21:02:17 <markmcclain> #startmeeting Networking 21:02:18 <openstack> Meeting started Mon Apr 7 21:02:17 2014 UTC and is due to finish in 60 minutes. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:02:21 <openstack> The meeting name has been set to 'networking' 21:02:30 <Swami> hi 21:02:31 <emagana> sorry I am late! 21:02:38 <emagana> Hello Folks! 21:03:02 <Sam-I-Am> hi emagana 21:03:24 <fawadkhaliq> hello everyone! 21:03:44 <emagana> Sam-I-Am: Hi There! 21:04:08 <markmcclain> #link https://wiki.openstack.org/wiki/Network/Meetings Agenda 21:04:23 <markmcclain> #topic Announcements 21:04:37 <markmcclain> We're in the process of closing out RC2 21:05:14 <markmcclain> We've have backported any of the committed fixes that were tagged -rc or -backport potential 21:05:16 <emagana> markmcclain: cheers! 21:05:40 <markmcclain> #topic Bugs 21:06:03 <markmcclain> Are there any critical bugs that necessitates RC2 staying open? 21:06:07 <salv-orlando> a delayed aloha from me. 21:06:28 <salv-orlando> parsed all high bugs this morning. None really critical for me. 21:06:36 <mestery> markmcclain: Did you merge the ODL bug into RC2 from rkukura? 21:06:47 * mestery looks for that bug. 21:07:06 <markmcclain> Otherwise we're likely to publish RC2 soon and barring any critical bugs this will be the final version 21:07:16 <markmcclain> mestery: I did and amotoki fixed my backport :) 21:07:27 <mestery> markmcclain amotoki: Thank you guys both very much! :) 21:07:38 <mestery> That was the main one I was tracking markmcclain. 21:07:54 <markmcclain> mestery: https://review.openstack.org/#/c/85511/ 21:08:01 <markmcclain> Any other bugs? 21:08:04 <mestery> markmcclain: thanks! 21:08:20 <amotoki> we already have the difference between icehouse and juno :) 21:08:27 <phil_h|2> ran into one today: https://bugs.launchpad.net/neutron/+bug/1303998 21:08:28 <uvirtbot> Launchpad bug 1303998 in neutron "vm fails with error vif_type=binding_failed using gre tunnels" [Undecided,New] 21:09:01 <phil_h|2> It is new not a rehash of the old one 21:09:28 <Sam-I-Am> which is interesting because i sent phil my working config files 21:10:06 <salv-orlando> phil_hi2: can you exclude it has not the same root cause as https://bugs.launchpad.net/neutron/+bug/1276391 21:10:09 <uvirtbot> Launchpad bug 1276391 in neutron "ML2 MechanismDriver.bind_port() called inside transaction" [High,In progress] 21:10:10 <markmcclain> phil_h|2: thanks looks like we'll need to triage this more 21:10:31 <phil_h|2> I will help all I can 21:10:40 <phil_h|2> it is a show stopper for me 21:11:01 <rkukura> phil_h|2: Can you check the “neutron agent-show” output for the L2 agents and make sure you see “gre” in tunnel_types? I can followup with you this evening or tomorrow if you like. 21:11:23 <phil_h|2> Yes, I get that 21:11:29 <phil_h|2> all show GRE 21:11:47 <markmcclain> ok the sooner we can get to the bottom of this the better 21:11:54 <salv-orlando> phil_hi2: can you share also neutron.log and your neutron.conf? 21:12:05 <markmcclain> I'll tag the bug right now as a blocker until we understand it better 21:12:09 <phil_h|2> yes 21:12:44 <markmcclain> For those that can discuss let's resume this after the meeting 21:13:03 <markmcclain> Last call for other RC critical bugs 21:14:34 <Sukhdev> markmcclain: it wil be good to include this https://bugs.launchpad.net/neutron/+bug/1227336 21:14:35 <uvirtbot> Launchpad bug 1227336 in neutron "ml2 plugin update operations" [Medium,In progress] 21:15:05 <Sukhdev> uvirtbot: talk about timing :-) 21:15:06 <uvirtbot> Sukhdev: Error: "talk" is not a valid command. 