21:03:28 <markmcclain> #startmeeting Networking 21:03:29 <openstack> Meeting started Mon Jun 3 21:03:28 2013 UTC. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:03:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:03:32 <openstack> The meeting name has been set to 'networking' 21:03:33 <danwent> armax: you're closest to arosen, can you kick him :) 21:03:45 <markmcclain> haha 21:03:50 <armax> hang on :) 21:04:13 <markmcclain> #link https://wiki.openstack.org/wiki/Network/Meetings 21:04:15 <armax> done 21:04:15 <gongysh> what a long leg ... 21:04:17 <arosen> Hi 21:04:17 <armax> :) 21:04:22 * salv-orlando if my leg was 6,000 miles long I could have kicked him 21:04:26 <dkehn> hi 21:04:34 <maroh_1> hi 21:05:05 <markmcclain> any changes to the agenda before we get going? 21:05:27 <markmcclain> #topic Announcements 21:05:40 <markmcclain> #info H1 was released last week 21:06:00 <garyk> hi, all 21:06:05 <markmcclain> H1 included the initial work commit for ML2 21:06:45 <markmcclain> if you have some spare cycles take a moment to test the builds and report back any findings 21:07:02 <markmcclain> H2 will be released July 18th 21:07:12 <amotoki> hi all 21:07:18 <markmcclain> #link https://launchpad.net/quantum/+milestone/havana-2 21:07:29 <markmcclain> we have 50 blueprints targeted 21:07:42 <markmcclain> and is way over our normal velocity 21:07:45 * salv-orlando is having a laugh 21:08:33 <markmcclain> I'd like to ask that the subteam leads take a look through the H2 blueprint list and refine the items that need to be in H2 21:08:50 <markmcclain> also we have glut of Medium blueprints 21:09:56 <markmcclain> With that let's run through the sub teams reports 21:09:58 <markmcclain> #topic API 21:10:04 <salv-orlando> hello 21:10:16 <salv-orlando> nothing major to report if not that we've merged a few patches. 21:10:49 <salv-orlando> I will refine blueprints targeting H-2 as suggested by markmcclain 21:11:01 <salv-orlando> and unless I've missed it, we do not have any major bug 21:11:34 <markmcclain> ok.. cool 21:12:12 <gongysh> we need more testing to find some 21:12:54 <markmcclain> we've got one bug I've highlighted, but we can talk about it when we talk Nova 21:13:07 <markmcclain> Any questions for salv-orlando? 21:13:25 <salv-orlando> gongysh: yeah the API lacks coverage especially when it comes to scale and performance 21:13:39 <salv-orlando> but that involves several components, not just the API layer. 21:14:15 <markmcclain> marun is also work on improving the API test coverage too 21:14:43 <markmcclain> #topic VPNaaS 21:15:00 <markmcclain> nati_ueno: how are things progressing 21:15:12 <mlavalle> markmcclain: we are doing our bit with Tempest also 21:15:13 <markmcclain> looks like StrongSwan is back? 21:15:34 <markmcclain> mlavalle: good point 21:16:33 <markmcclain> nati_ueno: around? 21:17:09 <markmcclain> We'll skip to Nova and come back to VPN when he returns 21:17:12 <markmcclain> #topic Nova 21:17:15 <nati_ueno> hi 21:17:17 <nati_ueno> sorry 21:17:29 <nati_ueno> May I report it ? 21:17:43 <markmcclain> nati_ueno: sure 21:17:56 <nati_ueno> VPN has two important bp for H2 21:17:59 <nati_ueno> markmcclain: Thanks 21:18:05 <nati_ueno> [1] CRUD API and DB model for IPSec https://blueprints.launchpad.net/quantum/+spec/vpnaas-python-apis 21:18:14 <nati_ueno> In review 21:18:14 <nati_ueno> Quantum Advanced Service Plugin for VPNaas (Patchset 5) https://review.openstack.org/#/c/29812/ 21:18:19 <nati_ueno> Quantum Client for VPNaaS (Patchset 3) https://review.openstack.org/#/c/29811/ 21:18:24 <nati_ueno> [2] Agent Impl https://blueprints.launchpad.net/quantum/+spec/ipsec-vpn-reference 21:18:35 <nati_ueno> About to finish investigation. I'm start implementing 21:18:46 <nati_ueno> As mark said, it looks like we can use strongswan 21:19:10 <nati_ueno> using nswrap script by Francois 21:19:16 <nati_ueno> That's all from me 21:19:58 <markmcclain> cool.. thanks for the update 21:20:08 <markmcclain> any VPNaaS questions? 