| openstackgerrit | Han Zhou proposed openstack/networking-ovn: WIP: Fix address set adding & removing race condition https://review.openstack.org/362494 | 00:04 |
|---|---|---|
| *** chandrav has quit IRC | 00:26 | |
| *** chandrav has joined #openstack-neutron-ovn | 00:27 | |
| *** yamamoto has quit IRC | 01:15 | |
| *** yamamoto has joined #openstack-neutron-ovn | 01:21 | |
| *** gongysh has joined #openstack-neutron-ovn | 01:23 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 01:32 | |
| *** banix has joined #openstack-neutron-ovn | 01:34 | |
| *** salv-orlando has quit IRC | 01:43 | |
| *** s3wong has quit IRC | 01:52 | |
| *** yamamoto has quit IRC | 02:05 | |
| *** gongysh has quit IRC | 02:11 | |
| flaviof | hi folks. I got a silly question. first, looking at the options column of my logical router table, I see th tthe chassis prop is getting set: https://gist.github.com/9254dd639c244e539132a23074118aff | 02:18 |
| flaviof | trouble is, I do not think that router is supposed to be a gateway router, as I don't even have br-provider in my setup. all I'm configuring is a plain old rtr: https://gist.github.com/e3080eb529675aba7c1a89f79dace060 | 02:20 |
| flaviof | thoughts? is this something I'm doing wrong, or a known bug? | 02:20 |
| *** mlavalle has joined #openstack-neutron-ovn | 02:43 | |
| *** armax has quit IRC | 02:46 | |
| *** banix has quit IRC | 02:46 | |
| *** banix has joined #openstack-neutron-ovn | 02:48 | |
| *** banix has quit IRC | 02:55 | |
| *** jckasper has quit IRC | 03:00 | |
| *** jckasper has joined #openstack-neutron-ovn | 03:00 | |
| *** jckasper has quit IRC | 03:04 | |
| *** yamamoto has joined #openstack-neutron-ovn | 03:05 | |
| *** yamamoto has quit IRC | 03:10 | |
| *** mickeys has quit IRC | 03:28 | |
| *** yamamoto has joined #openstack-neutron-ovn | 03:39 | |
| *** yamamoto has quit IRC | 03:45 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 03:45 | |
| *** salv-orlando has quit IRC | 03:53 | |
| *** armax has joined #openstack-neutron-ovn | 03:55 | |
| *** numans has joined #openstack-neutron-ovn | 04:03 | |
| *** numans has quit IRC | 04:04 | |
| *** numans has joined #openstack-neutron-ovn | 04:04 | |
| *** yamamoto has joined #openstack-neutron-ovn | 04:04 | |
| *** yamamoto has quit IRC | 04:18 | |
| *** mlavalle has quit IRC | 04:23 | |
| *** yamamoto has joined #openstack-neutron-ovn | 04:40 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 04:41 | |
| *** yamamoto has quit IRC | 05:00 | |
| openstackgerrit | ZongKai LI proposed openstack/networking-ovn: Add DB sync support for lrouter port networks https://review.openstack.org/358501 | 05:03 |
| *** armax has quit IRC | 05:22 | |
| *** yamamoto has joined #openstack-neutron-ovn | 05:41 | |
| *** armax has joined #openstack-neutron-ovn | 05:46 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 06:08 | |
| *** salv-orlando has quit IRC | 06:11 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 06:12 | |
| *** salv-orl_ has quit IRC | 06:17 | |
| *** salv-orlando has quit IRC | 06:17 | |
| *** pcaruana has joined #openstack-neutron-ovn | 06:26 | |
| *** armax has quit IRC | 06:28 | |
| *** cryptarium has joined #openstack-neutron-ovn | 06:29 | |
| *** jckasper has joined #openstack-neutron-ovn | 07:02 | |
| *** jckasper has quit IRC | 07:06 | |
| *** mmirecki has joined #openstack-neutron-ovn | 07:21 | |
| *** mmirecki has quit IRC | 07:30 | |