21:15:21 * markmcclain looking 21:15:47 <markmcclain> Sukhdev: concerned about stablity 21:16:03 <markmcclain> generally at this point we want small patches 21:16:15 <Sukhdev> markmcclain: please read my comments on the patch - I think we should consider it 21:16:24 <banix> markmcclain: https://review.openstack.org/#/c/83217/ 21:16:30 <salv-orlando> the patch is 'just' exception handling. Still, I see it at something that will be released with the first stable update 21:16:54 <salv-orlando> unless Sukhdev convinces me something is fundamentally broken and needs to be fixed 21:16:55 <rkukura> markmcclain: Only non-test code change is passing a flag to 3 existing function calls 21:17:33 <banix> the original patch is now marked as work in progress; the one I posted above is essentially 3 lines. 21:17:35 <markmcclain> rkukura: right…the release management team has requested that we have a really high bar for changes at this point 21:17:49 <rkukura> Sukhdev, markmcclain: I’d be happy to have that go in for RC2, but if not, I think it and the binding outside transaction patch could be backported for the 1st stable release 21:17:53 <markmcclain> and that unless critical we should wait 2014.1.1 21:18:20 <markmcclain> rkukura: agreed 21:18:39 <markmcclain> also for the first .1 release we'll have lots of test runs in the CI pipeline too 21:18:41 <salv-orlando> rkukura, mestery: how do feel about stability and reliability of ml2 as it is (ie: without these patches? 21:19:28 <mestery> salv-orlando: I think overall the ML2 stability has been good, but I'd be curious to hear about others that have deployed RC1 as well. :) 21:19:44 <rkukura> I’m not sure either of these are issues that have been hit in the gate, or reported from failures in deployments 21:20:06 <Sukhdev> salv-orlando, rkukura: It gives us opportunity to bake in the ML2 driver behaviour sooner than later - if it makes any sense to you 21:20:19 <markmcclain> ok, so we'll track them as important for stable/icehouse, but not a release blocker 21:20:45 <rkukura> Sukhdev: I agree we should get these in ASAP, but I don’t feel they block the RC 21:20:51 <salv-orlando> mestery: I was looking for somebody more knowledgeable than me for an opinion on ML2 quality. Look like we're good 21:20:52 <banix> markmcclain: makes sense 21:21:19 <salv-orlando> Sukhdev: Don't mind but you've not convinced me yet… I would still vote for stable/havana 21:21:23 <rkukura> Hopefully phil_h|2’s issue isn’t a big one ;) 21:21:30 <salv-orlando> however, I'm not part of the stable team so my opinion counts 0 ;) 21:21:46 <Sam-I-Am> salv-orlando: so you're unstable? :P 21:21:47 <markmcclain> rkukura: yeah agreed that is the one I think we need to understand quickly 21:21:56 <salv-orlando> Sam-I-Am: I would say bipolar 21:22:12 <markmcclain> Ok if anyone has any critical release blocking bugs please contact me directly 21:22:16 <markmcclain> #topic Docs 21:22:18 <markmcclain> emagana: hi 21:22:29 <emagana> markmcclain: Hi there! 21:22:36 <emagana> markmcclain: Hi there! 21:22:45 <emagana> sorry... new IRC client! 21:23:10 <emagana> to make it easy I updated the wiki with all the changes on the Docs 21:23:56 <emagana> Sam-I-Am did a great job with a lot of commits 21:23:57 <irenab> emagana: please update this one as Merged too: https://bugs.launchpad.net/openstack-manuals/+bug/1283334 21:23:58 <uvirtbot> Launchpad bug 1283334 in openstack-manuals "Add support to request vnic type on port" [Medium,Fix released] 21:24:05 <emagana> phil_h|2 and nick as well 21:24:21 <emagana> Thanks irenab, I will do that! 21:24:57 <Sam-I-Am> the ml2 sections in the install guide work, at least on a few systems 21:25:06 <emagana> Sam-I-Am has been reaching the right people for his questions, I think we are in good track 21:25:12 <Sam-I-Am> it was marked 'works' before the nova-net stuff :P 21:25:55 <Sam-I-Am> emagana: can we talk about OVS in the install guide? 