*** yamamoto_ has joined #openstack-neutron-ovn | 00:02 | |
openstackgerrit | Ramu Ramamurthy proposed openstack/networking-ovn: Add a dhcp acl to allow dhcp-client to dhcp-server https://review.openstack.org/273272 | 00:03 |
---|---|---|
*** dlundquist has quit IRC | 00:04 | |
*** dlundquist has joined #openstack-neutron-ovn | 00:05 | |
*** dlundquist has quit IRC | 00:05 | |
*** dlundquist has joined #openstack-neutron-ovn | 00:07 | |
*** manand has quit IRC | 00:17 | |
*** dlundquist has quit IRC | 00:17 | |
*** gangil1 has quit IRC | 00:21 | |
*** yamamoto_ has quit IRC | 00:27 | |
*** chenli has joined #openstack-neutron-ovn | 00:39 | |
*** azbiswas has quit IRC | 00:45 | |
openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: Vagrant: Fix various networking bits https://review.openstack.org/273162 | 01:01 |
*** Uday has joined #openstack-neutron-ovn | 01:06 | |
Uday | can someone let me know how to enable logging of OVN code? from devstack ovs-appctl vlog/list I do not see any of the ovn modules | 01:08 |
*** chandrav has joined #openstack-neutron-ovn | 01:21 | |
openstackgerrit | Ramu Ramamurthy proposed openstack/networking-ovn: Add a dhcp acl to allow dhcp-client to dhcp-server https://review.openstack.org/273272 | 01:23 |
*** yamamoto_ has joined #openstack-neutron-ovn | 01:23 | |
*** jckasper has quit IRC | 01:25 | |
*** jckasper has joined #openstack-neutron-ovn | 01:25 | |
*** jckasper has quit IRC | 01:30 | |
*** yamamoto_ has quit IRC | 01:37 | |
*** zhouhan has quit IRC | 01:46 | |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: tempest debugging, ignore .... https://review.openstack.org/273797 | 01:56 |
*** chandrav has quit IRC | 01:57 | |
russellb | Uday: you need to target the ovn daemon you're interested in ... ovn-appctl -t ovn-controller vlog/list | 01:58 |
russellb | err, ovs-appctl i mean | 01:58 |
*** gangil has joined #openstack-neutron-ovn | 02:12 | |
*** gangil has joined #openstack-neutron-ovn | 02:12 | |
*** yamamoto_ has joined #openstack-neutron-ovn | 02:23 | |
*** gangil has quit IRC | 02:28 | |
*** chandrav has joined #openstack-neutron-ovn | 02:38 | |
*** chandrav has quit IRC | 02:44 | |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: tempest debugging, ignore .... https://review.openstack.org/273797 | 02:48 |
*** azbiswas has joined #openstack-neutron-ovn | 02:51 | |
*** azbiswas has quit IRC | 03:13 | |
*** gangil has joined #openstack-neutron-ovn | 03:20 | |
*** gangil has joined #openstack-neutron-ovn | 03:20 | |
Sam-I-Am | russellb: should anything useful from ovn appear in neutron agent-list? | 03:21 |
russellb | Sam-I-Am: depends on the setup | 03:22 |
russellb | nothing from ovn | 03:22 |
russellb | just whatever neutron agents are in use | 03:22 |
Sam-I-Am | ok, i see the conventional agents in there | 03:22 |
russellb | yep | 03:22 |
Sam-I-Am | didnt know if there was sort of an equivalent command we can tell people to run to show that bits are up | 03:22 |
russellb | ovn-sbctl show | 03:23 |
russellb | will list all the hypervisors from OVN's perspective | 03:23 |
Sam-I-Am | yeah, i guess that works | 03:23 |
Sam-I-Am | what happens if a process dies? | 03:23 |
russellb | nothing | 03:24 |
russellb | OVN does nothing rather | 03:24 |
Sam-I-Am | so nothing is tracking whether things are checking in to their respective components? | 03:25 |
russellb | you could have systemd configured to restart ovn-controller, or some other monitoring thing yell at you | 03:25 |
russellb | correct, OVN is not doing anything like that right now | 03:25 |
Sam-I-Am | are there plans for it? | 03:25 |
russellb | it has come up in discussion, nobody working on anything though | 03:25 |
Sam-I-Am | or for that matter, any way to monitor traffic | 03:26 |
russellb | nothing new to ovn | 03:26 |
