15:00:32 #startmeeting neutron_l3 15:00:32 Meeting started Thu Nov 12 15:00:32 2015 UTC and is due to finish in 60 minutes. The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:33 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:35 The meeting name has been set to 'neutron_l3' 15:01:19 good mrning everybody, who is here today? 15:01:20 help 15:01:24 hi 15:01:27 hi 15:01:31 moo 15:01:34 #info 15:01:34 Hi 15:01:38 hi 15:01:40 #HELP 15:01:41 agenda for today is https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam#Agenda 15:02:01 #topic Annoucements 15:02:32 Mitaka-1 is approaching fast. It is Decmber 3rd if I remember correctly 15:02:57 Yikes 15:03:09 any other annoucements from the team? 15:03:48 ok, moving on, let's review bugs 15:03:57 #topic Bugs 15:04:31 we have a new bug https://bugs.launchpad.net/neutron/+bug/1478100 15:04:31 Launchpad bug 1478100 in neutron "DHCP agent scheduler can schedule dnsmasq to an agent without reachability to the network its supposed to serve" [High,In progress] - Assigned to Cedric Brandily (cbrandily) 15:05:17 it is assigned to ZZelle (Cedric) 15:05:36 I don't see him on-line today. No patchset has been submitted yet 15:06:06 I will monitor progress and follow up with Cedric 15:06:10 he was online earlier 15:06:13 there is a patch for that 15:06:24 https://review.openstack.org/#/c/205631/ 15:06:29 amuller: thanks! 15:06:39 I will update the bug 15:07:31 any ational comments amuller? 15:07:36 nope! 15:07:57 thanks for working on this! 15:08:15 next up is https://bugs.launchpad.net/neutron/+bug/1494351 15:08:15 Launchpad bug 1494351 in neutron "Observed StaleDataError in gate-neutron-dsvm-api tests if reference IPAM driver is used" [High,In progress] - Assigned to Pavel Bondar (pasha117) 15:08:59 pavel_bondar has been working on a fix https://review.openstack.org/#/c/223123/ 15:09:25 it is awaiting reviews from cores. It has several +1's since last week 15:09:43 so I encourage cores to take a look 15:10:04 Sorry I have not yet finished review of it. I have draft comments but need to wrap it up today. 15:10:43 thanks to haleyb. I pinged him last nigh and he reviewed it 15:11:25 pavel_bondar also closed a bug last week. So Those are all the bugs I found for today's meeting 15:11:26 yeah, i didn't +2 because some of that sql code is out of my comfort zone, need to get better at it 15:11:50 carl_baldwin: thanks for taking the time today 15:12:00 haleyb: It is a pretty heavy patch to review from that perspective. 15:12:21 any bugs that I missed from the team? 15:13:04 ok, let's move on 15:13:23 #topic Routed Network Segments 15:13:33 carl_baldwin: you are up 15:13:51 Hi 15:13:59 I've been working on some code. 15:14:19 I got the basic CRUD up for IpNetwork and I've been working on adding relationships. 15:14:32 #link https://review.openstack.org/#/c/242270/ 15:14:46 I'm almost done with an update to add those. 15:14:58 Hopefully today. I just have a few UTs to write. 15:15:17 We also need a patch started for adding host <-> Network mapping. 15:15:30 Any volunteers? 15:16:05 * tidwellr1 hears crickets 15:16:09 carl_baldwin: if a slow start is possible, I can help. I want to keep an eye on wrapping up DNS 15:16:13 I'm happy to help 15:16:17 * regXboi notes loud crickets 15:16:31 tidwellr1: mlavalle: Thanks. 15:16:54 tidwellr1: Would it be too much with your BGP work to start a host mapping patch? It needs some CRUD and DB. 15:17:12 not at all 15:17:19 tidwellr1: I just know you're pretty good at that stuff by now. :) 15:17:27 I can't crank on BGP 24/7 15:17:36 need to walk away from it at times 15:18:05 I'm happy to help on Routed Network Segments topic 15:18:09 I'm also thinking that we need to figure out if a Port can exist without a Network. Some have called this late binding. 15:18:18 tidwellr1: understood 15:19:25 rsxin: Hi, have you had a chance to review the spec and patch? 