14:00:47 <SridarK_> #startmeeting fwaas 14:00:47 <openstack> Meeting started Tue Jun 20 14:00:47 2017 UTC and is due to finish in 60 minutes. The chair is SridarK_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:49 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:52 <openstack> The meeting name has been set to 'fwaas' 14:01:00 <SridarK_> #chair yushiro xgerman 14:01:01 <openstack> Warning: Nick not in channel: xgerman 14:01:02 <openstack> Current chairs: SridarK_ xgerman yushiro 14:01:03 <xgerman_> o/ 14:01:05 <SarathMekala> hi all O/ 14:01:10 <vks1> hi all 14:01:15 <SridarK_> #chair xgerman 14:01:17 <openstack> Current chairs: SridarK_ xgerman yushiro 14:01:41 <SridarK_> Hi All i think folks are joining in 14:02:01 <doude> Hi 14:02:39 <SridarK_> ok lets get started 14:02:44 <SridarK_> #topic Pike 14:03:27 <SridarK_> lets get started with L2 changes 14:03:40 <SridarK_> i dont see chandanc online 14:03:53 <yushiro> SridarK_, yes.... Can I start to talk about L2-agent? 14:03:53 <SridarK_> SarathMekala: do u know if chandanc will be joining ? 14:04:07 <annp> hi, sorry for late 14:04:07 <SridarK_> yushiro: yes pls 14:04:11 <yushiro> OK 14:04:28 <SarathMekala> Chandan is on PTO for 10 days.. has some personal commitments 14:04:37 <SridarK_> SarathMekala: ok 14:04:47 <SarathMekala> he will be reachable on mail 14:04:51 <yushiro> #link https://review.openstack.org/#/c/323971/ 14:04:56 <SridarK_> SarathMekala: ok thx 14:05:14 <yushiro> L2-agent patch, I'll adding UTs. slow progress.. 14:05:25 <yushiro> #link https://review.openstack.org/#/c/475183/ 14:05:54 <yushiro> This is a patch that makes default fwg configurable(on/off) 14:06:26 <yushiro> I just added a parameter [fwaas] default_firewall_group = True|False 14:06:30 <SridarK_> yushiro: ok great 14:07:07 <yushiro> Hmm, However, current impl includes so complexity for state transition. 14:07:16 <SridarK_> yushiro: hmm 14:08:02 <yushiro> e.g. If we start def_fwg = False and create fwg named 'default'. After that, we changed 'def_fwg = True' ... 14:09:00 <yushiro> But, I'll finish current impl until it works correctly. 14:09:06 <xgerman_> it should be true by default — we want people to use it ;-) 14:09:15 <yushiro> folks, please feel free to comment my patch. 14:09:18 <SridarK_> yushiro: will having it so that this behavior change can only happen on startup make it easier ? 14:09:19 <yushiro> xgerman_, yeah :) 14:09:47 <SridarK_> i presume we do this via an ini file ? 14:09:53 <SridarK_> let me look at the patch 14:10:20 <xgerman_> well, we shoudl warn people from switching it on/off and telling them only NEWLY created VMs are affected 14:10:56 <xgerman_> also redip tweeted me he is running late 14:11:04 <yushiro> SridarK_, yes, please. current impl, a definition in ml2_conf.ini ('default_firewall_group' = True/False) are loaded by server side and agent side. 14:11:34 <yushiro> xgerman_, I see. I'll take care of warning messages for switching. 14:11:36 <SridarK_> yushiro: ok - it is important that we keep it simple and there is no ambiguity 14:11:45 <SridarK_> but yes lets discuss on gerrit 14:11:53 <xgerman_> +1 14:12:03 <yushiro> SridarK_, yes. 14:12:53 <SridarK_> yushiro: anything else u want to discuss ? 14:13:08 <yushiro> from l2 agent side, nothing. 14:13:13 <yushiro> SridarK_, please going on :) 14:13:13 <SridarK_> ok 14:14:00 <SridarK_> SarathMekala: I will reach out to chandanc but dont want to disrupt his PTO - to see if there is something we can pick up to help 14:14:17 <SridarK_> SarathMekala: if u get a chance to talk to him pls mention it as well 14:14:53 <SridarK_> I think the driver changes were in pretty good shape, there is probab more integration we will need to test 14:15:11 <yushiro> SarathMekala, SridarK_ ah, I'll also rebase ovs driver's patch because it raises an exception for referring constant values. 