[ovs-discuss] VLAN unable to ping

[Phil Daws]

----- Original Message -----
From: "Ben Pfaff" <blp at nicira.com>
To: "Phil Daws" <uxbod at splatnix.net>
Cc: discuss at openvswitch.org
Sent: Sunday, 8 December, 2013 3:57:26 PM
Subject: Re: [ovs-discuss] VLAN unable to ping

On Sun, Dec 08, 2013 at 03:55:39PM +0000, Phil Daws wrote:
> Ben, my apologies! I meant vnet1 and vnet3 which are trunked on tag id 2.
> 
> ovs-vsctl show
> 76d53621-7bf2-4af2-b24c-181a016317d7
>     Bridge "ovsbr0"
>         Port "vnet1"
>             trunks: [2]
>             Interface "vnet1"
>         Port "ovsbr0"
>             Interface "ovsbr0"
>                 type: internal
>         Port "vnet3"
>             trunks: [2]
>             Interface "vnet3"
>         Port "eth0"
>             Interface "eth0"
>         Port "vnet0"
>             trunks: [0, 1]
>             Interface "vnet0"
>         Port "vnet2"
>             trunks: [0, 1]
>             Interface "vnet2"
>     ovs_version: "1.10.2"

Did you configure the guest to use VLAN tag 2?  I didn't see any
configuration for that in your "ifconfig" output.
------------

getting a little further with this now.  The guests are both able to ping each other on their 192.168.1.X address as-well as the SAN 10.0.0.X one.  What they are unable to do is reach the outside world even though the default GW of 192.168.1.1 has been set.  The host is also unable to ping the guests 192.168.1.X addresses even though it itself has a an address within the same range.  Here is what I see:

uxbod-office sbin # ovs-vsctl show
76d53621-7bf2-4af2-b24c-181a016317d7
    Bridge "ovsbr0"
        Port "vnet3"
            trunks: [20]
            Interface "vnet3"
        Port "vnet2"
            trunks: [0, 10]
            Interface "vnet2"
        Port "ovsbr0"
            Interface "ovsbr0"
                type: internal
        Port "eth0"
            Interface "eth0"
        Port "vnet1"
            trunks: [20]
            Interface "vnet1"
        Port "vnet0"
            trunks: [0, 10]
            Interface "vnet0"
    ovs_version: "1.10.2"

vnet0 = 192.168.1.56
vnet1 = 10.0.0.10
vnet2 = 192.168.1.90
vnet3 = 10.0.0.20

the physical nic and the bridge have no trunks or tags applied so I thought that they would pass all traffic to the guests:

uxbod-office sbin # ovs-vsctl list port ovsbr0
_uuid               : 22601276-6040-4c57-a01e-b856e16e09c8
bond_downdelay      : 0
bond_fake_iface     : false
bond_mode           : []
bond_updelay        : 0
external_ids        : {}
fake_bridge         : false
interfaces          : [d38d649a-e529-4b69-81b3-e3c04de4923e]
lacp                : []
mac                 : []
name                : "ovsbr0"
other_config        : {}
qos                 : []
statistics          : {}
status              : {}
tag                 : []
trunks              : []
vlan_mode           : []
uxbod-office sbin # ovs-vsctl list port eth0
_uuid               : 39d344bb-8db5-4695-9817-28a0906ec0be
bond_downdelay      : 0
bond_fake_iface     : false
bond_mode           : []
bond_updelay        : 0
external_ids        : {}
fake_bridge         : false
interfaces          : [07ae5c32-5b0c-4b65-a6d1-8570d43d3ac6]
lacp                : []
mac                 : []
name                : "eth0"
other_config        : {}
qos                 : []
statistics          : {}
status              : {}
tag                 : []
trunks              : []
vlan_mode           : []

any thoughts please ? Thank you.