14:00:26 #startmeeting networking 14:00:27 Meeting started Tue Sep 25 14:00:26 2018 UTC and is due to finish in 60 minutes. The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:30 The meeting name has been set to 'networking' 14:00:41 o/ 14:00:45 hi 14:00:50 hi 14:01:01 hi 14:01:06 hi 14:01:10 hi 14:01:43 ok, let's get going 14:01:52 #topic Announcements 14:02:02 hi all 14:02:14 o/ 14:02:44 Our next milestone is Stein-1, October 22 - 26 14:02:59 o/ 14:03:16 The complete Stein release schedule is here: https://releases.openstack.org/stein/schedule.html 14:04:11 Yesterday I sent to the ML a summary of the meetings we had in Denver during the PTG: 14:04:17 #link http://lists.openstack.org/pipermail/openstack-dev/2018-September/135032.html 14:04:47 I hope I captured the discussions accurately 14:04:53 Thanks for that, it is very complete 14:05:46 We will keep a pointer to that summary in the announcements section of our weekly agenda, for rapid reference 14:06:11 Any other announcements? 14:06:53 ok, let's move on 14:07:00 #topic Blueprints 14:07:06 summary looks good, thanks 14:07:47 These are the blueprints that we are targeting at this point in the cycle: 14:07:54 #link https://launchpad.net/neutron/+milestone/stein-1 14:08:23 As a result of the PTG summary, I might update it over the next few days 14:08:50 Does anyone have any updates or requests for help in regards to blueprints? 14:10:02 I've been out of the loop, does anyone know the status of https://blueprints.launchpad.net/neutron/+spec/subnet-onboard ? 14:10:15 We have implemented the new solution as discussed at PTG 14:10:24 i need to know if we are ok with it 14:10:33 tidwellr_: that one is mine now, target for Stein indeed 14:11:03 https://review.openstack.org/#/c/603501/ 14:11:32 bcafarel: great! do you have a link to a review handy? 14:11:39 munimeha1: I took a look on Friday at the spec. I will look again today or tomorrow, but it seemed to me that we were close to converge on a solution 14:12:04 thanks 14:13:29 munimeha1: also, this week, per the PTG summary, I will push a patch with the evaluation criteria to bring TaaS to the Stadium. In the first version, all the questions will be "Fail" and then you and I start working on them one by one 14:13:58 we have to start from somewhere :-) 14:14:25 tidwellr_: https://review.openstack.org/#/c/348080/ (I have some local changes here on top, but that's mostly it atm) 14:14:32 munimeha1, spec need changes since with tap flow it will have data model impact 14:14:59 manjeet i updated that in version 18 14:15:27 munimeha1: thanks for the update :) 14:15:44 ok, anything else on blueprints? 14:15:45 cool 14:16:26 let's move on then 14:16:44 #topic Community goals 14:17:25 amotoki is off on a well deserved vacation this week, so no update on policy in code 14:17:47 njohnston: any updates worth mentioning on PY36? 14:18:07 The neutron-fullstack-python36 job is now voting 14:18:19 \o/ 14:18:44 njohnston: btw I sent https://review.openstack.org/#/c/604749/ -python36 job actually still runs 3.5 :) 14:18:59 it fails finding py36 though 14:19:07 I am going to ask the infra team if when we convert jobs from the legacy format tot he new zuulv3 format if we need to do anything to enable python 3 or if it comes by default 14:19:33 bcafarel: Thanks for that! I'll check on the failing to find python 3.6 issue with infra as well 14:19:46 good catch bcafarel! 14:20:25 I have one change up to convert the iptables_hybrid tempest job to the new format so I will see if I can use that to see if we get python 3 for free there 14:21:18 that's it for python 3 for now 14:21:28 thanks for the update, njohnston 14:22:05 slaweq: any updates on pre-upgrade checks? 