14:15:25 <SarathMekala> sure SridarK_ 14:15:30 <SridarK_> Let me look into this and yushiro we can sync on the integration with L2 Agent 14:15:41 <yushiro> +1 14:15:52 <SridarK_> i am sure we may need some minor tweaks on the plugin as well - i can certainly look at that 14:16:20 <SridarK_> We have about a month to Pike 3 milestone 14:16:28 <SridarK_> so we can try to pull things together 14:16:39 <SridarK_> ok lets move on 14:16:56 <SridarK_> SarathMekala: pls go ahead on Dashboard changes 14:17:27 <reedip_> hi guys sorry 14:17:29 <reedip_> just caME 14:17:33 <SarathMekala> Akihiro has created a new repo for horizon dashboard 14:17:38 <yushiro> reedip_, welcome :) 14:17:45 <SarathMekala> it got approved yesterday 14:17:54 <xgerman_> awesome!! 14:17:59 <yushiro> SarathMekala, great :) 14:18:00 <SridarK_> SarathMekala: cool so now u are unblocked 14:18:12 <SarathMekala> yes SridarK_ 14:18:21 <SarathMekala> I am sending out for a review now 14:18:38 <SarathMekala> should be available by the end of this meeting 14:18:50 <SarathMekala> most of the functionality is avialable 14:18:58 <SridarK_> SarathMekala: perfect - u will need more lead time as u want review from Horizon core (Akihiro included) 14:19:00 <SarathMekala> except adding and removing ports from FWG 14:19:13 <SarathMekala> yes SridarK_ 14:19:22 <yushiro> SarathMekala, OK, please feel free to assign me as a reviewer. 14:19:33 <SarathMekala> sure yushiro 14:19:45 <SridarK_> I think we have approval rights but u will get more meaningful and useful comments from Akihiro and other Horizon cores 14:19:47 <SarathMekala> will add all the team 14:20:02 <yushiro> +1 14:20:09 <SarathMekala> yes SridarK_, Akihiro has mentioned that he will review the changes 14:20:29 <SridarK_> SarathMekala: yes from past experience - u will get very good reviews from him 14:20:31 <reedip_> ok great 14:20:41 <SarathMekala> great :) 14:20:57 <SridarK_> and i think u have enough time now to get the changes in ahead of Feature Freeze 14:21:13 <SarathMekala> yes SridarK_ 14:21:28 <SarathMekala> I will send across instructions on installing the UI plugin 14:21:29 <SridarK_> And it will easier to do some testing with the dashboard 14:21:36 <SridarK_> SarathMekala: perfect 14:21:49 <SarathMekala> if anyone has bandwidth please install it and try it out 14:21:56 <SridarK_> +1 14:22:13 <yushiro> cool 14:22:34 <xgerman_> +1 14:22:41 <yushiro> SarathMekala, Could you tell me how to enable our dashboard? enable_plugin ? 14:23:10 <SarathMekala> thats yet to be worked out 14:23:39 <SarathMekala> for now we will have to manually copy and test it out 14:23:47 <yushiro> SarathMekala, I see :) 14:23:56 <SarathMekala> devstack updates have to be looked into 14:24:37 <yushiro> SarathMekala, OK, I'll try it copy and restart apache2 service 14:24:40 <SridarK_> ok just publish the steps to get it going and we can live with that for now and gradually work in the peripheral changes 14:25:04 <SridarK_> SarathMekala: anything else u would like to discuss ? 14:25:09 <SarathMekala> sure SridarK_ 14:25:13 <SarathMekala> thats it from my side 14:25:45 <SridarK_> ok 14:25:53 <SridarK_> #topic Stadium Compliance 14:25:58 <SridarK_> reedip_: pls go ahead 14:26:47 <xgerman_> I chatted with him last night 14:26:49 <xgerman_> https://review.openstack.org/#/c/471699/ 14:26:57 <xgerman_> is a revert and he offered to options: 14:27:13 <xgerman_> a) Revert the whole patch, rework it and then publish it again 14:27:13 <xgerman_> 11:01 PM b) Partially revert the patch (https://review.openstack.org/471720) and dedicate time to fix the fwaas and midonet in parallel 14:27:53 <xgerman_> the revert will fix midonet — and I am not sure if we should revert or fix the bug (b) 14:28:07 <xgerman_> Thoughts? 14:28:18 <reedip_> hi 14:28:25 <xgerman_> there you are :-) 14:28:39 <SridarK_> IMHO the complete revert is a last resort 14:28:49 <reedip_> Well considering that midonet gate is failing, we can go ahead and revert the whole thing 14:28:52 <hoangcx_> this patch is being discussing in neutron team meeting right now :-) 14:29:01 <xgerman_> how timely 14:29:47 <SridarK_> reedip_: what is ur assessment in terms of having a fix to get midonet back on track ? 