*** Oriz has quit IRC | 02:33 | |
*** dholler has quit IRC | 02:48 | |
*** dholler has joined #openvswitch | 03:01 | |
*** rcernin has quit IRC | 03:03 | |
*** rcernin has joined #openvswitch | 03:34 | |
*** rcernin has quit IRC | 03:38 | |
*** rcernin has joined #openvswitch | 04:08 | |
*** ralonsoh has joined #openvswitch | 05:37 | |
*** slaweq has joined #openvswitch | 06:00 | |
*** _lore_ has quit IRC | 06:06 | |
*** _lore_ has joined #openvswitch | 06:11 | |
*** eelco has joined #openvswitch | 06:13 | |
*** mdgray has joined #openvswitch | 07:17 | |
*** ricolin has quit IRC | 07:18 | |
*** ricolin has joined #openvswitch | 07:21 | |
*** rcernin has quit IRC | 07:56 | |
*** istokes has joined #openvswitch | 07:57 | |
*** anilvenkata has quit IRC | 08:35 | |
*** anilvenkata has joined #openvswitch | 09:16 | |
*** anilvenkata has quit IRC | 09:17 | |
*** eelco has quit IRC | 09:24 | |
*** anilvenkata has joined #openvswitch | 09:32 | |
*** eelco has joined #openvswitch | 09:59 | |
*** deadalnix has joined #openvswitch | 10:31 | |
*** istokes1 has joined #openvswitch | 11:08 | |
*** istokes has quit IRC | 11:10 | |
deadalnix | So, I've been playing with ovn-nbctl lsp-set-addresses port dynamic | 11:43 |
---|---|---|
deadalnix | dynamic_addresses are being set properly in the nb database | 11:43 |
deadalnix | but then, they don't seem to be used anywhere | 11:44 |
deadalnix | How is that info supposed to propagate? | 11:44 |
*** istokes1 has left #openvswitch | 12:32 | |
*** elvira has quit IRC | 12:48 | |
*** elvira has joined #openvswitch | 12:50 | |
*** bostondriver has joined #openvswitch | 13:06 | |
*** bdonnahue1 has joined #openvswitch | 13:15 | |
*** deadalnix has quit IRC | 13:18 | |
*** bdonnahue1 has quit IRC | 13:32 | |
*** eduardas has joined #openvswitch | 13:50 | |
*** dcbw has joined #openvswitch | 14:00 | |
*** deadalnix has joined #openvswitch | 14:21 | |
flaviof | deadalnix: see if this page helps you https://blog.oddbit.com/post/2019-12-19-ovn-and-dhcp/ | 14:30 |
deadalnix | flaviof, I read that, but that seems to cover just half of it. What about the MAC addresses? | 14:32 |
flaviof | deadalnix: propagation of mac addresses? | 14:43 |
deadalnix | ovn-nbctl lsp-set-addresses myport dynamic | 14:44 |
deadalnix | If you run this ^ | 14:44 |
deadalnix | OVN will assign a MAC and an IP to the port (granted you defined a subnet for the switch) | 14:44 |
deadalnix | But that MAC address doesn't seems to propagate through OVN's SB or anything, so I'm wondering what's up with it. | 14:46 |
*** rcernin has joined #openvswitch | 14:48 | |
flaviof | deadalnix: i do not think there will be a lot to see in the sb before that lp is considered bound at a given chassis. Is that what you are asking? If that is so, consider creating a port in a chassis and set the external id to be what you have in the nb. An ez way for playing with that may be by using: https://github.com/ovn-org/ovn-fake-multinode but there are other great options too. | 14:49 |
deadalnix | flaviof, no, you can create a port in a chassis, bind it to the switch and all, and the mac address appears nowhere outside of the NB. | 14:50 |
flaviof | deadalnix: did you look at the openflow rules? | 14:50 |
*** rcernin has quit IRC | 14:52 | |
deadalnix | I think the flow rule are correct, but nothing works, because the interface themselves don't have a mac. Obviously, the mac can be set manually and then it works, but how's one supposed to know what mac are going to be picked by the switch? That doesn't seems like a setup that can work in practice. | 14:52 |
flaviof | deadalnix: ah... ok. normally the cms tells ovn what the mac is. it is also the entity that creates the interface in ovs. if you do not have a way to convey the 'thing' that is creating the port what mac it should use, then indeed you would need to have that info somehow. either that is simply use "unknown". | 14:57 |
flaviof | s/is/or/ | 14:57 |
*** |subz3r0| has joined #openvswitch | 15:01 | |
*** |subz3r0| has joined #openvswitch | 15:01 | |
deadalnix | flaviof, setting all the mac explicitely works, but then I'm wondering why OVN tries to assign MACs at all? | 15:03 |
deadalnix | I must be missing something, because I don't see how that feature is usable at all. | 15:04 |
flaviof | deadalnix: I'm sure there is a use case that needed that somewhere.... but I do not see that ever used in Openstack, which is mostly what I know. I know very little! ;) | 15:11 |
*** tfherbert has quit IRC | 15:30 | |
*** eelco has quit IRC | 16:02 | |
*** elvira has quit IRC | 17:04 | |
*** deadalnix has quit IRC | 17:12 | |
*** mdgray has quit IRC | 17:17 | |
*** fdangelo_ has quit IRC | 17:18 | |
*** fdangelo_ has joined #openvswitch | 17:19 | |
*** ktraynor has quit IRC | 17:34 | |
*** fdangelo_ has quit IRC | 18:04 | |
*** _lore_ has quit IRC | 18:04 | |
*** jaicaa has quit IRC | 18:04 | |
*** donhw_ has quit IRC | 18:04 | |
*** stand has quit IRC | 18:04 | |
*** fdangelo_ has joined #openvswitch | 18:07 | |
*** _lore_ has joined #openvswitch | 18:07 | |
*** jaicaa has joined #openvswitch | 18:07 | |
*** donhw_ has joined #openvswitch | 18:07 | |
*** stand has joined #openvswitch | 18:07 | |
*** anilvenkata has quit IRC | 18:34 | |
*** eduardas has quit IRC | 18:44 | |
*** rcernin has joined #openvswitch | 18:48 | |
*** rcernin has quit IRC | 18:53 | |
*** ohama has quit IRC | 19:07 | |
*** ohama has joined #openvswitch | 19:09 | |
*** slaweq has quit IRC | 19:18 | |
*** fbl has joined #openvswitch | 19:27 | |
*** fbl has quit IRC | 19:33 | |
*** dholler has quit IRC | 20:23 | |
*** dcbw has quit IRC | 20:27 | |
*** ralonsoh has quit IRC | 20:37 | |
*** ricolin has quit IRC | 20:48 | |
*** rcernin has joined #openvswitch | 20:48 | |
*** ricolin has joined #openvswitch | 20:52 | |
*** rcernin has quit IRC | 20:53 | |
*** fbl has joined #openvswitch | 21:07 | |
*** bostondriver has quit IRC | 21:22 | |
*** bostondriver has joined #openvswitch | 21:24 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!