21:26:12 <emagana> So, I added to the wiki the lists of commits that has been completed. If somebody has a question just reach us out 21:26:27 <emagana> Sam-I-Am: I would prefer to have just ML2 21:26:41 <emagana> OVS and Linuxbridge should be completely removed 21:26:42 <Sam-I-Am> me too! just curious what the opinion is in here 21:26:58 <emagana> is anyone having a different opinion? 21:27:05 <emagana> speak now or .. you know the rest! 21:27:10 <Sam-I-Am> the migration issue is what prevent deprecation of ovs in icehouse, right? 21:27:43 <markmcclain> Sam-I-Am: we have the script now 21:27:54 <Sam-I-Am> markmcclain: thats what i saw 21:28:03 <Sam-I-Am> i dont see a reason for a new install to use ovs 21:28:07 <markmcclain> so I think we're ready to officially announce deprecation and removal in Juno 21:28:10 <salv-orlando> isn't it a bit too early to removed docs for ovs & lb plugins? 21:28:17 <mestery> OVS and LB should be deprecated in Juno right markmcclain? 21:28:28 <Sam-I-Am> salv-orlando: this would just be from the install guide 21:28:34 <markmcclain> We announced them as deprecated in Havana 21:28:37 <emagana> markmcclain: +1 21:28:43 <markmcclain> so we only need 1 more cycle of support 21:28:59 <mestery> Cool 21:29:00 <markmcclain> of course nothing stops us from including in Juno except the amount of deadcode 21:29:20 <amotoki> where is the best place to decsribe the step to migration from ovs/lb to ml2 in the doc? install-guide? 21:29:24 <rkukura> I think we need to make it completely clear in the docs, etc, that juno will not contain the old monolithic plugins, so migration must be done with icehouse before moving to juno 21:29:26 <markmcclain> my thinking is that we announce that this is the last release with OVS and LB plugins 21:29:33 <markmcclain> and schedule remove for J-2 21:29:35 <emagana> markmcclain: I think we should highlight the "deprecation" in the summit and remove the code for Juno release 21:29:41 <markmcclain> so that will give us 12 wks 21:29:58 <markmcclain> emagana: it will be in the release notes for Icehouse 21:30:04 <salv-orlando> the deprecation should be highligthed in release notes 21:30:08 <Sam-I-Am> amotoki: in the past, migration was in the ops guide 21:30:11 <emagana> markmcclain: I do agree! 21:30:14 <salv-orlando> sorry for repeating markmcclain 21:30:18 <emagana> salv-orlando: I will make sure that is the case! 21:30:22 <Sam-I-Am> amotoki: there is a bug for documenting havana->icehouse 21:30:31 <Sam-I-Am> amotoki: the install guide is primarily for new installs 21:31:00 <markmcclain> Upgrade docs are important too 21:31:13 <amotoki> Sam-I-Am: understand, but the ops-guide is not bound to a specific release in general. 21:31:13 <markmcclain> emagana: any other doc items to highlight/ 21:31:15 <markmcclain> ? 21:31:23 <emagana> markmcclain: Nothing else 21:32:12 <markmcclain> amotoki: I will include upgrade notes in the release notes 21:32:12 <Sam-I-Am> amotoki: maybe we need a migration guide :P 21:32:15 <markmcclain> emagana: thanks for the update 21:32:20 <Sam-I-Am> amotoki: because, moar guide 21:32:26 <markmcclain> #topic Design Summit 21:32:34 <markmcclain> #link http://summit.openstack.org/ 21:32:35 <amotoki> Sam-I-Am: agree. i can help on it. 21:32:55 <markmcclain> deadline is April 20th 21:32:59 <Sam-I-Am> amotoki: it would be good to have one or more devs from each project working on such a guide 21:33:23 <emagana> Sam-I-Am and amotoki: Let's have that conversation on the mailing list 21:33:41 <markmcclain> remember it is not first to file, so if you have a similar proposal feel file a duplicate and we'll sort it out 21:34:25 <markmcclain> also remember the design summit should include a healthy balance of sessions of items that impact the common core (ie testing, parity work) and those that innovate it 21:34:40 <markmcclain> Any questions on the design summit? 