15:19:35 sure 15:19:42 I will do tomorrow 15:20:09 it's night local time now, so tomorrow is ok? 15:20:23 rsxin: Please include your IRC nick somewhere on there so that I can associate it with your comments. 15:20:28 rsxin: Tomorrow is great. 15:20:52 rsxin: And welcome, it is good to meet you. I don't think we've met before (unless I just don't recognize your nick) 15:21:03 I will put my irc name on openstack account 15:21:32 yes, I am watch neutron for long time, but not join it 15:21:42 I want to be involved 15:22:05 rsxin: can you set your Real Name in irc as well? 15:22:43 sure 15:22:46 rsxin: Great. Adding review comments to interesting patches and participating in IRC is usually the best place to start! 15:23:54 rsxin: thanks for joining the meeting and the group! 15:23:56 Well, exploring how the late binding thing could work and doing the host / network mapping are the things that I'm pushing now. 15:24:16 That is all. 15:24:33 any more questions for carl_baldwin? 15:24:55 ok, let's move on 15:25:04 Thanks! 15:25:19 #topic BGP Dynamic Routing 15:25:35 tidwellr1, vikram: you are up 15:26:18 i have one question, in bp, it mentions associating router to bgp speaker, but you have not implemented it right? 15:26:23 vikram won't be around for a couple of weeks. I wish him and his family the best. 15:26:46 yes, absolutely a rough time 15:26:55 carl_baldwin: thanks for the clarification 15:27:28 Na_: routers are not associated with bgp_speaker 15:27:47 Na_: which BP are you looking at? 15:27:52 https://review.openstack.org/#/c/125401/21/specs/liberty/bgp-dynamic-routing.rst 15:27:55 it mentions here 15:28:20 i think this is the latest one, it is for liberty 15:29:01 Na_: we've evolved from this, so it is dated 15:29:06 4. Associate a ``Router`` to a ``BGPSpeaker``. That means the router will have dynamic routing capabilities that will let it advertise its routes 15:29:24 we have a devref going here https://review.openstack.org/#/c/196452/ 15:29:45 Na_: the devref is probably more accurate 15:30:08 ok, thanks 15:30:44 right now I'm pushing some changes through the review chain that change some code structure 15:31:27 tidwellr1: Should reviewers start hitting the chain? 15:31:30 I have 1 last patch set to finish with, then I'll be moving through with address scope awareness and CLI polish, then we can really hit it with reviews 15:31:32 tidwellr1: bgp dynamic routing has many patches, some have conflicts 15:32:12 Na_: I've resolved the conflicts except with this last review 15:32:15 which patch includes all the code changes? 15:33:04 Na_: https://review.openstack.org/#/c/216994/ is the last in the review chain, but it's in conflict and needs a re-spin anyway 15:33:40 tidwellr1: thanks 15:33:48 what about the vpn plan? 15:34:00 Na_: not for Mitka 15:34:10 have you started the design? 15:34:14 Na_: nope 15:34:32 Na_: this needs to all shake out and merge first imho 15:35:09 anyway, we're not ready to hit BGP with reviews yet, but getting close 15:35:21 so have to wait for long time for vpn 15:36:03 why not add vpn in mitka? 15:36:22 Na_: there's just not time in Mitaka, if you have ideas feel free to look at this work and file an RFE 15:36:31 tidwellr1: thanks. Ping us as soon as they're ready 15:36:34 if less of resource, i think you can get volunter 15:37:12 Mickey has filed RFE, but no response yet 15:37:35 Na_: do you have a link? 15:37:57 yes, give me 1 min 15:38:06 * tidwellr1 is done after this btw 15:39:00 mlavalle: we can move on 15:39:08 tidwellr1: thanks! 