21:02:16 <markmcclain> #startmeeting Networking
21:02:17 <openstack> Meeting started Mon Feb 24 21:02:16 2014 UTC and is due to finish in 60 minutes.  The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:02:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:02:20 <openstack> The meeting name has been set to 'networking'
21:02:27 <markmcclain> #link https://wiki.openstack.org/wiki/Network/Meetings Agenda
21:02:34 <emagana_> markmcclain: last minute changes
21:02:53 <markmcclain> Mostly the standard agenda with one special item this week
21:03:03 <markmcclain> #topic Announcements
21:03:24 <markmcclain> #info Icehouse-3 merge window closes at the end of the day on March 4th
21:04:43 <markmcclain> We have lots of in progress reviews and many will need to several iterations to ready for merging
21:04:59 <markmcclain> it is important both the proposers and reviewers be responsive this week
21:05:11 <markmcclain> #link https://launchpad.net/neutron/+milestone/icehouse-3
21:05:55 <markmcclain> Review for blueprints not in Icehouse-3 are being deferred to Juno
21:06:54 <markmcclain> #topic Bugs
21:07:08 <markmcclain> #link https://bugs.launchpad.net/neutron/+bugs?search=Search&field.importance=Critical&field.status=New&field.status=Confirmed&field.status=Triaged&field.status=In+Progress
21:07:25 <markmcclain> We have 4 open critical bugs.  Most of these we have been carrying for some time
21:07:48 <markmcclain> The incident rate of https://bugs.launchpad.net/neutron/+bug/1253896
21:08:29 <salv-orlando> markmcclain: I think we should reduce its criticality to high.
21:09:03 <markmcclain> salv-orlando: that's my thinking as weel
21:09:08 <markmcclain> well
21:09:21 <salv-orlando> same for bug 1273386 (not sure what to with it since the bug is actually still there)
21:09:40 <salv-orlando> but now we know it's not critical at least
21:09:58 <markmcclain> yeah… for 1273386 reducing makes sense too
21:10:06 <salv-orlando> Sorry for keeping posting on this topic, but also 1112912 an 1230407 should be retriaged
21:10:22 <salv-orlando> I think the first can be closed, and the latter perhaps is not anymore valid now
21:10:42 <markmcclain> I'm planning on including in the release notes that running an network node and nova-compute on same host when key injection is enabled should be avoided on 12.04
21:10:51 <salv-orlando> on the other hand, I would like to raise to critical the two top offenders for the full neutron tempest job
21:11:00 <salv-orlando> plus one bug which was filed by lifeless
21:11:30 <markmcclain> I missed the merge on #1112912
21:11:33 <markmcclain> I've updated the status
21:11:50 <salv-orlando> The bug I am talking about are:
21:11:50 <salv-orlando> 1
21:11:58 <salv-orlando> 1) bug 1283522
21:12:08 <salv-orlando> https://bugs.launchpad.net/neutron/+bug/1283522
21:12:20 <salv-orlando> 2) https://bugs.launchpad.net/neutron/+bug/1265495
21:12:50 <salv-orlando> can't find the one lifeless filed now...
21:12:56 <arosen> Hiya sorry running late.
21:13:15 <lifeless> salv-orlando: is it in this list - https://bugs.launchpad.net/tripleo/ ?
21:13:16 <markmcclain> ok looks like enikanorov_ is looking into 1283522
21:13:43 <salv-orlando> lifeless… the one about "queue pool limit x of y"...
21:14:27 <salv-orlando> bug 1282482 I think
21:14:37 <markmcclain> I pushed https://bugs.launchpad.net/neutron/+bug/1265495 back to critical
21:14:43 <armax> bug #1282421 might need attention too as it may add instability to the gate
21:14:53 <armax> I have a fix, but I am working on UT's
21:15:31 <markmcclain> armax: great
21:15:53 <armax> http://bugs.launchpad.net/neutron/+bug/1282421
21:16:04 <armax> looks line the bot is not working today?
