[RFE] Implement vlan transparent for openvswitch ML2 driver

[Bo Chi]

There's some progress in ovs community to add QinQ support, email discussion: [1]

This bug is to leverages the effort in the ovs community to add QinQ support in order to enable vlan transparent networks when using the openvswitch ML2 driver. This will allow VMs to send out tagged packets, rather than being dropped.

The thought is to use the vlan_transparent property of network, which is added before, but never used in openvswitch ML2 driver.

[1] https://www.mail-archive.com/search?a=1&l=dev%40openvswitch.org&haswords=%22qinq+tunneling%22&x=12&y=11&from=&subject=&datewithin=2m&date=2016-07-13&notwords=&o=newest

[Armando Migliaccio (armando-migliaccio)]

https://review.openstack.org/#/c/342755/

[Armando Migliaccio (armando-migliaccio)]

We'd need to emphasise the upstream dependencies as done similarly in [1]. I think a spec is in order to raise the visibility of this effort across the wider audience.

[1] https://review.openstack.org/#/c/344997/

[Armando Migliaccio (armando-migliaccio)]

We'll reassess this once Ocata opens up.

[Armando Migliaccio (armando-migliaccio)]

That will give us more time to understand how Open vSwitch is going to provide the required functionality.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Armando Migliaccio (<email address hidden>) on branch: master
Review: https://review.openstack.org/342755
Reason: This review is > 4 weeks without comment and currently blocked by a core reviewer with a -2. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and contacting the reviewer with the -2 on this review to ensure you address their concerns.

[Trevor McCasland (twm2016)]

Is it ok if I continue working on this by proposing a QinQ driver like Sean suggested in review?

[sean mooney (sean-k-mooney)]

ovs has now merged the support for QinQ

https://github.com/openvswitch/ovs/commit/f0fb825a3785320430686834741c718ff4f8ebf4
https://github.com/openvswitch/ovs/commit/fed8962aff57f552163ef718cc1b0db582f2295e

this will be released in ovs 2.8 in august prior to the release of pike.

[Ihar Hrachyshka (ihar-hrachyshka)]

Trevor, you can work on the patch but drivers will need to approve the RFE before we move with it. At this point, it's probably Queens material.

[Kevin Benton (kevinbenton)]

Hi,

I unassigned Trevor because what he is working on (QinQ driver) is not actually related to enabling VLAN transparency in the OVS agent. See comment on https://review.openstack.org/#/c/446047/

So this needs a contributor focused just on VLAN transparency.

[Kevin Benton (kevinbenton)]

Postponing for now until we have someone who can implement this. Considering our switches to use push/pop VLAN more, it may be as simple as detecting that we have a new enough OVS version.

[JD Lester (jdlester8)]

In my case I have a VNF (Customer edge router) that sends double tagged frames to its peering aggregation router. The current implementation drops the double tagged frame at the port the VNF is connected to on br-int because the port is configured by the ML2 OVS driver as an access port for the neutron network's local VLAN ID. I have questions as to how this case would be addressed.

Can OVS push a VLAN on a double tagged frame as a local VLAN then pop it at br-provider? AKA Is two VLANs the maximum number that can be on a frame? I attached a capture of the frame I am referring to.

Does the ML2 OVS driver require local VLAN to provider VLAN mappings for this case? Can the VNF connect directly to br-provider? AKA Is the mapping even necessary and whats the thought behind it? Can br-int be configured to pass provider VLANs to VNFs?

[Li Zhouzhou (lizhouzhou)]

I was implement vlan transparent for openvswitch Ml2 driver, which needs ovs-2.8.1.
Now, it's workly in my experiment enviroment.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/576687

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/pike
Review: https://review.openstack.org/623978

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/pike
Review: https://review.openstack.org/623980

[OpenStack Infra (hudson-openstack)]

Change abandoned by Li Zhouzhou (<email address hidden>) on branch: stable/pike
Review: https://review.openstack.org/623978
Reason: override

[Slawek Kaplonski (slaweq)]

This bug has had a related patch abandoned and has been automatically un-assigned due to inactivity. Please re-assign yourself if you are continuing work or adjust the state as appropriate if it is no longer valid.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Slawek Kaplonski (<email address hidden>) on branch: master
Review: https://review.openstack.org/576687
Reason: This review is > 4 weeks without comment, and failed Jenkins the last time it was checked. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and leaving a 'recheck' comment to get fresh test results.

[Slawek Kaplonski (slaweq)]

This bug has had a related patch abandoned and has been automatically un-assigned due to inactivity. Please re-assign yourself if you are continuing work or adjust the state as appropriate if it is no longer valid.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Slawek Kaplonski (<email address hidden>) on branch: stable/pike
Review: https://review.opendev.org/623980
Reason: This review is > 4 weeks without comment, and failed Jenkins the last time it was checked. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and leaving a 'recheck' comment to get fresh test results.

[Rodolfo Alonso (rodolfo-alonso-hernandez)]

Bug closed due to lack of activity, please feel free to reopen if needed.