21:20:16 <garyk> MarkAtwood: there is an open review for the host parameter to the plugin 21:21:04 <garyk> MarkAtwood: i saw Nova 21:21:18 <garyk> markmcclain: we talking about nova now? 21:21:24 <markmcclain> yes 21:22:06 <markmcclain> garyk: you're talking about the extension to notify quantum of the host? 21:22:20 <garyk> markmcclain: yes. i'll post the review in a sec 21:22:24 <markmcclain> ok 21:22:45 <garyk> markmcclain: https://review.openstack.org/#/c/29767/ 21:22:56 <garyk> this is really holding back the multi host feature 21:23:34 <gongysh> we need some core review on nova side 21:24:08 <markmcclain> right.. let me work on getting Nova cores to review it 21:24:32 <gongysh> great 21:24:33 <garyk> markmcclain: thanks 21:25:05 <markmcclain> I also wanted to ask who can look into this bug: https://bugs.launchpad.net/quantum/+bug/1160442 21:25:08 <uvirtbot> Launchpad bug 1160442 in quantum "when boot many vms with quantum, nova sometimes allocates two quantum ports rather than one" [High,Incomplete] 21:25:38 <garyk> markmcclain: this is on my to do list for tomorrow. 21:25:57 <emagana> markmcclain: I was able to reproduce that one 21:25:58 <markmcclain> ok.. also looks like arosen has been digging into to it too 21:26:14 <arosen> I've actually seen this occur on our cloud too but haven't had time to track it down :/ 21:26:21 <emagana> markmcclain: suppose to provide more logs to arosen and gary 21:26:43 <markmcclain> emagana: ok.. thanks for posting the first set 21:26:50 <arosen> I was supprised emagana was able to reproduce this with just one HV. I haven't been able to do that. 21:26:58 <garyk> markmcclain: arosen emagana: i'll work on this tomorrow. 21:26:59 <emagana> just got stuck into tasks but I will update the launchpad page 21:27:08 <gongysh> it seems we have already minds on it before I want to jump in. 21:27:28 <SumitNaiksatam> i had earlier posted the logs in the launchpad anwers 21:27:39 <SumitNaiksatam> difficult to figure out with just the logs 21:27:48 <SumitNaiksatam> need to put debug statement 21:28:23 <markmcclain> ok.. I just wanted to make sure we've got folks working on it as it's current rated high 21:28:32 <markmcclain> garyk: any update on Nova migration? 21:28:39 <garyk> markmcclain: sorry, no. 21:29:06 <markmcclain> ok 21:29:11 <markmcclain> Any other Nova items? 21:29:24 <garyk> markmcclain: there is a little update which may be worth exploring 21:29:48 <markmcclain> sure.. what's the update? 21:30:02 <garyk> markmcclain: it may be libvirt centric, but there may be a flag in libvirt whcih can be set which can enable a tap device to be moved from a linux bridge to an open vswitch. 21:30:41 <garyk> markmcclain: i need to investigate a little more. if so then mabe we need to move that support to nova and later this be leveraged to enable a migration. 21:30:46 <garyk> it is still early days 21:31:00 <markmcclain> oh cool.. that would help out 21:31:28 <markmcclain> #topic Security Groups 21:31:35 <markmcclain> arosen: any updates? 21:31:52 <arosen> no updates from me 21:31:57 <markmcclain> alright 21:32:02 <markmcclain> #topic FWaaS 21:32:11 <markmcclain> SumitNaiksatam: quick update? 