14:30:10 <reedip_> SridarK_ : Well the partial revert which I proposed passes midonet when tests are executed SERIALLY 14:30:25 <reedip_> but not parallely means there is some left over stuff which is causing the midonet gate to fail 14:30:31 <SridarK_> reedip_: but that is probab not an acceptable constraint 14:31:06 <reedip_> SridarK_ : Yes I know that the parallel execution should work in the gate 14:31:46 <reedip_> SridarK_ , xgerman : I agree complete revert would get us back in the stadium proposal and push back other patches 14:32:26 <reedip_> but letting midonet gate stop is not a very acceptable solution. We can always push our patch once we get another look at it 14:32:46 <reedip_> mlavelle might also look at it 14:33:31 <yushiro> reedip_, OK, so, we're going to revert https://review.openstack.org/#/c/471699/, right? 14:34:06 <yushiro> at the beginning of (a) 14:34:16 <reedip_> yushiro : considering the midonet gate is broken, I think so. However, I will still look into https://review.openstack.org/471720 14:34:34 <SridarK_> reedip_: ok 14:35:47 <SridarK_> I am wondering what makes it different for midonet 14:36:05 <reedip_> SridarK_ probably how the service is loaded 14:36:12 <SridarK_> since we are ok with neutron-lib 14:36:48 <reedip_> SridarK_ there were some modifications in the Unit tests as well , which might be the reason why the issue didnt materliazie 14:37:19 <reedip_> SridarK_ the thing is, if I am understanding correctly, that the extension once loaded is not unloaded , which I think yushiro also saw 14:37:22 <SridarK_> in the past too we have had issues since we have no way of telling, but usually working with yamamoto has resulted in the proper fix 14:38:45 <reedip_> SridarK_ let me discuss it once with him tomorrow morning and add him as a reviewer to the new patch for neutronlib migration 14:39:03 <yushiro> reedip_, +1 It'll so helpful !! 14:39:11 <xgerman_> +1 14:39:51 <SridarK_> reedip_: ok even if we can do a partial back out will be preferrable but if this needs more time then we may need to go with the revert and double back for the final fix 14:39:58 <SridarK_> reedip_: thx for following up on this 14:40:14 <xgerman_> +1 14:40:24 <xgerman_> when do we decide if we need to full revert? 14:40:33 <reedip_> right now ? 14:40:37 <SridarK_> reedip_: do u think we can allocate a time to do some debugging together ? 14:41:02 <reedip_> SridarK_ xgerman_ : if you guys have time, we can discuss this issue after meeting 14:41:05 <SridarK_> reedip_: do u want to talk to yamamoto and then we can go for it 14:41:23 <yushiro> reedip_, please sync with me tomorrow :) 14:41:26 <SridarK_> reedip_: right after mtg - i have a conflict 14:41:33 <SridarK_> can we try tomorrow ? 14:41:39 <reedip_> SridarK_ xgerman_ or tomorrow evening 0200 UTC ? 14:41:57 <reedip_> tomorrow morning * 14:42:15 <xgerman_> 10 am EST 14:42:18 <reedip_> It would sync up with me and yushiro 14:42:20 <xgerman_> works 14:42:58 <SridarK_> is that 10am EST or 10pm EST ? 14:43:02 <reedip_> 10 PM EST :) 14:43:09 <xgerman_> am 14:43:45 <SridarK_> xgerman_: 14:00 would be 10am 14:43:46 <reedip_> xgerman_ 0200 UTC is 10 AM EST ? 14:44:00 <reedip_> no not 1400 UTC, 0200 UTC :) 14:44:19 <xgerman_> Google told me wrong 14:44:28 <SridarK_> it works for me 14:44:34 <SridarK_> but xgerman_ it is a bit late for u 14:44:54 <xgerman_> I cna do it if I get the day right ;-) 14:44:59 <SridarK_> :-) 14:45:16 <SridarK_> ok so lets do that, i think it will be a good time for yushiro also 14:45:41 <reedip_> ok, done ... its 0730 for me, so I will leave office at 0630 :P 14:45:42 <SridarK_> so reedip_ u are saying in about 11hrs and 15 mins ? 14:45:52 <reedip_> Roughly 14:45:58 <SridarK_> ok cool 14:46:06 <SridarK_> lets ping each other on fwaas channel 14:46:19 <yushiro> 2:00 UTC is good for me 14:46:33 <SridarK_> i can set up a webex if screen sharing and voice makes it easier too 14:46:36 <yushiro> Ok, I'll keep on joining IRC. 14:47:00 <SridarK_> will be good to assess the damage and some plan for remediation - even if we do have to revert 14:47:38 <reedip_> Ok done 14:47:46 <SridarK_> reedip_: thx for staying on top of this 14:47:51 <reedip_> should we move on now , 13 minutes left 14:48:13 <SridarK_> yes, reedip_ anything else on any related stadium topic ? 