| *** chandrav has quit IRC | 07:31 | |
| *** fzdarsky|afk has joined #openstack-neutron-ovn | 07:41 | |
| *** fzdarsky|afk is now known as fzdarsky | 07:44 | |
| *** mickeys has joined #openstack-neutron-ovn | 07:59 | |
| *** mickeys has quit IRC | 08:04 | |
| *** roeyc has joined #openstack-neutron-ovn | 08:28 | |
| *** cryptarium has quit IRC | 08:36 | |
| openstackgerrit | ZongKai LI proposed openstack/networking-ovn: Skip get dhcpv4 options for ports owned by neutron https://review.openstack.org/362726 | 09:27 |
| *** yamamoto has quit IRC | 09:48 | |
| *** roeyc has quit IRC | 09:55 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 10:27 | |
| *** roeyc has joined #openstack-neutron-ovn | 10:35 | |
| *** yamamoto has joined #openstack-neutron-ovn | 10:42 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 10:52 | |
| *** salv-orl_ has quit IRC | 10:54 | |
| *** salv-orlando has quit IRC | 10:56 | |
| *** dkehn_ has quit IRC | 10:57 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 10:59 | |
| *** yamamoto has quit IRC | 11:02 | |
| *** fzdarsky is now known as fzdarsky|lunch | 11:03 | |
| *** jckasper has joined #openstack-neutron-ovn | 11:03 | |
| *** jckasper has quit IRC | 11:08 | |
| *** yamamoto_ has quit IRC | 11:08 | |
| *** dkehn_ has joined #openstack-neutron-ovn | 11:10 | |
| *** rtheis has joined #openstack-neutron-ovn | 11:12 | |
| *** yamamoto has joined #openstack-neutron-ovn | 11:16 | |
| *** yamamoto has quit IRC | 11:17 | |
| *** banix has joined #openstack-neutron-ovn | 11:26 | |
| *** yamamoto has joined #openstack-neutron-ovn | 11:29 | |
| *** banix has quit IRC | 11:29 | |
| *** banix has joined #openstack-neutron-ovn | 11:30 | |
| *** yamamoto has quit IRC | 11:33 | |
| numans | rtheis, is it better to wait for the proper fix in neutron ? or take this temporary solution https://review.openstack.org/#/c/362699/1 ? | 11:38 |
| rtheis | numans: unless we have a critical block, I think we are okay to wait for the proper fix. | 11:38 |
| *** banix has quit IRC | 11:39 | |
| numans | rtheis, even i think its better to wait. | 11:39 |
| rtheis | sounds good | 11:39 |
| *** banix has joined #openstack-neutron-ovn | 12:17 | |
| *** yamamoto has joined #openstack-neutron-ovn | 12:22 | |
| *** banix has quit IRC | 12:27 | |
| *** fzdarsky|lunch is now known as fzdarsky|afk | 12:48 | |
| *** fzdarsky|afk is now known as fzdarsky | 12:52 | |
| *** roeyc has quit IRC | 12:54 | |
| rtheis | russellb, mestery: networking-ovn Newton-3 deliverable: https://review.openstack.org/#/c/362847/ | 12:54 |
| rtheis | is anything else required? | 12:55 |
| *** roeyc has joined #openstack-neutron-ovn | 12:55 | |
| *** jckasper has joined #openstack-neutron-ovn | 13:09 | |
| *** mlavalle has joined #openstack-neutron-ovn | 13:18 | |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Fail deployment if q-dhcp and OVN_NATIVE_DHCP enabled https://review.openstack.org/361096 | 13:22 |
| mestery | rtheis: Ack | 13:36 |
| mestery | looking | 13:36 |
| *** woodburn has left #openstack-neutron-ovn | 13:46 | |
| numans | rtheis, mestery, it is good if this goes through - https://review.openstack.org/#/c/352209/ | 13:46 |
| rtheis | numans: that would be good to get in, I can hold the deliverable | 13:47 |
| numans | rtheis, thanks. | 13:55 |
| *** woodburn has joined #openstack-neutron-ovn | 14:11 | |
| *** roeyc has quit IRC | 14:15 | |