14:22:14 mlavalle: not too much yet 14:22:24 slaweq: cool, thanks 14:22:27 but there is topic about that in ML also: http://lists.openstack.org/pipermail/openstack-dev/2018-September/134249.html 14:22:33 I'm following it for now 14:22:53 it looks that most of the common code will be in new oslo repository so I will consume it from there 14:24:05 this seems good. You have a recipe to follow and mriedem is very thorough in this type of things 14:24:23 so it should cover all or close to all that you have to do 14:24:24 yes, I only need some time to start this patch :) 14:24:48 very good 14:25:18 ok, let's move on 14:25:24 #topic Bugs 14:26:11 I sent a summary of last weeks bugs to the ML 14:26:18 #link http://lists.openstack.org/pipermail/openstack-dev/2018-September/135008.html 14:26:41 * mlavalle was beaten by boden posting the pointer :-) 14:26:58 I can read minds... that comes along with being married 14:27:20 that's tru, LOL 14:27:33 any bugs that you might want to highlight? 14:27:45 the bugs in the first 2 lists of that email still require some attention by technical experts in the respective area 14:28:56 ok, I'll go over those two lists after the morning meetings. I encourage others to do the same 14:29:04 I will just note that it appears bug https://bugs.launchpad.net/neutron/+bug/1792925 breaks ironic's grenade job... I'm assuming that's high priority for them, but not sure 14:29:04 Launchpad bug 1792925 in neutron "q-dhcp crashes with guru meditation on ironic's grenade" [Undecided,New] 14:29:45 thanks for pointing that out. I will follow up with them 14:29:46 that's all I have on this topic 14:30:33 This week our deputy is njohnston 14:30:37 In end can we discuss these issues, i have a question about a two issues - we want solution/can develop for it in Stein-2, One is for trunk sriov port multiple vlans on a direct port, another is bonding of SRIOV VFs for VRRP(all direct ports with 00:00:00:00) mac so tenant can do HA with changing macs of interfaces. Nics support this on VF but Openstack don't allow it since we have random mac in neutron port. 14:31:09 munimeha1: yes, let's discuss that at the end 14:31:13 Thanks mlavalle, I am on it 14:32:04 and also pelase note that I updated our bugs deputy calendar. removed garyk, since he doesn't work upstream anymore 14:32:32 if any other volunteers one to add their name to the calendar, please speak up 14:32:43 it's only one week every 3 months 14:33:18 ok, let's move on 14:33:22 I could also be a bug deputy 14:33:45 rubasov: of course you can :-) 14:33:54 please add my name too 14:34:11 rubasov: will the first week of Decmber be good for you? 14:34:19 it's new to me but I'll try doing my best 14:34:45 yes 14:35:04 rubasov: we have a set of guidelines in the docs, that were originally created by njohnston if I remember correctly 14:35:20 and you can always yell for help 14:35:22 mlavalle: I'm aware 14:35:32 mlavalle: I'm still digesting some of it 14:35:38 cool 14:35:48 I'm sure I will ask a few times on that week :-) 14:35:58 let's move on then 14:36:05 #topic os-ken 14:36:56 first of all, please note that I just changed the title of this topic in our agenda: https://wiki.openstack.org/wiki/Network/Meetings#os-ken 14:37:15 hongbin: any updates worth mentioning this week? 14:37:26 hi 14:37:30 #link https://review.openstack.org/#/q/project:openstack/os-ken 14:37:43 above is the list of patches that are under review in that repo 14:38:13 as we discussed, we plans the first release after those patches are merged 14:38:13 ah nice. I will review them later today 14:38:21 mlavalle: thanks 14:38:31 there is a question raised in the code review though 14:38:59 slaweq raised a question about whether os-ken should move to storyboard? or stay at launchpad 14:39:32 throughts? 