russellb | yet | 03:26 |
Sam-I-Am | hmm, did your localnet patches merge recently? | 03:29 |
*** manand has joined #openstack-neutron-ovn | 03:29 | |
russellb | Sam-I-Am: nope | 03:30 |
Sam-I-Am | interesting. i can access my vm on the provider net, and there's also a dhcp server there. | 03:31 |
russellb | all on the same host? | 03:32 |
russellb | and are you using my localnet branch by chance? | 03:32 |
Sam-I-Am | well, sort of. accessing the instance from the host running the vagrant vms | 03:33 |
russellb | cool | 03:33 |
russellb | dhcp agent and the instance both on the same vagrant VM though? | 03:33 |
russellb | that would be broken before my patches, fixed after | 03:33 |
Sam-I-Am | let me check that | 03:34 |
Sam-I-Am | doh, nope... ended up on compute1 | 03:35 |
Sam-I-Am | thought it would end up on the controller, which has n-cpu on it (for some reason) | 03:35 |
russellb | 2 successful test runs in a row on https://review.openstack.org/#/c/273797/ | 03:37 |
* russellb tries not to get his hopes up .... | 03:37 | |
russellb | Sam-I-Am: have a good night! i'm headed to bed | 03:38 |
Sam-I-Am | woo | 03:38 |
Sam-I-Am | i'm going to go play some gee-tar | 03:38 |
Sam-I-Am | go jam out to some country western tunes | 03:38 |
Sam-I-Am | see ya tomorrow | 03:38 |
Sam-I-Am | instance now has internet access too | 03:39 |
Sam-I-Am | w00t | 03:39 |
*** allan_h has quit IRC | 03:56 | |
*** chandrav has joined #openstack-neutron-ovn | 04:01 | |
*** salv-orlando has joined #openstack-neutron-ovn | 04:07 | |
*** salv-orlando has quit IRC | 04:07 | |
*** salv-orl_ has quit IRC | 04:09 | |
*** Uday has quit IRC | 04:13 | |
*** jckasper has joined #openstack-neutron-ovn | 04:41 | |
*** numans has quit IRC | 04:41 | |
*** jckasper has quit IRC | 04:52 | |
*** jckasper has joined #openstack-neutron-ovn | 04:52 | |
*** salv-orlando has joined #openstack-neutron-ovn | 05:08 | |
*** salv-orlando has quit IRC | 05:13 | |
*** chandrav has quit IRC | 05:13 | |
*** dlundquist has joined #openstack-neutron-ovn | 05:19 | |
*** dlundquist has quit IRC | 05:24 | |
*** numans has joined #openstack-neutron-ovn | 05:31 | |
*** manand has quit IRC | 05:33 | |
*** gangil has quit IRC | 05:38 | |
*** gangil has joined #openstack-neutron-ovn | 05:38 | |
*** gangil has joined #openstack-neutron-ovn | 05:38 | |
*** chandrav has joined #openstack-neutron-ovn | 05:39 | |
*** gangil has quit IRC | 05:42 | |
*** azbiswas has joined #openstack-neutron-ovn | 05:49 | |
*** azbiswas_ has joined #openstack-neutron-ovn | 05:54 | |
*** azbiswas has quit IRC | 05:57 | |
*** chenli has quit IRC | 06:01 | |
*** chenli has joined #openstack-neutron-ovn | 06:02 | |
*** salv-orlando has joined #openstack-neutron-ovn | 06:34 | |
*** chandrav has quit IRC | 07:08 | |
*** salv-orlando has quit IRC | 07:26 | |
*** chandrav has joined #openstack-neutron-ovn | 07:33 | |
*** chandrav has quit IRC | 07:38 | |
*** azbiswas_ has quit IRC | 07:52 | |
*** chandrav has joined #openstack-neutron-ovn | 07:56 | |
*** chenli_ has joined #openstack-neutron-ovn | 08:31 | |
*** chenli has quit IRC | 08:32 | |
*** chenli_ is now known as chenli | 08:32 | |
*** fzdarsky has joined #openstack-neutron-ovn | 08:59 | |
*** fzdarsky has quit IRC | 09:12 | |
*** chenli has quit IRC | 09:25 | |
*** chandrav has quit IRC | 09:50 | |
*** yamamoto_ has quit IRC | 10:16 | |
*** fzdarsky has joined #openstack-neutron-ovn | 10:51 | |
*** fzdarsky has quit IRC | 10:53 | |
*** fzdarsky has joined #openstack-neutron-ovn | 10:53 | |
*** fzdarsky has quit IRC | 10:53 | |
*** fzdarsky has joined #openstack-neutron-ovn | 10:56 | |
*** salv-orlando has joined #openstack-neutron-ovn | 11:03 | |
*** salv-orlando has quit IRC | 11:04 | |
*** salv-orl_ has joined #openstack-neutron-ovn | 11:05 | |
*** salv-orl_ has quit IRC | 11:30 | |
*** salv-orlando has joined #openstack-neutron-ovn | 12:01 | |
*** salv-orlando has quit IRC | 12:06 | |
*** rtheis has joined #openstack-neutron-ovn | 12:27 | |
*** yamamoto has joined #openstack-neutron-ovn | 12:39 | |
*** yamamoto has quit IRC | 12:43 | |
*** yamamoto has joined #openstack-neutron-ovn | 12:48 | |
*** yamamoto has quit IRC | 12:59 | |
*** yamamoto has joined #openstack-neutron-ovn | 13:05 | |
*** salv-orlando has joined #openstack-neutron-ovn | 13:12 | |
*** yamamoto has quit IRC | 13:31 | |
Sam-I-Am | mornings | 13:33 |
