17:02:07 #startmeeting service_chaining 17:02:08 Meeting started Thu Feb 11 17:02:07 2016 UTC and is due to finish in 60 minutes. The chair is cathy_. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:02:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:02:12 The meeting name has been set to 'service_chaining' 17:02:18 hi everyone 17:02:21 hi 17:02:21 hi 17:02:39 o/ 17:02:40 sorry I am a little late 17:03:12 I see prithiv LouisF johnsom. Anyone else? 17:04:41 i don't see vikram 17:04:48 any topic you have in mind for today's discussion? 17:05:03 cathy_: I'm here as well, I've some question on ODL support and timeline 17:05:18 sridhar_ram: hi 17:05:26 cathy_ You asked last week about OVN/dragonflow. I asked around our group and I found that there is more interest here in OVN at the moment. 17:05:45 johnsom: thanks for the info! 17:06:17 Yes, I think we should consider integration of newtroking-sfc with OVN in our second phase 17:07:08 Some of us have started to investigate what we need to do for integration with OVN. 17:07:23 johnsom: has there been any work to extend ovn to support sfc? 17:07:54 LouisF Sorry, I don't know. Personally I have not explored this space much. 17:08:30 johnsom: thx 17:09:03 sridhar_ram: please go ahead with your questions 17:09:09 cathy_: sure.. 17:09:43 cathy_: regarding neutron-sfc's ODL support.. I'd like to know who owns this piece in the neutron, is anyone active looking into it ? 17:10:57 sridhar_ram: Integration of SDN controllers with networking-sfc is also something we will consider in our second phase. OVN is one of the controller, ODL too. 17:11:47 cathy_: how does your timeline looks ? second phase == mid-mitaka / newton ? 17:11:56 pcarver originally proposed this work item. 17:12:41 the context for this question, as you know, is possible integration of Tacker with neutron-sfc and still bring in ODL support 17:12:46 I am afraid that we will not be able to get that done in mitaka timeline. second phase means after mitaka release. 17:12:59 Sure, that is possible 17:13:15 , just the timeline might be after mitaka release cycle. 17:14:06 cathy_: okay.. I'm trying to explore an option where Tacker can avoid going directly to ODL instead flow thru' neutron-sfc 17:14:38 there is tremendous interest in OPNFV / ODL community to expose their SFC features to VNF Manager like Tacker 17:14:55 we need to enable them.. sooner the better 17:15:10 sridhar_ram: I think that is the right way to go since networking-sfc will integrate with different SDN Controllers. ONOS Controller integration with networking-sfc is already completed. 17:15:23 ODL controller can be done in a similar way 17:15:36 cathy_: ONOS has SFC support ? 17:15:43 sridhar_ram: yes 17:16:04 cathy_: who did that ? any idea about the effort involved ? 17:16:32 There is an ONOS driver plug into networking-sfc's southbound common driver interface 17:16:57 sridhar_ram: vikram and mohan etc. did it. 17:17:11 that's quite promising ... 17:17:21 I can check with vikram about the effort and get back to you later 17:17:31 sridhar_ram: https://github.com/openstack/networking-onos 17:17:50 LouisF: thanks for posting the link 17:18:32 LouisF: thanks. will look into that.. qq, is that specific SFC or a ONOS mech driver ? 17:18:33 pcarver: are you there? 17:19:02 sridhar_ram: that is networking-sfc to onos driver 17:19:18 srithar_ram : https://github.com/openstack/networking-onos/tree/master/networking_onos/services/sfc 17:19:18 LouisF: sounds good 17:19:42 mohankumar: thx 17:19:50 cathy_: here is one line of thought.. tacker has some code to talk to ODL SFC.. we can take that as a seed and somehow speed up neutron-sfc ODL integration .. that would be awesom 17:20:14 sridhar_ram: the networking-sfc to odl driver would be similar 17:20:25 LouisF : yes 17:20:32 LouisF: sure, make sense... 17:21:07 mohankumar: that onos driver has been working and tested for some time - right? 