21:32:14 <SumitNaiksatam> hi 21:32:18 <SumitNaiksatam> more progress in the past week 21:32:36 <SumitNaiksatam> RajeshMohan has posted a patch for the iptables driver for the ref impl 21:32:48 <SumitNaiksatam> KC has started work on client and CLI 21:32:56 <markmcclain> cool 21:33:00 <SumitNaiksatam> i need to do more work on the API patch 21:33:04 <markmcclain> ok 21:33:10 <SumitNaiksatam> both iptables driver and cli is being documented in the same fwaas google doc as before: https://docs.google.com/document/d/1PJaKvsX2MzMRlLGfR0fBkrMraHYF0flvl0sqyZ704tA 21:33:32 <SumitNaiksatam> RajeshMohan: u there for an update? 21:33:45 <RajeshMohan> you covered it all - nothing from me 21:34:06 <SumitNaiksatam> ok, i think there is good info in the google doc 21:34:14 <SumitNaiksatam> pl check and comment 21:34:40 <gongysh> is the link in the BP desc? 21:34:44 <SumitNaiksatam> we might have to rework the review patch since it currently is not based off the API patch 21:34:53 <SumitNaiksatam> gongysh: yeah, its in the bp 21:35:09 <SumitNaiksatam> we are actually using the same google doc that we used for the other fwaas bp 21:35:14 <gongysh> ok, I don't want to scan the meeting log to find the doc. 21:35:28 <SumitNaiksatam> just want to make sure that we find all info in one place 21:35:28 <RajeshMohan> The description for IPTables BP has the li nk 21:35:34 <markmcclain> SumitNaiksatam: ok 21:35:56 <SumitNaiksatam> thats the quick update 21:36:04 <markmcclain> SumitNaiksatam: can you update the blueprint assignees with person leading the dev work? right now you're still listed for a few of them 21:36:27 <SumitNaiksatam> yeah, i will be posting patches for those (plugin and agent) 21:36:41 <SumitNaiksatam> other bps already assigned 21:36:48 <markmcclain> oh.. I that was farmed out to others on the team 21:36:51 <amotoki> SumitNaiksatam: gongysh: each BP has the spec link, but there is no list which covers all of FWaaS.... It is better to have a summary page. 21:36:55 <markmcclain> we're good then 21:37:19 <SumitNaiksatam> plugin and agent might be smaller patches so i was thinning of rolling them into the api patch, might make it easier to test 21:38:29 <SumitNaiksatam> amotoki: summary page is already present 21:38:36 <SumitNaiksatam> just finding the link, one sec 21:38:37 <markmcclain> yeah it nicer to be able to test end to end 21:38:43 <markmcclain> I think the summary is here: 21:38:43 <markmcclain> https://blueprints.launchpad.net/quantum/+spec/quantum-fwaas 21:39:03 <SumitNaiksatam> https://wiki.openstack.org/wiki/Quantum/FWaaS/HavanaPlan 21:39:15 <SumitNaiksatam> markmcclain: thanks, that too 21:39:43 <markmcclain> ok any other FWaaS questions? 21:39:44 <amotoki> SumitNaiksatam: thanks, good summary! 21:40:22 <markmcclain> SumitNaiksatam: thanks for the update 21:40:41 <markmcclain> #topic LBaaS 21:40:49 <markmcclain> LBaaS is bit behind schedule 21:41:02 <markmcclain> to get back on track the team is going to focus on getting this blueprint done 21:41:02 <markmcclain> https://blueprints.launchpad.net/quantum/+spec/multi-vendor-support-for-lbaas-step0 21:41:26 <markmcclain> once this step is complete others who want to write drivers will have stable base to work from 21:41:53 <markmcclain> This is currently in review: https://review.openstack.org/#/c/28245/ 21:42:22 <markmcclain> Also once this step is complete, LBaaS team will resume working on some of the other items on on the roadmap 21:42:27 <markmcclain> Any LBaaS questions? 21:42:54 <nati_ueno> Is there any conclusion for inherit VS composition ? 