*** yamamoto has joined #openstack-neutron-ovn | 13:36 | |
*** yamamoto has quit IRC | 13:38 | |
russellb | morning | 13:42 |
*** yamamoto has joined #openstack-neutron-ovn | 13:47 | |
*** yamamoto has quit IRC | 13:53 | |
*** yamamoto has joined #openstack-neutron-ovn | 13:55 | |
*** jckasper has quit IRC | 13:58 | |
*** jckasper has joined #openstack-neutron-ovn | 13:58 | |
*** jckasper has quit IRC | 14:02 | |
*** yamamoto has quit IRC | 14:05 | |
*** yamamoto has joined #openstack-neutron-ovn | 14:14 | |
*** yamamoto has quit IRC | 14:17 | |
*** yamamoto has joined #openstack-neutron-ovn | 14:22 | |
*** jckasper has joined #openstack-neutron-ovn | 14:29 | |
*** yamamoto has quit IRC | 14:29 | |
*** flaviof has joined #openstack-neutron-ovn | 14:34 | |
*** jckasper has quit IRC | 14:43 | |
*** jckasper has joined #openstack-neutron-ovn | 14:43 | |
*** yamamoto has joined #openstack-neutron-ovn | 14:43 | |
*** roeyc has joined #openstack-neutron-ovn | 14:48 | |
*** yamamoto has quit IRC | 14:50 | |
*** ajo has quit IRC | 14:51 | |
*** yamamoto has joined #openstack-neutron-ovn | 15:03 | |
*** yamamoto has quit IRC | 15:03 | |
*** shettyg has joined #openstack-neutron-ovn | 15:05 | |
*** yamamoto has joined #openstack-neutron-ovn | 15:08 | |
*** yamamoto has quit IRC | 15:08 | |
*** dlundquist has joined #openstack-neutron-ovn | 15:23 | |
*** jckasper has quit IRC | 15:46 | |
*** jckasper has joined #openstack-neutron-ovn | 15:47 | |
*** allan_h has joined #openstack-neutron-ovn | 15:51 | |
*** jckasper has quit IRC | 15:51 | |
*** alhu_ has joined #openstack-neutron-ovn | 15:55 | |
*** allan_h has quit IRC | 15:57 | |
*** alhu_ has quit IRC | 16:02 | |
*** allan_h has joined #openstack-neutron-ovn | 16:03 | |
openstackgerrit | Numan Siddique proposed openstack/networking-ovn: (WIP) Support port security API extension https://review.openstack.org/269219 | 16:06 |
Sam-I-Am | moo. | 16:07 |
* mestery yawns | 16:09 | |
Sam-I-Am | hows life? | 16:09 |
russellb | it's Friday! | 16:12 |
Sam-I-Am | yup. and a jelly donut found me. mmmm. | 16:12 |
Sam-I-Am | 'health torus' | 16:12 |
*** numans has quit IRC | 16:13 | |
*** jckasper has joined #openstack-neutron-ovn | 16:13 | |
* russellb imagines a walking jelly donut wandering around looking for you | 16:15 | |
russellb | kinda creepy | 16:15 |
mestery | Friday | 16:16 |
mestery | Love it | 16:16 |
mestery | I explained OVN to 40 IBM execs on my birthday. | 16:17 |
mestery | I've never done that before | 16:17 |
mestery | That was fun :) | 16:17 |
russellb | \o/ | 16:17 |
russellb | how'd that go | 16:17 |
mestery | Super well | 16:17 |
mestery | The ship has been turned | 16:17 |
mestery | If OVN doesn't work for IBM public cloud, it's all on me now :P | 16:17 |
mestery | We are full steam ahead with OVN baby! | 16:17 |
russellb | better you than me | 16:18 |
russellb | :-p | 16:18 |
mestery | lol | 16:18 |
Sam-I-Am | mestery: cool. how much alcohol was involved? :) | 16:18 |
Sam-I-Am | i mean you, not them. | 16:18 |
mestery | lol :) | 16:18 |
*** salv-orlando has quit IRC | 16:19 | |
*** roeyc has quit IRC | 16:24 | |
*** roeyc has joined #openstack-neutron-ovn | 16:25 | |
*** azbiswas has joined #openstack-neutron-ovn | 16:29 | |
flaviof | mestery game on! | 16:31 |