21:16:10 <armax> s/line/like
21:16:17 <salv-orlando> the bot appears to not be running. Hey openstack are you there?
21:16:39 <salv-orlando> looks like it's not
21:17:35 <markmcclain> yeah not sure what is going on with linking bugs
21:18:15 <markmcclain> re 1273386 looks like we still have don't have full crash dump for the kernel team to use
21:18:48 <markmcclain> but I agree that moving high makes sense now that we have reasonable workaround for the upstream issue
21:18:52 <salv-orlando> yeah I got pinged for providing a repro
21:19:09 <markmcclain> Any other bugs to discuss?
21:19:13 <salv-orlando> but I had other things to tend to last week unfortunately
21:19:28 <salv-orlando> I don't think there is any other bugs which deserves the team's attention here.
21:19:46 <salv-orlando> For a full list of bugs to fix for having the full jobs working decently, see the mailing listy
21:21:00 <markmcclain> yeah.. I include a full tempest further down the agenda
21:21:12 <markmcclain> #topic Docs
21:21:15 <markmcclain> emagana_: hi
21:21:22 <emagana_> markmcclain: Hi there!
21:21:55 <emagana_> Basically, want to share with the team the work that some folks are doing for improving and cleaning the Installation Guide
21:22:10 <markmcclain> cool
21:22:48 <markmcclain> now that we're getting close to the feature freeze have we begun make sure any changes with DocImpact have a corresponding doc bug?
21:22:56 <emagana_> There was some feedback (complains) about Installation Guide, therefore some work is in progress to make it very EASY to follow
21:23:15 <emagana_> markmcclain: let me take that in a minute
21:23:44 <emagana_> So, for the new installation guide: https://wiki.openstack.org/wiki/Documentation/InstallGuideChanges we have decided to only post information for OpenSource components
21:24:02 <emagana_> Any vendor specific will be move to the Apendix section!
21:24:30 <markmcclain> ok.. makes sense
21:24:33 <emagana_> There are some tickets opened agains the installation guide to add vendor-specific info, please do not work on this ticket becasue installation guide will change
21:25:08 <emagana_> I would encourage all vendors to have their own documentation in a public place that we can include in the guide
21:25:41 <markmcclain> ok
21:26:05 <emagana_> markmcclain: about your previous question. Yes, this is the time were we start mapping features with Docs.. WIll be sure we are covering anything new!
21:26:28 <salv-orlando> emagana_: I think it might be better to mark them as invalid. Many people working on vendor specific docs are not in this room
21:26:41 <markmcclain> salv-orlando: good point
21:26:46 <salv-orlando> and I would for sure forget to pass on the information to the rest of my team ;)
21:26:57 <Sukhdev> emagana_: will you provide some guidelines for the vendors to follow?
21:27:09 <emagana_> salv-orlando: good point, thanks!
21:27:46 <emagana_> Sukhdev: The wiki above will be the main point of reference
21:28:07 <emagana_> It hasn't been decided how the communication with the vendor will be done!
21:28:14 <emagana_> I will bring it to the docs meeting
21:28:39 <emagana_> markmcclain: I am done!
21:29:21 <Sukhdev> emagana_: so, if I want to add anything vendor specific, can I go and update this wiki? will that be OK?
21:30:13 <markmcclain> emagana_: I think you'll probably want to send out a message tagged for Neutron to the ML since many have strict filtering in place
21:30:15 <emagana_> Sukhdev: Installation Guide should not contain vendor specific information in main section but the Apendix will be available for that. Yes, please add in the Wiki anything that you want to be covered
21:30:26 <markmcclain> emagana_: thanks for the update
21:30:38 <emagana_> markmcclain: Good idea, I will do that!
