#opendaylight-group-policy: gbp_status_arch

Meeting summary

1. agenda (tbachman, 17:00:12)
   1. https://meetings.opendaylight.org/opendaylight-group-policy/2015/gbp_status_arch/opendaylight-group-policy-gbp_status_arch.2015-02-27-18.02.html Minutes from last week’s meeting (tbachman, 17:01:16)
   
   
2. Trello (tbachman, 17:03:32)
   1. alagalah, martin_sunal, and others have been working on the neutron provider (tbachman, 17:10:51)
   2. Worked out a lot of the mappings, but still some more work to be done (tbachman, 17:11:02)
   3. yapeng moved the port name from openstack-endpoint renderer to the generic endpoint; moved the URI in openstack, and changed the port name (tbachman, 17:11:53)
   4. yapeng says the code has been merged to the master branch, for kilo, and will be back-ported to the Juno branch (tbachman, 17:12:10)
   5. yapeng says both openstack-gbp and regular gpb were validated against yapeng’s work (tbachman, 17:12:32)
   6. Sanjay asks if the augmentation for the port has been added to the northbound (tbachman, 17:12:55)
   7. alagalah says he’ll address that in a minute (tbachman, 17:13:03)
   8. yapeng says from the openstack side, the change has been completed and merged — hoping to be completed/approved by the end fo the day (tbachman, 17:13:28)
   9. alagalah says that having this back-ported to Juno will allow us to remove the old openstack-gbp code from ODL (tbachman, 17:14:31)
   10. alagalah says that we’re still in the same state with port-name — we didn’t want to break what yapeng was working on; alagalah saw the patch to change the port name and will try to look at it over the weekend (tbachman, 17:15:55)
   11. alagalah has added support for EPs being in multiple EPGs, but there are lots of implications; control plane part is done, data plane still being worked on (tbachman, 17:17:17)
   12. alagalah says when an EP was only part of a single EPG, network containment was simple — defined by EPG; with multi-EPG, this is defined by the EP. (tbachman, 17:17:52)
   13. Sanjays asks about moving network location from Endpoint (tbachman, 17:29:38)
   14. alagalah says the EPG’s NetworkDomainID is more generic — the EP has an L2 and L3 context (tbachman, 17:31:14)
   15. alagalah introduces the architecture proposed for multi-EPG support (tbachman, 17:51:07)
   16. alagalah says that the EP will be able to specify the network domain, which would override the network domain in the EPG; this is important when an EP can belong to mulitple EPGs (tbachman, 17:51:40)
   17. Sanjay asks what can be done when an EP doesn’t have a network domain — doesn’t want to touch a bunch of EPs in order to update a domain (tbachman, 17:52:05)
   18. alagalah says this is being looked at — not sure of the specifics just yet (tbachman, 17:52:16)
   19. alagalah says this could be handled by a network domain EPG (tbachman, 17:52:42)

Action items

1. (none)

People present (lines said)

1. tbachman (26)
2. odl_meetbot (4)
3. alagalah (0)