21:42:55 <garyk> i have started t review https://review.openstack.org/#/c/28245/ 21:43:14 <garyk> once again i am sleeping. i have a really minor nit with the last pacth 21:43:42 <markmcclain> composition is something we really should look at, but we need to do this on a project wide basis 21:44:17 <nati_ueno> markmcclain: I agree. so this is related to VPN code also. We should do composition in new code ? or wait some big refactoring? 21:44:39 <markmcclain> I expect there is going to be some necessary refactoring 21:44:41 <gongysh> all mixin into components? 21:45:06 <markmcclain> that is ahead in the future for us to align with the other projects that moving to Pecan & WSME 21:45:41 <markmcclain> the current wsgi framework which we inherited is planned for deprecation in Oslo and other projects 21:45:59 <markmcclain> I expect to talk about this in Hong Kong more than what we touched on in Portland 21:46:23 <markmcclain> nati_ueno: to answer your question specifically, I think we should strive to be consistent for now 21:46:35 <nati_ueno> so you are expecting we will work on the refactoring in next release? 21:46:45 <nati_ueno> markmcclain: OK so we should choose inherit for now? 21:46:46 <markmcclain> the nice thing about drivers is that we should be able to refactor with minimal disruption to vendor impls 21:47:27 <salv-orlando> nati_ueno: I think the gist of the discussion is: whatever you choose there's going to a major refactoring (likely to happen over 2 release cycles) 21:47:34 <markmcclain> nati_ueno: refactoring will be a large task 21:48:07 <nati_ueno> I got it 21:48:35 <markmcclain> salv-orlando is correct.. this will be a long term process 21:48:47 <markmcclain> #topic ML2 21:49:27 <rkukura> the initial ml2 code was merged, but we still need core reviewers for the devstack patch: https://review.openstack.org/#/c/27576/ 21:50:08 <rkukura> without this, the ml2 plugin isn't going to get sufficient testing as various merges change it 21:50:09 <markmcclain> garyk: can you take a look at it? 21:50:32 <garyk> markmcclain: sure 21:50:51 <markmcclain> thanks! 21:51:12 <markmcclain> Any other ML2 updates? 21:51:13 <rkukura> salv-orlando had looked at an earlier version, so I'd appreciate if he takes another look 21:51:13 <nati_ueno> so new feature for ovs and linuxbridge plugin should go to ml2? 21:51:31 <salv-orlando> rkukura: np. That patch worked well for me. 21:51:44 <markmcclain> nati_ueno: ideally once we're at fully parity yes 21:51:45 <gongysh> will we set ML2 plugin as default one in devstack? 21:52:12 <rkukura> To get to full parity, we need https://blueprints.launchpad.net/quantum/+spec/ml2-gre 21:52:16 <nati_ueno> markmcclain: so kyle is going to add VXLan for ovs, this work should go to ml2 or not? 21:52:30 <markmcclain> gongysh: eventually yes.. as there will be only one OpenSource plugin 21:52:51 <nati_ueno> rkukura: ah I got it, after the ml2-gre, we will quit to get new feature for ovs and linuxbridge. 21:53:00 <rkukura> I've been focusing on trying to make sure the agent side of the various tunnel-related work will all work with ml2 21:53:28 <markmcclain> nati_ueno: there will be a few bits of overlap 21:53:35 <rkukura> So there is vxlan work going on right now for both agents, plus work on partial mesh and l2-population 21:54:23 <rkukura> All of these potentially effect the tunnel management RPCs that ml2 needs to implement for gre 21:54:37 <markmcclain> the vxlan work for the current plugin is relatively small 21:54:39 <nati_ueno> markmcclain: rkukura: OK please let me know when we quit to add new feature for ovs and lb. 21:54:44 <markmcclain> I