21:30:52 <markmcclain> #topic Nova Parity
21:31:14 <markmcclain> beagles updated the agenda
21:31:48 <markmcclain> he had a conflict and could not make the meeting today
21:31:53 <markmcclain> #topic Tempest
21:32:07 <markmcclain> #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027797.html
21:32:27 <markmcclain> sdague started a thread on making the final push for full tempest testing by Icehouse-3
21:32:46 <markmcclain> salv-orlando dug further into the gap
21:32:58 <markmcclain> #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027862.html
21:33:11 <markmcclain> salv-orlando: want to share more on what you found?
21:33:31 <salv-orlando> basically the two bugs I want to be moved to critical are the ones we *must* fix
21:33:53 <salv-orlando> then there are a plethora of minor bugs both in tempest and/or neutron, which appear to have less frequency
21:34:06 <salv-orlando> but I think we have enough people on the team and we should be able to fix all of them
21:34:40 <salv-orlando> in particular I think the one enikanorov assigned to himself might be due to agents reporting while api requests are being processed
21:34:56 <markmcclain> yeah I was wondering the same thing myself
21:34:57 <salv-orlando> so I was wondering if we should still think about separating API and RPC server
21:35:19 <markmcclain> I'm leaning that direction
21:35:35 <markmcclain> arosen and carl_baldwin have both done work to make this easier
21:35:46 <marun> carl_baldwin has a patch that will give us that capability: https://review.openstack.org/#/c/72565/
21:36:17 <markmcclain> right this will enable multiple processes
21:36:22 <carl_baldwin> Doesn't totally separate RPC and API yet.
21:36:37 <marun> if it's running in a different process, how is it now separating it?
21:36:42 <markmcclain> but there is still one process that is potentially doing both
21:36:47 <salv-orlando> that patch will allow to have some processes which only process APIs and others which only process RPC requests from the message bug?
21:36:48 <carl_baldwin> They are still run under the same parent process.
21:36:52 <salv-orlando> bug/bus
21:36:52 <marun> uh
21:36:58 <marun> what does that have to do with ti?
21:37:25 <marun> in the case of the api_workers = 1, that would mean the only task of the parent process would be making sure there is a live process handling api requests.
21:37:46 <marun> other than that, it would be handling rpc
21:38:14 <salv-orlando> ok, let's move that to review then. No need to keep using the meeting's time
21:38:41 <markmcclain> yeah.. let's see what implementing this will do to address the testing issue
21:39:04 <markmcclain> might get a win by running more than one RPC process in the gate… other services do this
21:39:15 <markmcclain> salv-orlando: Anything else on the full run?
21:39:33 <salv-orlando> nothing which is not already on the mailing list.
21:40:25 <markmcclain> salv-orlando: Thanks for looking into the final gaps and id'ing the bugs
21:40:29 <markmcclain> mlavalle: hi
21:40:35 <mlavalle> hi
21:40:52 <mlavalle> we have continue making steady progress with api testing
21:40:58 <markmcclain> great
21:41:20 <mlavalle> after the message I sent to the ML last week, Tempest core are doing more reviews of api patches
21:41:42 <mlavalle> in fact, 4 were merged over the pat 7 days
21:41:49 <markmcclain> good news
21:41:56 <mestery> Great news mlavalle!
21:42:04 <mlavalle> I updated the agenda with the pointers
21:42:30 <mlavalle> so I still feel condindent that by the end of icehouse, we will be done
21:42:43 <markmcclain> yeah same here
21:42:49 <markmcclain> thanks for updating the agenda
21:43:00 <mlavalle> that's all I have, unless there are questions
21:43:16 <mlavalle> mestery: thanks :-)
21:43:23 <markmcclain> mlavalle: thanks for the report
21:43:33 <markmcclain> #topic L3
21:43:35 <markmcclain> carl_baldwin: hi
21:44:56 <markmcclain> Looks like he's no longer around… moving on since we're running short on time
21:45:01 <markmcclain> #topic IPv6
21:45:03 <markmcclain> sc68cal:  hi
21:45:08 <sc68cal> hello
21:45:19 <sc68cal> I'd like to thank salv-orlando for his comments on some of the IPv6 reviews
21:45:26 <sc68cal> very helpful
21:45:43 <sc68cal> bear with me - i have prepared remarks
21:45:46 <salv-orlando> We might use another core reviewer there. The patches are fairly mature and I think once the comments are addressed they should be good to go.