21:35:42 <markmcclain> #topic Tempest 21:35:44 <markmcclain> mlavalle: hi 21:35:47 <mlavalle> hi 21:35:59 <mlavalle> we have continued reviewing api tests 21:36:17 <markmcclain> great 21:36:18 <mlavalle> last week we merged another on. We have merged 18 of 27 that we are tracking 21:36:37 <mlavalle> and we have 2 that only require one more +2 21:36:52 <mlavalle> last week we also recovered all the abandoned patchsets 21:37:00 <markmcclain> that's good news 21:37:05 <mlavalle> the bad news is that a new one showed up 21:37:21 <mlavalle> but we only have that one abandoned and I will tak to the author 21:38:00 <mlavalle> big thanks to HenryG and SumitNaiksatam whoe helped track down the correct FWaaS test 21:38:16 <SumitNaiksatam> mlavalle: np, thanks for reaching out 21:38:18 <markmcclain> cool good to see everyone helping out 21:39:05 <markmcclain> Any Tempest questions? 21:39:15 <mlavalle> Can nati_ueno give a review to https://review.openstack.org/#/c/83627/? 21:39:23 <nati_ueno> markmcclain: sure 21:39:28 <nati_ueno> mlavalle: sure. 21:39:29 <nati_ueno> sorry mistake 21:39:32 <mlavalle> and that's all I have 21:39:39 <markmcclain> nati_ueno: thanks 21:39:43 <markmcclain> mlavalle: thanks for the update 21:39:49 <mlavalle> nati_ueno: thanks 21:39:51 <amotoki> mlavalle: is there a list to review? just go to gerrit? 21:40:06 <markmcclain> So the current thinking is that we'll enable full tempests for Neutron after RC2 is cut 21:40:17 <mlavalle> amotoki: I will send a list to the ML with the 8 patchsets that haven't merged yet 21:40:26 <HenryG> We need some strategy to avoid duplicate work in tempest. 21:40:38 <amotoki> mlavalle: nice. 21:40:56 <markmcclain> HenryG: what duplicate work? 21:41:35 <HenryG> I have encountered many reviews that are the same APIs being tested. 21:41:56 <HenryG> The fwaas was just the latest example. 21:42:42 <mlavalle> HenryG: in my perception, I think there was some duplication, but not a lot 21:42:56 <mlavalle> and I rather have that :-) 21:43:31 <HenryG> OK, if I am the only one then no problem :) 21:43:36 <mlavalle> in any case, I can send periodically a list to the ML with the tests identified and that will allow contributors to see if there is duplication 21:43:47 <markmcclain> mlavalle: that sounds like a good plan 21:43:52 <salv-orlando> it's not uncommon to have several people working at the same thing. This often happens with bugs as well 21:44:16 <markmcclain> yeah we had 4 proposals for the same bug this cycle :) 21:44:27 <sc68cal> Gotta game those stackalytics metrics 21:44:38 <sc68cal> ;) 21:44:57 <salv-orlando> sc68cal: tip for that… fix docstrings and other spelling errors 21:45:17 <markmcclain> So enabling the full tempest test on Neutron is likely to uncover a few more errors 21:45:25 <enikanorov> sc68cal: i've seen this! :> 21:45:26 <enikanorov> ;) 21:45:38 <markmcclain> and could increase the pain for short-term 21:46:28 <markmcclain> but activating this now will pay off as we can fix these items earlier than enabling in J-1 21:46:41 <markmcclain> Any other testing items? 21:46:56 <HenryG> Is grenade also going to start voting? 21:46:56 <amotoki> to make it easy to review in tempest, it seems better we have "neutron" or "network" in summary line :) 21:47:12 <markmcclain> HenryG: we're not quite ready to enable grenade, but close 21:47:13 <amotoki> it helps we search similar reviews. 