mestery | flaviof: :P | 16:31 |
*** azbiswas_ has joined #openstack-neutron-ovn | 16:32 | |
russellb | flaviof: ! | 16:33 |
*** rtheis_ has joined #openstack-neutron-ovn | 16:34 | |
*** azbiswas has quit IRC | 16:35 | |
*** rtheis has quit IRC | 16:36 | |
flaviof | hi fella. managed to get irc on my home laptop. I miss my 16gb power horse laptop already. | 16:36 |
russellb | heh | 16:37 |
russellb | my work laptop only has 8 GB | 16:37 |
russellb | i opted for the ultra portable option | 16:37 |
Sam-I-Am | 'only' 8 gb | 16:37 |
Sam-I-Am | remember a few years ago... | 16:37 |
russellb | :) | 16:37 |
* Sam-I-Am gets his old man hat on | 16:37 | |
Sam-I-Am | back in my day, laptops had 8 mb | 16:37 |
russellb | for real | 16:37 |
flaviof | Sam-I-Am LOL | 16:37 |
Sam-I-Am | and we loved it | 16:38 |
flaviof | well... that was all before multiple Vagrant VMs and Openstack; I'm sure | 16:38 |
* flaviof need to head out. | 16:39 | |
russellb | yeah, i just don't use vagrant ... i'm usually using cloud VMs | 16:39 |
russellb | i need to see if i can make the vagrant setup work against an openstack cloud | 16:39 |
russellb | or just build something else, based on ansible or something | 16:39 |
russellb | all of that sounds like effort | 16:39 |
flaviof | russellb all doable; easier than you think. We can use Vagrant with Ansible | 16:40 |
flaviof | gotta go... really. ttyl8r | 16:40 |
*** flaviof has quit IRC | 16:41 | |
mestery | russellb: I used Vagrant against the HP cloud while I was there, and it was AWESOME | 16:42 |
mestery | russellb: So yes, it's possible and yes you should look into it :) | 16:42 |
russellb | yeah. i should figure that out. | 16:42 |
mestery | Rigth | 16:42 |
mestery | Then we can add OpenStack provider support to our vagrant setup ;) | 16:42 |
russellb | i thought maybe if i just complained in a chat room it would get solved by someone else | 16:42 |
russellb | isn't that how the internet works? | 16:42 |
mestery | That's what I thought | 16:42 |
mestery | Certainly that's what Sam-I-Am told me | 16:42 |
Sam-I-Am | what did i say? | 16:43 |
mestery | You were explaining how the internet works | 16:43 |
mestery | Mention something in a chat room | 16:43 |
mestery | And then it happens | 16:43 |
mestery | :) | 16:43 |
Sam-I-Am | isnt it a series of tubes? | 16:44 |
russellb | yup | 16:44 |
mestery | TUBES! | 16:44 |
Sam-I-Am | and dump trucks | 16:44 |
Sam-I-Am | especially when there's mtu problems | 16:44 |
mestery | GIANT dump trucks | 16:45 |
Sam-I-Am | jumbo :) | 16:45 |
Sam-I-Am | you cant fit a jumbo dump truck into a space for a regular dump truck, which is how mtu works | 16:45 |
Sam-I-Am | maybe i should use this in the presentation... | 16:45 |
mestery | lol | 16:46 |
*** dlundquist has quit IRC | 16:48 | |
russellb | anyone know where subunut-generate comes from? | 16:54 |
russellb | devstack failing on centos 7 for me on that ... | 16:55 |
mestery | Not me I'm afraid | 16:56 |
*** roeyc has quit IRC | 16:57 | |
russellb | so many yaks | 16:58 |
mestery | lol | 16:59 |
*** dlundquist has joined #openstack-neutron-ovn | 17:01 | |
*** dlundquist has quit IRC | 17:02 | |
russellb | os-testr obviously | 17:05 |
*** jckasper has quit IRC | 17:14 | |
*** jckasper has joined #openstack-neutron-ovn | 17:17 | |
*** jckasper has quit IRC | 17:22 | |
openstackgerrit | Merged openstack/networking-ovn: Update translation setup https://review.openstack.org/273788 | 17:25 |
openstackgerrit | Merged openstack/networking-ovn: Updated from global requirements https://review.openstack.org/272803 | 17:25 |
openstackgerrit | Merged openstack/networking-ovn: Vagrant: Fix various networking bits https://review.openstack.org/273162 | 17:25 |
openstackgerrit | Aaron Rosen proposed openstack/networking-ovn: Test run-1 https://review.openstack.org/274178 | 17:47 |