| *** roeyc has joined #openstack-neutron-ovn | 14:22 | |
| *** banix has joined #openstack-neutron-ovn | 14:23 | |
| *** roeyc has quit IRC | 14:52 | |
| openstackgerrit | ZongKai LI proposed openstack/networking-ovn: Skip get dhcpv4 options for ports owned by neutron https://review.openstack.org/362726 | 14:54 |
| *** armax has joined #openstack-neutron-ovn | 15:09 | |
| *** numans has quit IRC | 15:12 | |
| *** brad_behle has joined #openstack-neutron-ovn | 15:16 | |
| *** chandrav has joined #openstack-neutron-ovn | 15:24 | |
| *** dkehn_ has quit IRC | 15:26 | |
| *** banix has quit IRC | 15:31 | |
| *** banix has joined #openstack-neutron-ovn | 15:31 | |
| *** mickeys has joined #openstack-neutron-ovn | 15:32 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 15:40 | |
| *** dkehn_ has joined #openstack-neutron-ovn | 15:46 | |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Deployment fixes for DHCP and metadata support https://review.openstack.org/350237 | 15:48 |
| *** banix has quit IRC | 15:50 | |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Deployment fixes for DHCP and metadata support https://review.openstack.org/350237 | 15:52 |
| *** s3wong has joined #openstack-neutron-ovn | 16:02 | |
| *** yamamoto has quit IRC | 16:15 | |
| *** armax has quit IRC | 16:16 | |
| *** banix has joined #openstack-neutron-ovn | 16:19 | |
| Sam-I-Am | rtheis: mind if i poke at 350237? | 16:28 |
| rtheis | Sam-I-Am: go ahead | 16:30 |
| Sam-I-Am | rtheis: thx | 16:40 |
| Sam-I-Am | some wording for the most part | 16:40 |
| Sam-I-Am | rtheis: you around? | 16:56 |
| rtheis | yes | 16:56 |
| Sam-I-Am | got a logic question | 16:56 |
| Sam-I-Am | devstack/lib/networking-ovn L200-ish | 16:57 |
| Sam-I-Am | trying to figure out why we only enable isolated md when we're using the conventional l3 agent | 16:58 |
| Sam-I-Am | or so it looks | 16:58 |
| Sam-I-Am | isolated md basically enables md for non-routed (read: provider) networks | 16:59 |
| rtheis | force_metadata should be a superset of enable_isolated_metadata | 16:59 |
| Sam-I-Am | i've never had to set force_metadata before | 16:59 |
| rtheis | so with no conventional L3, we'll have to have DHCP do the work | 16:59 |
| Sam-I-Am | interestingly, if you enable isolated md, the l3 agent never provides md | 17:00 |
| rtheis | That may be due to devstack forcing config drive by default up until late last week | 17:00 |
| *** s3wong has quit IRC | 17:01 | |
| Sam-I-Am | thus, in any env with instances that boot on a provider net, we need allow isolated md | 17:01 |
| Sam-I-Am | rtheis: moo? | 17:05 |
| rtheis | Sam-I-Am: still here... just looking over the neutron code | 17:05 |
| rtheis | I think force_metadata should give you want you want with enable_isolated_metadata | 17:06 |
| Sam-I-Am | tl;dr - if you create a network and enable dhcp, it always has a dhcp instance running before you plug it into a router. if you enable isolated md, the dhcp agent handles md. thus, the dhcp agent will end up providing md to all networks because it comes before a router port. | 17:06 |
| rtheis | but if a router is added later then I think you lose host routes | 17:07 |
| Sam-I-Am | not that i've seen | 17:07 |
| Sam-I-Am | unless something changed, the route to 169.254.169.254 always points to the dnsmasq ip addr for that network | 17:08 |
| Sam-I-Am | because the instance hears from dhcp first, and dhcp hands that route out... i think | 17:08 |
