed to react to events in a certain order: https://review.openstack.org/#/c/541766/20:38
mlavallelast week they tried to implement something with SQLAlchemy events and I said no way20:38
oansonYes, we had some strange behaviour because some events were called before others as well.20:39
mlavalleso I thought all this might be useful for you20:39
mlavalleand of course you can provide valuable feedback20:39
oansonSure. I'll go over it. (Also on the referenced patch, which looks interesting as well)20:40
mlavalleI am glad we connected20:40
mlavallewe ended up having the meeting at the time I thought it was originally :-)20:40
mlavalleSee you in Dublin20:40
oansonHeh :)20:40
oansonYep. Looking forwards to seeing you!20:41
oansonThanks for updating me!20:41
mlavallethe other thing I wrote in that message I was going to send you....20:41
mlavalleis that I cannot commit to make it to all the odd week meetings20:42
mlavalleBut I'll do my best to attend some20:42
mlavalleI think you and your team have much to give to the wider Neutron community20:42
oansonThanks for the feedback. It would be great to have you.20:43
oansonThe truth is I wanted to contribute back to Neutron as well20:43
oansonIt's just ends up being many plates in the air20:44
mlavallefinally, I am looking to develop neutron cores20:44
mlavalleif you or anyone in your team can increase reviews count on the Neutron side, we need you20:44
oansonLet me see what I can do. I remember feeling that I don't understand the code I was reading last time I tried to review. But that was a long while back. Let me try again :)20:46
mlavallecool20:46
mlavallelet you go to rest and enjoy your family20:46
oansonNah. The kid is long asleep by now. :)20:46
oansonEnjoy your afternoon!20:47
mlavallehow old is he / she?20:49
oansonHe is 2 years old (and change).20:50
mlavalleenjoy him! I have a boy of 28 and a girl of 2420:50
oansonThanks. I am! :)20:51
*** ChanServ has quit IRC21:02
*** ChanServ has joined #openstack-dragonflow21:10
*** barjavel.freenode.net sets mode: +o ChanServ21:10
*** mlavalle has quit IRC23:25
*** mlavalle has joined #openstack-dragonflow23:27
*** mlavalle has quit IRC23:29

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!