Open Stack

On Tue, 9 Feb 2021 at 15:35, roshan anvekar <roshananvekar at gmail.com> wrote:
>
> Is it a hard requirement for br-ex to be configured on physical interface itself?? Or is it okay to configure it on a sub-interface too??
Hi Roshan,
* if you plug an interface into a bridge (OVS or Linux), then you will
no longer be able to use the IP configuration on that interface. That
is likely why you've lost access to the gateway.
* as Sean said, if you are using VLAN provider networks then plugging
a VLAN interface into br-ex will result in double encapsulation. You
can plug the bond in directly, and still make use of a VLAN interface
on the bond for the external API.
Mark
>
> Thanks in advance.
>
> ---------- Forwarded message ---------
> From: roshan anvekar <roshananvekar at gmail.com>
> Date: Tue, Feb 9, 2021, 5:54 PM
> Subject: [stein][neutron][vlan provider network] How to configure br-ex on virtual interface
> To: OpenStack Discuss <openstack-discuss at lists.openstack.org>
>
>
> Hello,
>
> Below is my scenario.
>
> I am trying to use a single physical bonded interface ( example: bond1) both for external TLS access traffic and for provider network too ( These are 2 different vlan networks with separate vlan ids) . Bond0 is used as api_interface for management traffic. Bond2 is used for storage traffic.
>
> For this I created 2 virtual interfaces on this bond1 and used it accordingly while deploying through kolla-ansible.
>
> Before deployment the gateways for both vlan networks were accessible.
>
> Post deployment I see that qdhcp-id router is created on one of the controllers. Post creation of this DHCP agent, the gateway is inaccessible.
>
> Also the vms created are not getting IPs through DHCP agent so failing to be accessible.
>
> I had configured br-ex on virtual interface of provider vlan network and not on physical interface directly.
>
> Please let me know if I am going wrong in my network configurations.
>
> Regards,
> Roshan
>
>
>