[rdo-list] [TripleO] Setting NeutronBridgeMappings just for compute nodes in NFV deployment

Dan Sneddon dsneddon at redhat.com
Mon Apr 3 19:03:50 UTC 2017


On 04/03/2017 06:03 AM, Samuel Monderer wrote:
> Hi Marius,
> 
> Attached is the controller node template.
> I've also added below the ovs-vsct show results for both the compute and
> control node
> For some reason there is also a bridge called "undef" in both the
> controller and the compute
> 
> [heat-admin at overcloud-controller-0 ~]$ sudo ovs-vsctl show
> 00ce4561-418d-4161-b907-3b2222930f35
>     Manager "ptcp:6640:127.0.0.1"
>     Bridge br-int
>         Controller "tcp:127.0.0.1:6633 <http://127.0.0.1:6633>"
>         fail_mode: secure
>         Port br-int
>             Interface br-int
>                 type: internal
>     Bridge undef
>         Port undef
>             Interface undef
>                 type: internal
>     Bridge br-mng
>         fail_mode: standalone
>         Port "vlan204"
>             tag: 204
>             Interface "vlan204"
>                 type: internal
>         Port "vlan202"
>             tag: 202
>             Interface "vlan202"
>                 type: internal
>         Port br-mng
>             Interface br-mng
>                 type: internal
>         Port "vlan203"
>             tag: 203
>             Interface "vlan203"
>                 type: internal
>         Port "eno2"
>             Interface "eno2"
>         Port "vlan201"
>             tag: 201
>             Interface "vlan201"
>                 type: internal
>     Bridge br-up
>         Port br-up
>             Interface br-up
>                 type: internal
>     Bridge br-ex
>         fail_mode: standalone
>         Port br-ex
>             Interface br-ex
>                 type: internal
>         Port "eno3"
>             Interface "eno3"
>     ovs_version: "2.5.0"
> [heat-admin at overcloud-controller-0 ~]$
> 
> 
> [heat-admin at overcloud-compute-0 ~]$ sudo ovs-vsctl show
> 2f6bccd5-9653-42df-a1db-8cb5c3b0293a
>     Manager "ptcp:6640:127.0.0.1"
>     Bridge br-down
>         fail_mode: standalone
>         Port br-down
>             Interface br-down
>                 type: internal
>         Port "em3"
>             Interface "em3"
>     Bridge br-up
>         fail_mode: standalone
>         Port br-up
>             Interface br-up
>                 type: internal
>         Port "em4"
>             Interface "em4"
>     Bridge br-ex
>         fail_mode: standalone
>         Port "vlan204"
>             tag: 204
>             Interface "vlan204"
>                 type: internal
>         Port "vlan202"
>             tag: 202
>             Interface "vlan202"
>                 type: internal
>         Port "vlan201"
>             tag: 201
>             Interface "vlan201"
>                 type: internal
>         Port br-ex
>             Interface br-ex
>                 type: internal
>         Port "em2"
>             Interface "em2"
>     Bridge br-int
>         Controller "tcp:127.0.0.1:6633 <http://127.0.0.1:6633>"
>         fail_mode: secure
>         Port br-int
>             Interface br-int
>                 type: internal
>     Bridge undef
>         Port undef
>             Interface undef
>                 type: internal
>     ovs_version: "2.5.0"
> [heat-admin at overcloud-compute-0 ~]$
> 
> 
> 
> 
> On Fri, Mar 31, 2017 at 1:01 PM Marius Cornea <marius at remote-lab.net
> <mailto:marius at remote-lab.net>> wrote:
> 
>     Hi Samuel,
> 
>     Could you please share the nic template used for the controller node?
>     The bridges should be created according to the nic template so I would
>     like to see how it looks.
> 
>     Thanks!
> 
>     On Thu, Mar 30, 2017 at 4:59 PM, Samuel Monderer
>     <smonderer at vasonanetworks.com <mailto:smonderer at vasonanetworks.com>>
>     wrote:
>     > Hi,
>     >
>     > I would like to test my VNF on an openstack environment deployed using
>     > TripleO.
>     > For that I need to setup flat provider networks that go only to
>     the compute
>     > node.
>     > I tried deploying an overcoud with additional neutron bridge
>     mappings as
>     > following
>     >
>     > Added NeutronBridgeMappings: "
>     > datacentre:br-ex,physdown;br-down,physup:br-up" to network
>     environment file
>     > Added following resources to compute.yaml file
>     >
>     >             -
>     >               type: ovs_bridge
>     >               name: br-down
>     >               mtu: 9000
>     >               members:
>     >                 -
>     >                   type: interface
>     >                   name: ens4f0
>     >                   # force the MAC address of the bridge to this
>     interface
>     >                   primary: true
>     >             -
>     >               type: ovs_bridge
>     >               name: br-up
>     >               mtu: 9000
>     >               members:
>     >                 -
>     >                   type: interface
>     >                   name: ens4f1
>     >                   # force the MAC address of the bridge to this
>     interface
>     >                   primary: true
>     >
>     > This will create the intended OVS bridges on the the compute node
>     with the
>     > interfaces connected.
>     >
>     > The unwanted side effect it that ii also created the bridges on the
>     > controller node without any interfaces connected to it
>     >
>     > Is there a way to have the undercloud create the bridges only on
>     the compute
>     > nodes and not on the controller nodes???
>     >
>     > Samuel
>     >
>     > _______________________________________________
>     > rdo-list mailing list
>     > rdo-list at redhat.com <mailto:rdo-list at redhat.com>
>     > https://www.redhat.com/mailman/listinfo/rdo-list
>     >
>     > To unsubscribe: rdo-list-unsubscribe at redhat.com
>     <mailto:rdo-list-unsubscribe at redhat.com>
> 
> 
> 
> _______________________________________________
> rdo-list mailing list
> rdo-list at redhat.com
> https://www.redhat.com/mailman/listinfo/rdo-list
> 
> To unsubscribe: rdo-list-unsubscribe at redhat.com
> 

You've got a syntax error in your bridge_mappings (semicolon instead of
colon). This is causing your "undef" bridge.

Also, it is expected behavior that if you define a bridge in a bridge
mapping, but you don't include that bridge in the NIC template, that you
will end up with a bridge with no interfaces attached. It shouldn't hurt
anything.

-- 
Dan Sneddon         |  Senior Principal Software Engineer
dsneddon at redhat.com |  redhat.com/openstack
dsneddon:irc        |  @dxs:twitter




More information about the dev mailing list