[ovs-dev] [ovn-controller-vtep V4 1/6] ovn-sb: Remove the "Gateway" table from the ovn-sb schema.

Russell Bryant rbryant at redhat.com
Thu Jul 16 13:25:29 UTC 2015


On 07/16/2015 03:56 AM, Alex Wang wrote:
> In a gateway like the VTEP L2 gateway, physical vlans belonging to
> the same logical network form a "logical switch".  Each logical switch
> has a dedicated tunnel key and will keep records of all MACs learned
> from the owned vlans.  So user can just send packet to a "logical
> switch" and the gateway will figure out the output port and vlan tag
> automatically.
> 
> Therefore, it is not really necessary to keep record of the vlan map
> for each gateway physical port in the OVN_Southbound database using
> "gateway_ports".
> 
> Thusly, this commit removes the "Gateway" table from the OVN_Southbound
> database.  In the "Chassis" table, the "gateway_ports" column is replaced
> by "vtep_logical_switches" column which maps the vtep logical switch name
> in the gateway to a logical port name.
> 
> Signed-off-by: Alex Wang <alexw at nicira.com>

> +      <column name="vtep_logical_switches">
> +        Maps from the name of a vtep logical switch on the gateway to a
> +        logical port name.  The logical port name must be unique; one way
> +        to do this is by concatenating the chassis name and the logical
> +        switch name.  To include one vtep logical switch to an OVN logical
> +        network, the corresponding logical port should be created in the
> +        Logical_Port table and be attached to some Logical_Switch in the
> +        OVN_Northbound database.
>        </column>

"one way to do this ..." suggests that's there could be another way, but
the implementation specifies exactly this way.  I would change "one way
to do this is by" to "this is done by".

I also think it would be useful to include this same instruction in the
documentation for ovn-nb, since that's what a CMS developer would be
reading to figure out their integration.  I'm OK with adding this later
in another patch, though.

I still plan on proposing some changes that make it so we don't encode
special information in the port name, but this patch (and series) don't
make that any more difficult to do later.

Acked-by: Russell Bryant <rbryant at redhat.com>

-- 
Russell Bryant



More information about the dev mailing list