openstackgerrit | Aaron Rosen proposed openstack/networking-ovn: Test run-2 https://review.openstack.org/274179 | 17:47 |
*** dlundquist has joined #openstack-neutron-ovn | 17:54 | |
*** dlundquist1 has joined #openstack-neutron-ovn | 17:56 | |
*** dlundquist has quit IRC | 17:59 | |
*** dlundquist1 has quit IRC | 18:03 | |
*** dlundquist has joined #openstack-neutron-ovn | 18:05 | |
*** gangil has joined #openstack-neutron-ovn | 18:16 | |
*** gangil has joined #openstack-neutron-ovn | 18:16 | |
*** s3wong has joined #openstack-neutron-ovn | 18:28 | |
*** azbiswas_ has quit IRC | 18:29 | |
*** salv-orlando has joined #openstack-neutron-ovn | 18:34 | |
openstackgerrit | Merged openstack/networking-ovn: Make OVS related logs multi-stack friendly https://review.openstack.org/272265 | 18:45 |
*** arosen has joined #openstack-neutron-ovn | 18:46 | |
openstackgerrit | Aaron Rosen proposed openstack/networking-ovn: Test run-3 https://review.openstack.org/274200 | 18:47 |
openstackgerrit | Aaron Rosen proposed openstack/networking-ovn: Test run-4 https://review.openstack.org/274201 | 18:47 |
arosen | there we go... Hopefully we'll get some more failures out of those too. | 18:48 |
*** jckasper has joined #openstack-neutron-ovn | 18:54 | |
russellb | arosen: still haven't had a failure in my patch to force removing dhcp ports ahead of time *shrug* | 18:56 |
russellb | ¯\_(ツ)_/¯ | 18:57 |
russellb | ¯\_(ツ)_/¯ | 18:57 |
russellb | ¯\_(ツ)_/¯ | 18:57 |
russellb | ¯\_(ツ)_/¯ | 18:57 |
russellb | oops. | 18:57 |
russellb | lol.. | 18:57 |
*** pcaruana has joined #openstack-neutron-ovn | 18:57 | |
arosen | lol | 18:57 |
*** jckasper has quit IRC | 18:58 | |
*** jckasper has joined #openstack-neutron-ovn | 19:00 | |
*** arosen1 has joined #openstack-neutron-ovn | 19:03 | |
russellb | i *may* have been looking at scripting my irc client while devstack was running earlier ... | 19:04 |
*** arosen has quit IRC | 19:05 | |
russellb | and i *may* have accidentally loaded several copies of a plugin instead of reloading it. | 19:05 |
*** arosen2 has joined #openstack-neutron-ovn | 19:05 | |
*** arosen1 has quit IRC | 19:08 | |
*** pcaruana has quit IRC | 19:24 | |
*** salv-orlando has quit IRC | 19:32 | |
*** pcaruana has joined #openstack-neutron-ovn | 19:37 | |
*** dlundquist has quit IRC | 19:40 | |
*** azbiswas has joined #openstack-neutron-ovn | 19:45 | |
*** chandrav has joined #openstack-neutron-ovn | 19:51 | |
*** manand has joined #openstack-neutron-ovn | 19:55 | |
*** pcaruana has quit IRC | 20:06 | |
*** pcaruana has joined #openstack-neutron-ovn | 20:09 | |
*** gangil has quit IRC | 20:12 | |
*** dlundquist has joined #openstack-neutron-ovn | 20:16 | |
openstackgerrit | Merged openstack/networking-ovn: Update DevStack testing documentation https://review.openstack.org/273685 | 20:16 |
*** dlundquist has quit IRC | 20:18 | |
*** pcaruana has quit IRC | 20:19 | |
*** dlundquist has joined #openstack-neutron-ovn | 20:21 | |
*** dlundquist has quit IRC | 20:22 | |
*** dlundquist has joined #openstack-neutron-ovn | 20:32 | |
*** salv-orlando has joined #openstack-neutron-ovn | 20:32 | |
*** salv-orlando has quit IRC | 20:39 | |
*** salv-orlando has joined #openstack-neutron-ovn | 20:44 | |
*** dlundquist has quit IRC | 20:45 | |
*** gangil has joined #openstack-neutron-ovn | 20:45 | |
*** gangil has joined #openstack-neutron-ovn | 20:45 | |
russellb | Sam-I-Am: I have a patch coming that i'm not sure if you'll love or hate ... | 20:56 |
russellb | I think it's neat for testing, but you may hate that it's not realistic usage :) | 20:57 |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Further reflect that OVN L3 is the default. https://review.openstack.org/274234 | 20:57 |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Use ovn-sbctl show. https://review.openstack.org/274235 | 20:57 |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Provide a method for access to private net. https://review.openstack.org/274236 | 20:57 |