21:45:48 <carl_baldwin> hi
21:46:03 <sc68cal> i can wait if we want to go back to l3
21:46:14 <markmcclain> let's continue with v6
21:46:17 <sc68cal> k
21:46:23 <sc68cal> We have one high priority fix that I'd like to see get merged by the merge window, IPv6 allow RAs from certain routers https://review.openstack.org/#/c/72252/ . My concern is that one reviewer has identified a larger issue with how neutron creates routers for IPv6 subnets, that is not addressed by the aforementioned review. I've attempted to get him to clear his -1 on the patch and file a bug for the issue he has identified, because it is
21:47:35 <sc68cal> The hairpin bug is looking pretty good - thanks to mestery for +1'ing the review
21:47:45 <sc68cal> just need one more nova core to sign off
21:47:48 <markmcclain> I'll look at 72252
21:47:57 <sc68cal> https://review.openstack.org/#/c/56381/
21:47:57 <mestery> sc68cal: No worries :)
21:48:00 <markmcclain> yeah we're close on the hairpin review
21:48:25 <sc68cal> yes- we've been here before though, so I will have toes and fingers crossed for the verify step
21:48:35 <sc68cal> find a couple rabbits feet too
21:48:55 <sc68cal> otherwise I have a new patch that addresses the suggestions that salv-orlando made -  https://review.openstack.org/#/c/52983/ - I will probably post tomorrow
21:49:24 <sc68cal> again - thanks to salv-orlando for looking closely at the attributes - catching the allow_put = false
21:49:43 <sc68cal> and the alembic rebaes.
21:49:47 <sc68cal> *rebase
21:49:58 <sc68cal> that's it - I yield the rest of my time
21:50:07 <markmcclain> yeah.. alembic rebases are not so fun this time of year
21:50:41 <markmcclain> I'll do the same thing I did during Havana and make sure they get sequenced correctly after approval
21:50:47 <markmcclain> sc68cal: thanks for the udpate
21:50:51 <markmcclain> #topic L3
21:50:56 <carl_baldwin> markmcclain: hi
21:51:00 <markmcclain> hi
21:51:06 <carl_baldwin> First meeting this Thursday at 1500.
21:51:13 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
21:51:23 <markmcclain> great
21:51:26 <carl_baldwin> DVR:  Documentation on agent design for changes to L2 (ML2) and L3 is forthcoming
21:51:32 <carl_baldwin> Watch for that from Swami
21:51:40 <carl_baldwin> HA Routers: patches in review
21:51:48 <carl_baldwin> #link https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/l3-high-availability,n,z
21:51:58 <Swami> Will send an update to ML when the docs are ready
21:52:09 <carl_baldwin> markmcclain: That is all I have.
21:52:24 <markmcclain> Swami: will be on the lookout for them
21:52:28 <markmcclain> carl_baldwin: thanks for the update
21:52:31 <markmcclain> #topic ML2
21:52:38 <markmcclain> mestery and rkukura: hi
21:52:42 <mestery> hi!
21:52:46 <Swami> markmcclain: Sure will be there
21:53:02 <rkukura> hi (I'll let mestery go 1st)
21:53:05 <mestery> So, a high priority item is the DB migration work addressed here: https://blueprints.launchpad.net/neutron/+spec/ml2-deprecated-plugin-migration
21:53:19 <markmcclain> marun is active on that one
21:53:22 <mestery> marun and rkukura started looking into this today and marun is going to drive this one
21:53:22 <mestery> Yes
21:53:37 <mestery> That's our highest priority item given it lets us deprecate LB and OVS plugins
21:53:50 <mestery> Outside of that, 2 of our SR-IOV patches merged, we have one left: https://blueprints.launchpad.net/neutron/+spec/ml2-binding-profile
21:54:13 <mestery> We have two port binding bugs being worked as well: https://bugs.launchpad.net/neutron/+bug/1276395 https://bugs.launchpad.net/neutron/+bug/1276391
21:54:30 <mestery> rkukura has proposed a detailed summary of the issues which those address here: http://lists.openstack.org/pipermail/openstack-dev/2014-February/026344.html
21:54:38 <markmcclain> are they bugs that would block I-3 or something that would block the final release?