17:21:23 LouisF : yes 17:21:24 cathy_: just for the record, tacker doesn't have any preference to go directly to ODL, we prefer to use neutron-sfc for all backends 17:22:39 sridhar_ram: OK, I think Tacker should remove the direct interface to ODL, otherwise there are two paths to ODL in Tacker which is confusing and also not compliant with general Neutron architecture 17:23:30 sridhar_ram: overall agree, it was just the we were cooking at the same time.. the good news is tacker --> odl-sfc code hasn't landed 17:24:12 sridhar_ram: I see 17:24:27 cathy_: so, there is some wiggle room... 17:24:42 sridhar_ram: :-) 17:24:50 sridhar_ram: just want to know ODL already supports SFC ? i had same thought some weeks before too 17:25:00 https://wiki.opendaylight.org/view/Service_Function_Chaining:Main 17:25:03 cathy_: ... if we can speed up ODL work here, perhaps using the WIP code already in Tacker we all go home happy :) 17:25:15 mohankumar: no yet 17:25:31 mohankumar: we will add a ODL driver to integrate with networking-sfc 17:25:32 mohankumar: Yes, ODL has quite elaborate support for SFC 17:25:57 sridhar_ram: Ahh .. ok 17:26:20 mohankumar: ODL-SFC is a fairly active group for last many year.. use to lot of GBP, now we have non-GBP sfc options 17:26:45 cathy_: fyi, ODL now has "netvirtsfc" support .. so no need for GBP to do SFC 17:27:10 *use to do lot using GBP 17:27:23 sridhar_ram: I will take a look at the non-GBP option 17:28:23 cathy_: I'd request if your and this team help to get ODL sooner.. ! 17:28:54 sridhar_ram: will try our best:-) 17:29:07 cathy_: thanks.. 17:29:30 Regarding openstack-manual int the networking guide, I think we need to add networking-sfc there. 17:29:47 johnsom: what do you think? 17:30:58 johnsom: AFAIK you have been working on other Neutron subprojects. Do the other subprojects update the openstack-manual? 17:31:46 LouisF: mohankumar do you think we should updtae that manual? 17:32:39 cathy_ , we may need to update ..but not sure 17:33:16 prithiv: any idea? 17:33:34 does any documentation team will support us to add ? 17:34:02 cathy_ Sorry, stepped away. Yes, typically the neutron project teams submit the manual updates 17:34:13 mohankumar: not sure. 17:34:16 At least the advanced services I work on do. 17:34:34 cathy_: we see where text needs to get added 17:34:56 cathy_: and engage with the doc teams 17:35:31 johnsom: thanks for the info. I think we need to submit the update patch ourselves or shall we ask the documentation team to do this? 17:36:06 We submit the patches, they correct us... grin 17:36:19 johnsom: Ok, thanks. 17:37:55 We got request on doing a release of networking-sfc based on liberty code base. IS there any back porting rule saying we should not do this way? 17:38:39 johnsom: LouisF mohankumar any idea on this? 17:38:56 cathy_ : no idea 17:39:30 There are rules about back porting, there is a doc somewhere on the wiki. Let me see if I can find it for you. 17:39:53 cathy_ Some doc examples from our team: https://review.openstack.org/#/c/226599/ https://review.openstack.org/224929 17:40:52 cathy_ This page talks to backport guidelines: http://docs.openstack.org/project-team-guide/stable-branches.html 17:41:27 johnsom: thanks. I remember reading this before. But it seems it applies to projects that already delivered a liberty release before. Anyway I will take a look at these. Thanks! 17:41:36 Yes, trye 17:41:38 true 17:42:18 I guess you could do a local "stable/liberty" branch if it doesn't require changes to the other projects 17:42:39 It would be a bit odd though. 17:42:48 johnsom: OK, thanks. that is in line with my thought. 17:42:52 johnsom: we have a stable/liberty branch in the repo 17:43:23 it does not require changes to other projects, so we should be Ok 17:44:11 So by release you were talking about a pypi release? If you already have a stable/liberty you must have tagged that already. 17:44:49 I assume that it is not against the backport rules if networking-sfc does a first release on Liberty code base. johnsom what do you think? 17:45:00 johnsom: yes, pypi release 17:45:13 Oh, ok. Yeah, no problem there 17:45:28 johnsom: great, thanks for the info! 17:45:47 That's all from my side. 17:47:31 starting from next meeting, we will start discussing new features planned for second phase. 17:47:53 If no other topic, I will end today's meeting early. 17:48:04 bye everyone 17:48:06 ok bye 17:49:02 igordcard: you joined when we will end the meeting:-) 17:49:17 #endmeeting