russellb | ... it's that last patch in particular | 20:57 |
azbiswas | russellb: I am about to propose a design document for Floating IP support in OVN. Should I add it in networking-ovn under design/docs as a .rst? | 20:57 |
* Sam-I-Am perks up | 20:57 | |
*** arosen2 has quit IRC | 20:57 | |
russellb | azbiswas: yes, that sounds good, at least for the networking-ovn parts. There is a doc/source/design/ directory | 20:58 |
russellb | azbiswas: does it require some ovn changes too? | 20:58 |
russellb | that part should be discussed / proposed on the ovs dev list | 20:59 |
azbiswas | yes, I've put the changes in that document for tracking purposes. I will propose the changes in the ovs-dev as well | 20:59 |
Sam-I-Am | russellb: hmmmmmmmm | 20:59 |
russellb | azbiswas: sounds good, looking forward to it! | 21:00 |
Sam-I-Am | russellb: so, in conventional neutron-land, if you disable snat support in a router, you just need a route pointing back to the private network from somewhere. in the case of testing, usually the controller node on which provider net resides. | 21:00 |
Sam-I-Am | this looks like a workaround for issues with l3 in ovn rather than something permanent | 21:01 |
Sam-I-Am | right now i cant route to the private net because the router ip doesn't arp | 21:01 |
russellb | Sam-I-Am: here's a thought ... we might be able to combine these 2 things | 21:02 |
russellb | we can take your setup, where you've set up a provider net and neutron routers ... | 21:02 |
* russellb thinks | 21:02 | |
russellb | nevermind. | 21:02 |
Sam-I-Am | durrr? | 21:02 |
Sam-I-Am | the patches for vagrant deploy a what-should-work | 21:03 |
Sam-I-Am | or cause vagrant to deploy... | 21:03 |
Sam-I-Am | in preparation for the 'work' | 21:03 |
russellb | you're right that this is a workaround. | 21:03 |
russellb | i think the problem is actually even more than the arp issue | 21:03 |
russellb | that's the problem if you have one host | 21:03 |
Sam-I-Am | i have 3 hosts | 21:04 |
russellb | if you have multiple, more is missing | 21:04 |
Sam-I-Am | controller + 2 compute | 21:04 |
* russellb nods | 21:04 | |
Sam-I-Am | cant ping the router gw on the provider net | 21:04 |
russellb | right | 21:04 |
Sam-I-Am | the problem with this workaround is it really doesn't seem to accomplish much and people may assume thats the way it works and implement it everywhere... in which case you get snowflakes. | 21:05 |
russellb | another problem is that the router, which is distributed, resides on every host | 21:05 |
russellb | and that router only has 1 IP | 21:05 |
russellb | so ... | 21:05 |
Sam-I-Am | someone is going to be an early adopter, use that workaround, and somewhere down the road things break | 21:05 |
russellb | ok so let's back up. | 21:05 |
Sam-I-Am | beep beep beeeeep | 21:06 |
russellb | we have this devstack thing. with Neutron L3 we have a way to talk to the VMs we create. all is happy. | 21:06 |
russellb | we switched it to default to OVN L3, which is great in a sense, because we want to focus on OVN native stuff as much as possible, as that's where we're headed | 21:06 |
russellb | ... but it's not terribly useful if we don't tell people how to talk to their VMs | 21:06 |
russellb | that's what i want to put in the doc | 21:07 |
russellb | and isn't there right now | 21:07 |
Sam-I-Am | i can see that. its a fine line between providing people with a workaround, them understanding its a workaround, and people losing faith in your product due to workarounds. | 21:07 |
Sam-I-Am | so i suggest we decorate this addition with some strong language | 21:08 |