21:54:44 <mestery> Other than that, we're focused on getting new MDs upstream as well.
21:55:04 <mestery> markmcclain: I'm not sure, let me ponder that.
21:55:12 <mestery> That's all for ML2 given the time left here.
21:55:28 <rkukura> markmcclain: They are needed by many of the existing and new MDs
21:55:29 <markmcclain> ok we can figure out that out over the next day or so
21:56:03 <markmcclain> rkukura: good to know
21:56:06 <mestery> thanks rkukura
21:56:16 <rkukura> hope to have a patch for 1276395 toay or tomorrow, and the other soon after
21:56:29 <markmcclain> perfect
21:56:32 <markmcclain> mestery, rkukura: thanks for the update
21:56:40 <markmcclain> #topic SSL related blueprints
21:56:53 <markmcclain> So this is the special topic that was added to agenda
21:56:53 <nati_ueno> Including VPN?
21:56:59 <markmcclain> nati_ueno: yes
21:57:10 <salv-orlando> woo the bonus track!
21:57:19 <markmcclain> in three minutes or less :)
21:57:35 <nati_ueno> Can we do discussion in the #openstack-neutron?
21:57:48 <markmcclain> In general I've got significant security concerns on the current SSL work
21:58:06 <markmcclain> sensitive pieces of data are either transmitted or stored in the clear
21:58:17 <enikanorov_> same as for lbaas i guess
21:58:18 <markmcclain> nati_ueno: I think the ML is better forum
21:58:27 <nati_ueno> markmcclain: gotcha
21:58:39 <markmcclain> So I wanted to give a heads up that I'll be starting a thread about delaying both to Juno
21:58:50 <markmcclain> both teams have done a lot of work on design and code
21:59:10 <markmcclain> but I'd rather delay and ensure we're secure over releasing something even marked experimental
22:00:14 <markmcclain> enikanorov_: yes this applies to both subteams
22:00:32 <nati_ueno> Let's discuss on ML about this. If current design is issue, let's delay it to the Juno
22:01:11 <enikanorov_> nati_ueno: that seems to be the issue with the lack of security store and secure messaging
22:01:22 <nati_ueno> Same discussion will be applied to the IPsec because we are saving password on the db
22:01:29 <nati_ueno> I thought db is a kind of secure storeage
22:01:41 <nati_ueno> but if this assumption is wrong, we should if the ipsec too
22:01:54 <markmcclain> nati_ueno: that's part of the discussion
22:02:02 <nati_ueno> markmcclain: gotcha
22:02:16 <nati_ueno> I wrote discussion paper https://docs.google.com/presentation/d/1HVA6XJss2tiHu0Z1v9IpehLsTg6VdbAKUTNJCHFeAz8/edit#slide=id.p
22:02:22 <nati_ueno> I'll send this for the openstack-dev
22:02:26 <nati_ueno> so let's discussion on the thread
22:03:01 <markmcclain> yeah.. I'll kick if off with a big explanation on what I'm thinking and then we can go from there
22:03:11 <markmcclain> We've hit the time limit for this week
22:03:33 <markmcclain> Thanks to everyone for stopping in.. talk to everyone on IRC or the mailing list
22:03:33 <salv-orlando> I ok for the sake of the meeting… any decision related to SSL blueprint is then deferred to the Mailing list?
22:03:51 <markmcclain> salv-orlando: yes
22:04:08 <markmcclain> #endmeeting