think it makes sense to still review it 21:54:46 <rkukura> Its been suggested to have a weekly IRC meeting on all this 21:55:07 <markmcclain> rkukura: do we need a weekly meeting 21:55:13 <rkukura> Also, I filed 5 BPs today for H-2 21:55:25 <markmcclain> or one time meeting to get the next steps organized and then move back to the ML? 21:55:29 <markmcclain> *mailing list 21:55:47 <rkukura> We can't do all that without more people involved, so a weekly meeting might make sense 21:56:01 <markmcclain> k 21:56:02 <rkukura> Of the BPs, only the GRE one is high priority 21:56:07 <markmcclain> ok 21:56:39 <markmcclain> Anything else? 21:56:50 <rkukura> So I'd really like to get names on most of these soon, or take them off h-2 21:56:57 <gongysh> Why is GRE so different? 21:57:04 <rkukura> that covers it for now - see the agenda for links, etc. 21:57:14 <rkukura> GRE is needed for parity right now. 21:57:30 <gongysh> I mean the implementation 21:57:32 <markmcclain> gongysh: we'll have to talk about GRE offline as we're running short on time 21:57:33 <rkukura> There's a BP for VXLAN, which would be needed for parity if it gets in openvswitch first 21:57:37 <gongysh> ok 21:58:12 <markmcclain> #topic CLI 21:58:28 <markmcclain> A tarball for 2.2.2a was created last week 21:58:40 <gongysh> I still cannot see it in https://pypi.python.org/pypi/python-quantumclient 21:58:56 <gongysh> any problems found with the tarball? 21:58:57 <markmcclain> gongysh: since it was tagged as an alpha 21:59:05 <markmcclain> it won't show up in PYPI 21:59:09 <markmcclain> http://tarballs.openstack.org/python-quantumclient/python-quantumclient-2.2.2a.tar.gz 21:59:19 <markmcclain> is the link for folks that want to test 21:59:54 <markmcclain> I hope to push it out to PyPI later today 22:00:14 <markmcclain> #topic Testing 22:00:19 <markmcclain> mlavalle: Quick update? 22:00:44 <mlavalle> I will send an email to the ML with what will be delivered for Havana in Tempest 22:00:48 <mlavalle> this week 22:01:02 <markmcclain> awesome.. we'll be on the lookout for it 22:01:08 <markmcclain> #topic Horizon 22:01:16 <mlavalle> I am also cleaning up gate-tempest-devstack-vm-quantum-full 22:01:27 <mlavalle> and I need help from arisen this week 22:01:31 <mlavalle> arosen 22:02:01 <arosen> mlavalle: sure. If this on the security group stuff and ping? 22:02:05 <markmcclain> cool.. will this make the gate more stable? 22:02:10 <mlavalle> yes, security groups 22:02:17 <arosen> mlavalle: lets talk in #openstack-dev after and i'm happy to help. 22:02:20 <nati_ueno> markmcclain++ 22:02:32 <nati_ueno> mlavalle: let's me help too 22:02:40 <mlavalle> arosen: ;-) 22:02:57 <mlavalle> nati_ueno: ;-) 22:03:00 <markmcclain> amotoki: Looks like you've got assignees for the Quantum related blueprints in Horizon 22:03:15 <amotoki> yes. i updated Wiki. no big progress from last week. H-2 started, so i will check the target and progress of each blueprint this week. 22:03:57 <amotoki> that's all from me this week. 22:04:59 <markmcclain> great.. thanks for the update 22:05:04 <markmcclain> #topic Open Discussion 22:05:47 <garyk> i am going to crash. good night all 22:05:49 <nati_ueno> markmcclain: IMO, we should also manage service framework progress 22:05:52 <markmcclain> garyk: night 22:05:59 <nati_ueno> garyk: have a good dream :) 22:06:36 <emagana> ciao ciao 22:06:49 <markmcclain> thanks for the all of the hard work on H1 22:07:31 <markmcclain> next week we'll look that blueprints since 50 is way too many 22:07:39 <markmcclain> have a great week everyone 22:07:41 <markmcclain> #endmeeting