14:39:54 that's a good question (which I fully expect slaweq to raise). and I think we should skip launchpad and go straight to storyboard 14:40:02 +1 14:40:09 leapfrog 14:40:21 IMHO it might be good, small project to start playing with storyboard a bit 14:40:30 and we will don't need to migrate it later 14:40:31 exactly right 14:40:55 + no need to add additional ingredients in the curent neutron launchpad soup :) 14:41:07 LOL, the soup 14:41:41 ok, i will look into how to setup the storyboard later 14:42:03 hongbin: diablo_rojo can help 14:42:29 hongbin: I will copy you in an email thread that slaweq and I have been exchanging with diablo_rojo 14:42:39 she is more than willing to help 14:42:46 mlavalle: njohnston ack, thanks 14:43:11 ok, moving on 14:43:21 #topic neutron-lib 14:43:51 boden: any updates this week? 14:44:04 I do have a few things... I will try to be quick 14:44:24 First there's a new release of lib proposed: https://review.openstack.org/#/c/604908/ looks good to me, but others can have a looksee 14:44:42 added to my pile 14:45:30 next, during the PTG we talked about identifying "current' projects and getting them updated for zuul/local testing... I started that work... details in https://etherpad.openstack.org/p/neutron-sibling-setup under the "List of networking..." section 14:45:58 I would ask folks to keep an eye out for related patches and help out when possible 14:46:28 ok, I'll lookout for those 14:46:42 we also discussed how to test with neutron-lib master... there's a discussion in https://review.openstack.org/#/c/602748/ I was hoping cores could weigh in on 14:47:23 will look at it later today 14:47:28 or tomorrow 14:47:31 added to my pile 14:47:52 tap-as-a-service is not considered as neutron sibling? 14:47:54 there's also a discussion in https://review.openstack.org/#/c/596959/ that the callback payload patches are waiting on... I added some details there a week or so ago, but waiting for some concensus 14:48:32 I'll chime in 14:48:44 munimeha1 it is not.. there's no neutron-lib in requirements https://github.com/openstack/tap-as-a-service/blob/master/requirements.txt 14:49:24 finally, we talked about adding something to consumer projects that indicates they want neutron-lib updates... mlavalle were you still planning to send a note about that or should I? 14:50:10 boden: yes.... I wrote the summary first. and it is one of my todo's there. I'll tackle them this week 14:50:27 boden: i'm sure taas requires neutron-lib. i guess it's just a missing req. 14:50:36 but if you want to do it yourself, please feel free 14:50:47 mlavalle whichever 14:51:13 yamamoto I guess so and the topic mlavalle and I are just talking about is related 14:51:35 that's enough lib fun for now since I just added a few hrs of work to mlavalle :) 14:51:45 LOL 14:51:57 thanks for the update 14:52:01 mlavalle I can send the ML note for the lib topic 14:52:10 boden: cool, go ahead 14:52:39 let's move on 14:52:47 #topic On demand agenda 14:52:52 hi 14:54:09 go ahead annp 14:54:29 mlavalle, thanks. 14:54:34 In order to support SNAT log extension, we inend to use libnefilter_log (Which is used for firewall logging) to capture packet log. So we’re planing to move libnetfilter_log to neutron-lib 14:54:46 But, neutron-lib doesn’t allow eventlet. 14:54:58 Currently, libnetfilter_log is using eventlet.green.zmp for communicating between libnetfilter_log process (privilege process) and log-application( un-privilege process) 14:55:10 So I’d like to get your advice: How do I deal with this issue or we should place libnetfilter_log in neutron? 14:55:36 Iwould say that Neutron will have to have its own implementation of libnefilter_log 14:56:01 mlavalle, I see. Thanks. 14:56:01 at least for the time being 14:56:16 we could leave a comment in the code 14:56:27 and if neutron-fwaas does the same logging then I guess we will need to live with code duplication 14:56:50 yeap, for the time being. Let's leave a todo comment 14:57:14 ok, moving on.... 14:57:19 hi 14:57:19 mlavalle, yes. I will add todo comment. you read my mind. :-) 14:57:50 munimeha1: Please file a couple of RFEs for those two things as soon as possible 14:58:30 if you file them today and we can have some exchanbge of thought over the next few days, we'll see if we can discuss them in the next drivers meeting 14:58:47 thanks, will do 14:59:07 thank you 14:59:30 ok team, we ran out of time. Thanks for attending. Have a great week! 14:59:35 #endmeeting