14:48:49 <SridarK_> ok lets move on 14:48:55 <reedip_> SridarK_ Not as of now 14:49:19 <SridarK_> #topic Performance Improvement 14:49:32 <SridarK_> annp_: hoangcx_: things u would like to discuss ? 14:49:50 <SridarK_> i think we missed this due to time on last few mtgs 14:49:50 <hoangcx_> SridarK_: Oops, No. I think we are Ok with FWaaS v1. 14:50:17 <hoangcx_> Now we are proposing the same solution to Neutron (without configurable) 14:50:22 <SridarK_> and how are things panning out with SG - are we making the changes 14:50:29 <SridarK_> ah ok u read my mind :-) 14:50:46 <yushiro> :) 14:51:05 <hoangcx_> It is getting good comment from Kevin that we don't need to support configuratable in neutron 14:51:21 <SridarK_> hoangcx_: ok great 14:51:22 <hoangcx_> Just use Netlink solution. Now, we are adopting 14:51:51 <SridarK_> Sounds good 14:52:01 <yushiro> hoangcx_, coooooool 14:52:02 <hoangcx_> this is the work: https://review.openstack.org/#/c/470912/ 14:52:11 <SridarK_> +1 14:52:22 <SridarK_> #topic Open Discussion 14:52:50 <SridarK_> So priorities for this week - resolution of neutron-lib issues and L2 integration 14:53:02 <SridarK_> possibly for next 2 weeks 14:53:05 <yushiro> Yes. 14:53:19 <SridarK_> we have roughly 4 - 5 weeks 14:53:55 <xgerman_> also looking ahead PTG is coming and other teams are already starting their etherpads 14:53:56 <reedip_> ok 14:54:04 <SridarK_> Is anyone planning or will have budget for PTG ? 14:54:26 <doude> Hi, I just like to informe you I started to work on a patch that will permits to use a driver for fwaas service plugin which is not based on the Neutron DB 14:54:27 <yushiro> yes, try to get travel support program :) 14:54:31 <reedip_> I may not have the budget for the PTG 14:54:38 <SridarK_> doude: hi 14:54:41 <reedip_> doude : link ? 14:54:47 <yushiro> Maybe I can.. 14:54:58 <SridarK_> doude: what is the use case ? 14:55:02 <doude> I'm working on that to add supprt of the fwaas extension for Contrail neutron core plugin 14:55:35 <doude> which is not use the ML2 framework or the Neutron db model 14:55:45 <reedip_> doude : better to have a bug ( in case you havent logged one yet ) 14:55:48 <doude> we don't want to duplicate the source of trust 14:56:00 <xgerman_> #link https://etherpad.openstack.org/p/fwass_ptg_denver 14:56:10 <reedip_> doude : any reason for moving away from ML2 framework? 14:56:13 <SridarK_> doude - Edouard ? 14:56:19 <yushiro> xgerman_, thanks!! 14:56:20 <xgerman_> so we can collect whose attending and who is not and/or skip the PTG all together 14:56:24 <doude> I did not yet pushed the code, it's still a very early step 14:56:31 <doude> I'll in the week I hoe 14:56:35 <doude> hope* 14:56:59 <SarathMekala> doude is this for Open Contrail? 14:57:29 <yushiro> Maybe SarathMekala, and chandan knows about Contrail?? 14:57:41 <SridarK_> doude: ok cool - welcome - i think we spoke earlier at BCN ? 14:57:42 <doude> Not for the new networking-opencontrail project but for the actual contrail neutron plugin 14:58:19 <doude> new project networking-opencontrail will be based on the ML2 framework 14:58:39 <doude> but not it just bootstrapped in the last summit 14:58:49 <reedip_> ok 14:58:56 <SridarK_> ok almost at time 14:58:58 <doude> me I wok on the core driver actually used and deployed in our clients 14:59:03 <reedip_> doude : do share the link 14:59:06 <SridarK_> lets continue more next mtg 14:59:08 <reedip_> or add us in the patch 14:59:17 <SridarK_> ^^ +1 14:59:21 <doude> yes reedip_, we discuss about that bug https://review.openstack.org/#/c/474524/1 14:59:22 <yushiro> doude, hope to join next meeting :) 14:59:33 <SridarK_> ok folks lets call it - we are at time 14:59:40 <doude> yes I'll and I hope to have some code to show 14:59:40 <SridarK_> have a great week and thx for joining 14:59:47 <reedip_> see u tomorrow 14:59:51 <SridarK_> #endmeeting