russellb | then maybe we shouldn't have defaulted to OVN L3 yet? | 21:08 |
Sam-I-Am | and someone tracks when we can remove it | 21:08 |
russellb | or that | 21:08 |
Sam-I-Am | i'd say no | 21:08 |
Sam-I-Am | it doesn't work | 21:08 |
russellb | say no to what | 21:08 |
Sam-I-Am | the gate job wont pass with it | 21:08 |
Sam-I-Am | i guess it depends what we want people to test | 21:08 |
Sam-I-Am | if you want people to be able to deploy a conventional arch like OVS supports and go "ooo shiny" then we keep the conventional L3 agent default | 21:09 |
russellb | depends on the person | 21:09 |
russellb | thing #1 is just wanting to see it "work" | 21:09 |
Sam-I-Am | if they're working on ovn, then you want the native l3 | 21:09 |
Sam-I-Am | we can document this in the devstack stuffs | 21:09 |
russellb | I think this doc is targeted at "ooh shiny" | 21:09 |
Sam-I-Am | however, theres a class of people who despise the neutron l3 agent and what they really want to see is ovn do dvr-ish things without being dvr | 21:10 |
Sam-I-Am | that is a HUGE selling point | 21:10 |
russellb | heh, well, this gets us closer ... | 21:10 |
russellb | maybe i should combine this with multiple networks | 21:10 |
russellb | so you can see routing working | 21:10 |
Sam-I-Am | multiple priv nets? | 21:10 |
russellb | i guess | 21:10 |
Sam-I-Am | i was debating doing that in the vagrants | 21:10 |
russellb | provider net isn't going to work | 21:11 |
Sam-I-Am | i sort of want to enable all shinies in the vagrants and then see them come to life :) | 21:11 |
Sam-I-Am | i suspect what we do in vagrant is eventually what we're going to do in the gate, only right now we cant break the gate | 21:11 |
russellb | i want someone to make the vagrants work against an openstack cloud :) | 21:12 |
Sam-I-Am | you know we could make a second gate job with native l3 | 21:12 |
Sam-I-Am | nothing hurts there | 21:12 |
russellb | true, but most tests depend on floating IPs | 21:12 |
russellb | the tests that do network connectivity checks anyway | 21:12 |
russellb | IIRC | 21:12 |
Sam-I-Am | yeah i didnt say it would work for a while :) | 21:12 |
Sam-I-Am | vagrant against openstack - you mean deploy in the cloud? | 21:13 |
russellb | yes | 21:13 |
Sam-I-Am | vs vbox | 21:13 |
russellb | yes | 21:13 |
russellb | that works in theory, i just need to get our stuff working with it | 21:13 |
Sam-I-Am | multi-node will have the usual cloud net problems | 21:14 |
russellb | what problem is that, mtu? | 21:14 |
Sam-I-Am | mtu and cloud networks have mac filtering | 21:14 |
Sam-I-Am | you have to tunnel the br-provider among hosts | 21:15 |
Sam-I-Am | or mmmaaaaybe use a tunnel-type prov net, but that might not test what we want | 21:15 |
Sam-I-Am | put a vtep on the primary node | 21:15 |
russellb | ah, right | 21:15 |
russellb | well thanks for the solid feedback, you're good at bringing things back to practical / reality :) | 21:16 |
Sam-I-Am | lol... more like battle scars :) | 21:16 |
Sam-I-Am | think about it a bit, let me know what you want to do with a) this workaround doc b) default devstack of native l3 c) additional gate job for native l3 d) vagrant openstackses | 21:17 |
russellb | that's a lot of things to think about | 21:18 |
russellb | to be honest, i've mostly left the vagrant setup up to you and kyle and just rubber stamp changes | 21:18 |
russellb | because i haven't used it yet | 21:18 |
Sam-I-Am | its becoming sentient :) | 21:18 |
russellb | seems that way | 21:18 |
Sam-I-Am | i was only trying to use it for mtu testing, but then realized that wasnt going to work | 21:18 |
Sam-I-Am | so i've taken to making it super robust... perhaps what we'd want to see in a gate job somewhere. | 21:18 |
Sam-I-Am | or some CI/CD system | 21:18 |