| rtheis | I was seeing it point to router gateway port | 17:08 |
| Sam-I-Am | if l3 is handling md, there's no special route. it just intercepts that ip address in iptables and mangles it to the md service. | 17:10 |
| Sam-I-Am | guess we could test to see if any of this changed since i last stumbled across it | 17:11 |
| Sam-I-Am | anyway, i was just seeing if we could simplify the logic | 17:11 |
| rtheis | It is a bit of a mixed case ... where dhcp handles the host route but points to gateway port if not isolated | 17:12 |
| rtheis | I'll work through the tests again and let you know | 17:15 |
| *** yamamoto has joined #openstack-neutron-ovn | 17:16 | |
| Sam-I-Am | ok. i'll just clarify the comments for now. | 17:16 |
| rtheis | ok | 17:16 |
| *** yamamoto has quit IRC | 17:22 | |
| *** pcaruana has quit IRC | 17:25 | |
| *** s3wong has joined #openstack-neutron-ovn | 17:32 | |
| *** chandrav has quit IRC | 17:34 | |
| *** mlavalle has quit IRC | 17:53 | |
| *** chandrav has joined #openstack-neutron-ovn | 17:57 | |
| Sam-I-Am | rtheis: is there a reason the list of extensions in the relnote are not using a list? | 18:01 |
| rtheis | no good reason | 18:02 |
| Sam-I-Am | hmmm ok | 18:02 |
| rtheis | they really should, but I probably messed that up | 18:02 |
| Sam-I-Am | i'll do it | 18:03 |
| *** banix has quit IRC | 18:05 | |
| *** banix has joined #openstack-neutron-ovn | 18:06 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 18:08 | |
| *** salv-orlando has quit IRC | 18:11 | |
| rtheis | thx | 18:15 |
| openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: Deployment fixes for DHCP and metadata support https://review.openstack.org/350237 | 18:21 |
| Sam-I-Am | rtheis: think i fixerated it | 18:22 |
| *** tongli has joined #openstack-neutron-ovn | 18:30 | |
| rtheis | thx | 18:32 |
| *** mickeys has quit IRC | 18:32 | |
| *** salv-orl_ has quit IRC | 18:49 | |
| rtheis | russellb: are you around ? | 18:51 |
| *** basilAB has quit IRC | 18:52 | |
| *** mestery has quit IRC | 18:52 | |
| *** banix has quit IRC | 18:55 | |
| *** mestery has joined #openstack-neutron-ovn | 18:57 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 18:59 | |
| *** banix has joined #openstack-neutron-ovn | 19:00 | |
| *** basilAB has joined #openstack-neutron-ovn | 19:01 | |
| rtheis | Sam-I-Am: the reason that I needed to use force_metadata is because once a router interface is added then the host route uses the gateway IP instead of the DHCP IP. However, that won't route with native L3 router. | 19:11 |
| Sam-I-Am | rtheis: what happens without force_metadata... it'll try to rely on bits in the l3 agent? | 19:12 |
| rtheis | yes | 19:12 |
| rtheis | that is what I was seeing | 19:13 |
| Sam-I-Am | hmm, interesting. so the way it used to work where enabling isolated metadata always overrides magic in l3? | 19:13 |
| Sam-I-Am | er, no longer is the case? | 19:13 |
| *** armax has joined #openstack-neutron-ovn | 19:15 | |
| rtheis | must not be ... I'm going to run another test to verify. Need to setup another deployment ... | 19:16 |
| *** armax has quit IRC | 19:20 | |
| *** mickeys has joined #openstack-neutron-ovn | 19:33 | |
| *** rtheis_ has joined #openstack-neutron-ovn | 19:33 | |
| *** armax has joined #openstack-neutron-ovn | 19:35 | |
| *** rtheis has quit IRC | 19:35 | |
| *** mickeys has quit IRC | 19:39 | |
