Using OpenvSwitch, all devices with name "tapxxx" attached to OVS bridge can not be synchronized successfully when ovs_use_veth is set to True
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
neutron |
Expired
|
Undecided
|
Unassigned |
Bug Description
OpenStack will create a device attached to the OVS integration bridge as a dhcp server, so whenever you create a network in OpenStack, there will be a tap device created and attached to the OVS integration bridge such as below.
stack@vm:
67b6d3bf-
Bridge br-int
Port "tap865468b3-57"
tag: 7
Port "eth1"
Port "tapbd8e5831-c9"
ovs_version: "1.4.6"
stack@vm:
20: lo: <LOOPBACK,
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
24: ns-any: <BROADCAST,
link/ether 56:b7:ff:07:dc:85 brd ff:ff:ff:ff:ff:ff
50: tap865468b3-57: <BROADCAST,
link/ether 00:50:56:9b:95:48 brd ff:ff:ff:ff:ff:ff
inet 10.0.0.2/24 brd 10.0.0.255 scope global tap865468b3-57
inet 169.254.169.254/16 brd 169.254.255.255 scope global tap865468b3-57
inet6 fe80::250:
valid_lft forever preferred_lft forever
And when you remove the integration bridge carelessly, you can create the bridge with same name manually, and restart the dhcp agent to get all tap devices back.
But the devices can not be set correctly when ovs_use_veth is set to True. If ovs_use_veth is set to True, there will be two peer devices created, one is named ns-xxxxx, another one is named tapxxxxx. We need to make sure tapxxxxx could be attached to the ovs integration bridge automatically using the scenario above.
Changed in neutron: | |
assignee: | nobody → Yang Yu (yuyangbj) |
tags: | added: ovs |
I'm not sure 'And when you remove the integration bridge carelessly' is a valid cloud scenario