21:47:45 <HenryG> amotoki: +1 21:48:03 <markmcclain> #topic Juno 21:48:23 <salv-orlando> markmcclain: re grenade there is a patch in review for making it voting 21:48:36 <salv-orlando> seems failure rate improved a lot over the weekend 21:48:37 <markmcclain> #undo 21:48:38 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x2171e50> 21:48:57 <jlibosva> markmcclain: salv-orlando voting was merged few hours ago 21:48:59 <markmcclain> salv-orlando: awesome I had missed that review 21:49:12 <markmcclain> jlibosva: thanks glad it is now voting 21:49:16 <salv-orlando> jlibosva: thanks a lot! 21:49:43 <mestery> I've seen a failure already on a patch where grenade failed in fact. :) 21:50:07 <markmcclain> It is big deal for us to have grenade voting 21:50:23 <markmcclain> thanks to jlibosva and all that assisted 21:50:33 <Sam-I-Am> is grenade the right place to check configs used for gating? 21:51:14 <markmcclain> Sam-I-Am: configuration compatibility? 21:51:28 <markmcclain> or just the configs we use? 21:51:39 <salv-orlando> Sam-I-Am: if youre interested in config files, every job has a directory with them 21:51:44 <Sam-I-Am> just the configs you use... it seems to be a good way to check things for the install guide 21:52:03 <markmcclain> then salv-orlando is right it should be available in any tempest based job 21:52:07 <markmcclain> #topic Juno 21:52:16 <Sam-I-Am> salv-orlando: yeah, i've been using gate-tempest-dsvm-neutron 21:52:25 <markmcclain> Speaking of config compatibility… this landed in the last week https://review.openstack.org/#/c/84367/ 21:52:48 <markmcclain> We have no removed the last vestiges of Quantum compatibility from the tree 21:52:52 <markmcclain> s/no/now/ 21:53:04 * sc68cal dances a jig 21:53:18 <markmcclain> the only remaining references are in vendor backend code 21:53:23 <salv-orlando> please observe a minute of silence in memory of quantum 21:53:45 <Sam-I-Am> a quantum leap? 21:53:57 * mestery waits for all the Quantum jokes to land. :) 21:54:29 <markmcclain> haha 21:54:47 <amotoki> we still have many "quantum" in XML API :) 21:55:03 <markmcclain> amotoki: we have to keep the quantum namespace for XML 21:55:16 <markmcclain> because it is baked into our API spec 21:55:33 <amotoki> markmcclain: yes :) 21:55:43 <markmcclain> Speaking of Juno I think we as a community should consider dropping XML in K 21:55:58 <markmcclain> or possibly J, but I will put that on next week's agenda 21:56:20 * sc68cal will pack his asbestos undies 21:56:27 <mestery> sc68cal: Ha! :) 21:56:37 <markmcclain> yeah should be a fun discussion 21:56:50 <markmcclain> I will begin removing the temporary -2s this week 21:57:00 <salv-orlando> I think we should do the following 21:57:09 * mestery hopes we can get George Reese to attend the Neutron meeting next week. 21:57:14 <salv-orlando> deliberately inject a bug into XML serialization 21:57:18 <salv-orlando> wait for a few weeks 21:57:27 <salv-orlando> nobody complains --> nobody uses it 21:57:36 <markmcclain> salv-orlando: there were bad bugs in XML for months that nobody noticed 21:58:15 <markmcclain> especially around serialization 21:58:19 <salv-orlando> markmcclain: so we do not even need to do this. It's probably already bitrotting 21:58:44 <markmcclain> salv-orlando: some would argue that is the case 21:58:50 <markmcclain> so anyway should be fun next week 21:59:17 <markmcclain> Lastly if you've got a review that -2'd by me and abandoned, please restore so that I can remove the -2 21:59:46 <markmcclain> Ok we've hit our time limit for this week 22:00:37 <markmcclain> Thanks to everyone for their work testing RC1 and please test RC2. If you find any release critical issues, please contact me directly so that I can notifiy everyone who needs to track it 22:00:53 <markmcclain> Have a good week and talk to everyone on the mailing list and IRC 22:00:55 <markmcclain> #endmeeting