| *** zkassab has joined #openstack-neutron-ovn | 19:42 | |
| *** zkassab has quit IRC | 19:44 | |
| *** fzdarsky is now known as fzdarsky|afk | 19:52 | |
| *** zkassab has joined #openstack-neutron-ovn | 20:20 | |
| *** zkassab has quit IRC | 20:20 | |
| rtheis_ | Sam-I-Am: Enabling isolated metadata did override the magic in l3 until I restarted the DHCP agent and it saw the router port. Then it set a host route for that port and deferred to l3 magic. | 20:25 |
| *** rtheis_ is now known as rtheis | 20:26 | |
| Sam-I-Am | rtheis: thats... interesting. | 20:26 |
| rtheis | yes, and a bit confusing | 20:26 |
| Sam-I-Am | i dont like indeterminate things :/ | 20:26 |
| rtheis | me either | 20:26 |
| Sam-I-Am | wonder if its worth trying to fix, if thats even possible | 20:27 |
| Sam-I-Am | what does force_metadata do in that case? | 20:27 |
| Sam-I-Am | or whatever that was | 20:27 |
| rtheis | force_metadata will ensure that DHCP is always used | 20:28 |
| Sam-I-Am | even when a router is present? | 20:28 |
| rtheis | yes | 20:28 |
| Sam-I-Am | ok, so that helps a bit | 20:28 |
| Sam-I-Am | i suspect the help strings on that might need some improvement | 20:28 |
| rtheis | probably | 20:29 |
| Sam-I-Am | isolated metadata isnt exactly easy to understand either | 20:29 |
| rtheis | yeah | 20:29 |
| Sam-I-Am | so i guess force_metadata should always be used to make sure md proxies a consistent way? | 20:30 |
| Sam-I-Am | iff isolated metadata | 20:30 |
| rtheis | yes, we could do that | 20:31 |
| rtheis | I'll update the patch | 20:32 |
| *** banix has quit IRC | 20:33 | |
| *** mickeys has joined #openstack-neutron-ovn | 20:36 | |
| *** mickeys has quit IRC | 20:41 | |
| *** mickeys has joined #openstack-neutron-ovn | 20:46 | |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Deployment fixes for DHCP and metadata support https://review.openstack.org/350237 | 20:50 |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Add OVS transaction mutate support https://review.openstack.org/351861 | 21:02 |
| *** azbiswas has joined #openstack-neutron-ovn | 21:06 | |
| openstackgerrit | Richard Theis proposed openstack/networking-ovn: Deployment fixes for DHCP and metadata support https://review.openstack.org/350237 | 21:16 |
| *** rtheis has quit IRC | 21:20 | |
| openstackgerrit | Merged openstack/networking-ovn: gaterc: Avoid running entire tempest test suite on grenade job https://review.openstack.org/359086 | 21:21 |
| openstackgerrit | Merged openstack/networking-ovn: Cleanup gate stack user configuration for functional tests https://review.openstack.org/360492 | 21:24 |
| *** salv-orlando has quit IRC | 21:26 | |
| *** armax has quit IRC | 21:32 | |
| *** chandrav has quit IRC | 21:46 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 21:49 | |
| *** banix has joined #openstack-neutron-ovn | 21:53 | |
| *** mlavalle has joined #openstack-neutron-ovn | 21:55 | |
| *** chandrav has joined #openstack-neutron-ovn | 22:08 | |
| *** shettyg1 has quit IRC | 22:12 | |
| openstackgerrit | Amitabha Biswas proposed openstack/networking-ovn: BugFix: L3 scheduler scheduling distributed routers https://review.openstack.org/363305 | 22:50 |
| *** azbiswas has quit IRC | 22:51 | |
| *** armax has joined #openstack-neutron-ovn | 23:00 | |
| *** salv-orlando has quit IRC | 23:21 | |
| *** banix has quit IRC | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!