15:39:09 * pavel_bondar just figured out that meeting is in progress, catching up with log 15:39:18 moving on 15:39:26 #topic DNS 15:39:35 https://bugs.launchpad.net/neutron/+bug/1509431 15:39:36 Launchpad bug 1509431 in neutron "Enhance BGP Dynamic Routing with L2VPN/L3VPN support" [Undecided,New] 15:39:44 tidwellr1:https://bugs.launchpad.net/neutron/+bug/1509431 15:40:04 thanks Na_ we'll take a look 15:40:33 I got https://review.openstack.org/#/c/212213/ to pass Jenkins tests 15:40:53 and I got several reviews. Working on it 15:41:11 * carl_baldwin adds it back to radar 15:41:32 I have also been adding the code for ports on provider networks, where we want to make the ip_address public 15:42:20 the way I am approaching it is that I am going to consider the ip address public if the following conditions are met 15:42:33 1) the netwrok has a dns_domain assigned 15:42:44 2) the port has a dns_name assigned 15:43:29 the network has a segment outside of the range assigned to tenant networks. This tells me the network was created explicitely as a provider network 15:43:38 condition 3^^^^ 15:44:40 I should push the code over the next few days 15:44:58 and that's all I have today. Thanks for the reviews 15:44:58 mlavalle: Interesting. Condition 3 is the only thing that sets it apart as "public". It sounds a little bit heuristicky but I'd have to think about it to be sure. 15:45:45 ok, let's discuss in channel 15:46:01 this is what I wanted to trigger, feedback! 15:46:23 any more questions or comments? 15:46:54 mlavalle: Sounds like progress. I'll need it to sink it a bit. 15:47:14 carl_baldwin: thanks. let's move on 15:47:33 #topic Address Scopes 15:47:55 Hi 15:48:11 I'm looking to get the server-side RPC patches merged. 15:48:48 carl-baldwin: what's left to tie up? 15:49:05 https://review.openstack.org/#/c/189948/19 15:49:18 tidwellr1: You've been very good at reviewing them. Thanks. 15:49:36 The link is where the chain starts. It is a simple linear chain. 15:49:51 hi carl 15:50:02 steve_ruan: hi 15:50:16 I have a question how address scope data path work 15:50:27 could u spare me some time after this meeting 15:51:43 steve_ruan: Actually today could be kind of difficult. I'm actually traveling in a bit with my family. How would Monday do? 15:51:54 * neiljerram is sorry to be late 15:51:55 sure 15:52:06 I will send u a email 15:52:27 steve_ruan: Thanks. 15:52:43 *has to drop off now* 15:52:54 johnbelamaric: bye 15:52:57 carl_baldwin: is this review chain complete ie it contains all the reviews needed to tie off address scopes? 15:53:04 johnbelamaric: thanks! 15:53:06 tidwellr1: if i want to advertise tenant network, the subnets of the tenant network must be in the address scope, right? 15:53:12 tidwellr1: I believe so. 15:53:43 carl_baldwin: sweet 15:53:49 tidwellr1: steve_ruan: It is the data path patch at the end that I have the most concern about. 15:54:34 in https://review.openstack.org/#/c/180267/9/specs/liberty/address-scopes.rst 15:54:54 networks in same address scope will not use nat 15:55:08 steve_ruan: That is right. 15:55:18 if networks connect to different routers, how it happen 15:56:06 for example, in ovs, all traffic goto be-ex without nat? 15:56:07 carl baldwin: the address scope is significant when multiple tenants share one address scope, right? 15:56:56 steve_ruan: This is a change only to the routers. No change to the layer 2 stuff. 15:57:17 Na_: yes that is significant. 15:57:21 NAT is done in router 15:57:40 Have this already in the patch? 15:57:56 I will check the patch 15:58:01 steve_ruan: Yes, it is in the final patch in the series. The one I really need eyeballs on. 15:58:04 steve_ruan: Thanks. 15:58:10 thanks 15:58:13 steve_ruan: I'll look for your feedback. 15:58:25 carl_baldwin: so do you allow only one tenant use the address scope, not multiple? 15:58:31 mlavalle: We're about out of time. Was there anything else? 15:58:46 Na_: There are shared address scopes. 15:58:57 carl_baldwin: nope, let's move on 15:59:26 thank you all for showing up to the meeting, we are out of time 15:59:39 *is back just in time for the end* :) 15:59:39 bye 15:59:43 bye 15:59:50 #endmeeting