russellb | Sam-I-Am: i'm really not excited about switching anything back to Neutron L3 agent beyond the gate | 21:19 |
russellb | especially after you dug into how it "worked" | 21:19 |
russellb | it's not something we'll ever support in any real sense | 21:19 |
russellb | it's just a hack for testing | 21:19 |
Sam-I-Am | yeah, so thats a general problem with devstack | 21:19 |
russellb | which then leads to what you said about a native l3 gate job .... | 21:20 |
Sam-I-Am | if you set q_use_providernet_blah, you'll get the *real* way things should work | 21:20 |
Sam-I-Am | except that breaks the gate due to gnarlyness in devstack that i'm hoping neutron-lib fixes, but we might need to address it sooner in neutron-legacy | 21:21 |
*** gangil has quit IRC | 21:21 | |
Sam-I-Am | looking further down the line we probably need multinode gate jobs too | 21:21 |
Sam-I-Am | hopefully we can ditch the conventional l3 agent job by then and only run multinode (since it actually matters) | 21:22 |
russellb | ++ | 21:23 |
russellb | Sam-I-Am: do you know off hand how to link to a section in .rst ? | 21:23 |
russellb | specifically, I want to create a link to the "OVN L3" section | 21:23 |
Sam-I-Am | internal ref? | 21:24 |
Sam-I-Am | just do `thing-stuff`_ and at the target do .. _thing-stuff: | 21:25 |
Sam-I-Am | if its broken, running tox -e docs should warn you | 21:25 |
Sam-I-Am | i'm fixing a bunch of broken link crap in some specs now... which i'm curious why the specs gate job is not barking | 21:25 |
Sam-I-Am | so yeah, always more layers of the onion :) | 21:26 |
*** gangil has joined #openstack-neutron-ovn | 21:30 | |
*** gangil has joined #openstack-neutron-ovn | 21:30 | |
openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: Vagrant: Consistency with upstream docs https://review.openstack.org/274254 | 21:35 |
Sam-I-Am | let the bike shedding begin | 21:35 |
*** dlundquist has joined #openstack-neutron-ovn | 21:36 | |
Sam-I-Am | aight, i'm going to see daylight for a bit | 21:45 |
russellb | enjoy | 21:49 |
*** chandrav has quit IRC | 21:49 | |
switchcade | open... daylight? | 21:53 |
* switchcade slinks back into the corner | 21:53 | |
mestery | lol | 21:53 |
russellb | ZING | 21:53 |
*** chandrav has joined #openstack-neutron-ovn | 22:02 | |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Use ovn-sbctl show. https://review.openstack.org/274235 | 22:12 |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Further reflect that OVN L3 is the default. https://review.openstack.org/274234 | 22:12 |
openstackgerrit | Russell Bryant proposed openstack/networking-ovn: doc: Provide a method for access to private net. https://review.openstack.org/274236 | 22:12 |
openstackgerrit | Amitabha Biswas proposed openstack/networking-ovn: Floating IP support design (WIP). https://review.openstack.org/274274 | 22:17 |
*** salv-orlando has quit IRC | 22:18 | |
*** dlundquist1 has joined #openstack-neutron-ovn | 22:19 | |
openstackgerrit | Amitabha Biswas proposed openstack/networking-ovn: WIP: Floating IP support design. https://review.openstack.org/274274 | 22:20 |
*** dlundquist has quit IRC | 22:20 | |
*** s3wong_ has joined #openstack-neutron-ovn | 22:22 | |
*** azbiswas has quit IRC | 22:22 | |
*** azbiswas has joined #openstack-neutron-ovn | 22:23 | |
*** s3wong has quit IRC | 22:23 | |
*** azbiswas has quit IRC | 22:28 | |
*** arosen has joined #openstack-neutron-ovn | 22:31 | |
*** rtheis_ has quit IRC | 22:34 | |
*** s3wong_ is now known as s3wong | 22:34 | |
*** shettyg has quit IRC | 23:00 | |
*** yamamoto has joined #openstack-neutron-ovn | 23:08 | |
*** yamamoto has quit IRC | 23:12 | |
*** salv-orlando has joined #openstack-neutron-ovn | 23:18 | |
*** salv-orlando has quit IRC | 23:49 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!