Open Stack

Why do you have two physical bridged networks on your compute node?
El 14/12/2014 22:51, "Gonzalo Aguilar Delgado" <gaguilar at aguilardelgado.com>
escribió:

> Hi Adrián,
>
> In my case I learnt the bad way also. And now I purchased some Intel pci
> cards and add it to servers that have not much reliable cards.
>
> I post my bridge config:
>
>    Bridge "br0"
>         Port "phy-br0"
>             Interface "phy-br0"
>         Port "br0"
>             Interface "br0"
>                 type: internal
>     Bridge br-int
>         fail_mode: secure
>         Port "qvob27f1522-73"
>             tag: 1
>             Interface "qvob27f1522-73"
>         Port int-br-ex
>             Interface int-br-ex
>         Port br-int
>             Interface br-int
>                 type: internal
>         Port "qvo9da5cf98-f1"
>             tag: 1
>             Interface "qvo9da5cf98-f1"
>         Port "int-br0"
>             Interface "int-br0"
>         Port "qvo377f7953-d2"
>             tag: 1
>             Interface "qvo377f7953-d2"
>         Port "eth0"
>             Interface "eth0"
>         Port "qvod1b2e6dc-7f"
>             tag: 1
>             Interface "qvod1b2e6dc-7f"
>     Bridge br-ex
>         Port br-ex
>             Interface br-ex
>                 type: internal
>         Port phy-br-ex
>             Interface phy-br-ex
>     ovs_version: "2.0.2"
>
>
> El dom, 14 de dic 2014 a las 10:41 , Adrián Norte Fernández <
> adrian at bashlines.com> escribió:
>
> I had a similar problem because a cheap realtek card.
>
> Do you have logs? Can you post the result of ovs-vsctl show command on
> compute node?
> El 14/12/2014 22:36, "Gonzalo Aguilar Delgado" <
> gaguilar at aguilardelgado.com> escribió:
>
>> Hi Adrian,
>>
>> It's ubuntu trusty and network cards are intel, broadcom and marvell
>>
>> Marvell Technology Group Ltd. 88SE912x
>> Broadcom Corporation NetXtreme II BCM5706 Gigabit Ethernet
>> Intel Corporation 82574L Gigabit Network
>>
>> But I don't think it's the problem. It seems more a problem on config or
>> maybe a bug, since it doesn't fail on any other circumstances. And I'm
>> running ceph! It needs reliable network.
>>
>> What are you looking at?
>>
>>
>> El dom, 14 de dic 2014 a las 9:11 , Adrián Norte Fernández <
>> adrian at bashlines.com> escribió:
>>
>> Which SO and network cards are you using?
>> El 14/12/2014 19:35, "Gonzalo Aguilar Delgado" <
>> gaguilar at aguilardelgado.com> escribió:
>>
>>> Hi all,
>>>
>>> Question is. Why I have to restart neutron-plugin-openvswitch-agent to
>>> recover networking from new machines?
>>>
>>>
>>> I found a problem creating machines in a new node I've built. And
>>> realized that this happens on all my nodes of current version.
>>>
>>> neutron 2.3.4
>>> nova      2.17.0
>>>
>>> My logs shows that neutron assign vlan 2 for the new machine and this
>>> causes the machine to not have connectivity.
>>>
>>> ---------------------------------------
>>> 2014-12-14 17:57:07.321 12501 INFO
>>> neutron.plugins.openvswitch.agent.ovs_neutron_agent [-] Assigning 2 as
>>> local vlan for net-id=727b0415-f066-40d9-8147-183e07dd7b23
>>> 20
>>> ---------------------------------------
>>>
>>>
>>> If I shut down the machine and restart neutron plugin agent for ovs:
>>>
>>> #> restart neutron-plugin-openvswitch-agent
>>>
>>>
>>> ---------------------------------------
>>>  ...  assigning 1 as local vlan for
>>> net-id=727b0415-f066-40d9-8147-183e07dd7b23
>>>
>>> ---------------------------------------
>>>
>>>
>>> And the machine recovers connectivity.
>>>
>>> In the even of faillure (after creating it):
>>>
>>> If I start the machine without ip and set it manually I can try to ping
>>> it. It says it's not reachable but after restating the plugin automatically
>>> recovers connectivity. So no restart or further configuration is made.
>>>
>>>
>>> Maybe I have something wrong configured but it seems like a bug.
>>>
>>>
>>> Can someone point me to the right place?
>>>
>>>
>>> _______________________________________________
>>> Mailing list:
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>> Post to     : openstack at lists.openstack.org
>>> Unsubscribe :
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20141214/a5e8bfc0/attachment.html>