[ovs-discuss] CentOS 6.4 warning: "RTNETLINK answers: File exists" while network restart.

Gurucharan Shetty shettyg at nicira.com
Wed Mar 13 20:54:11 UTC 2013


On Wed, Mar 13, 2013 at 9:54 AM, Gurucharan Shetty <shettyg at nicira.com> wrote:
> On Wed, Mar 13, 2013 at 7:33 AM, Denis Iskandarov
> <d.iskandarov at gmail.com> wrote:
>> Again occurred warning from CentOS 6.3, reported here:
>> http://openvswitch.org/pipermail/discuss/2012-October/008322.html
>>
>> Everything works. No ip disconnects.
>> In my opinion it's seems to be like warning message that device with such
>> MAC address.
>> But if systems saying me something i want to understand it, if it is just
>> warning or can lead to something critical on production env with hundreds of
>> virtual interfaces.
>>
>> Environment:
>> VMware Workstation 9.0.2
>> Centos 6.4 x86_64; 2.6.32-358.0.1.el6.x86_64
>> OpenvSwitch 1.9.0 with applied patch:
>> http://www.mail-archive.com/dev@openvswitch.org/msg17555.html
>> Logs of log/messages and sh -vx outputs attached.
>>
>> Network scripts:
>> DEVICE=ovs0
>> ONBOOT=yes
>> DEVICETYPE=ovs
>> TYPE=OVSBridge
>> NM_CONTROLLED=no
>> BOOTPROTO=static
>> IPADDR=192.168.0.111
>> NETMASK=255.255.255.0
>> HOTPLUG=no
>>
>> DEVICE=eth0
>> ONBOOT=yes
>> DEVICETYPE=ovs
>> TYPE=OVSPort
>> OVS_BRIDGE=ovs0
>> NM_CONTROLLED=no
>> BOOTPROTO=none
>> HOTPLUG=no
>>
>>
>> service network restart
>> Shutting down interface eth0:                              [  OK  ]
>> Shutting down interface ovs0:                              [  OK  ]
>> Shutting down loopback interface:                          [  OK  ]
>> Bringing up loopback interface:                            [  OK  ]
>> Bringing up interface eth0:                                [  OK  ]
>> Bringing up interface ovs0:  RTNETLINK answers: File exists [  OK  ]
>
> What happens if you change your bridge name to "breth0" instead of
> ovs0. Most probably this error will go away.
>
> I am guessing here. Configuring ovs0 tries to reconfigure eth0 (the
> scripts run in alphabetical order). But since eth0 is already
> configured, we see this error. It is harmless.

For anyone else interested in the reason for the error, it looks like
the error comes from trying to add zeroconf route twice in
/etc/sysconfig/network-scripts/ifup-eth. If someone does not want this
error show up (and don't want default zeroconf route), they can add
the following line in the ifcfg-* config.

NOZEROCONF=yes

>
>>
>>
>> Mar 13 16:51:24 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --if-exists del-port ovs0 eth0
>> Mar 13 16:51:24 ovs-cent64 kernel: device eth0 left promiscuous mode
>> Mar 13 16:51:24 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --if-exists del-br ovs0
>> Mar 13 16:51:24 ovs-cent64 kernel: device ovs0 left promiscuous mode
>> Mar 13 16:51:24 ovs-cent64 kernel: lo: Disabled Privacy Extensions
>> Mar 13 16:51:24 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --may-exist add-br ovs0
>> Mar 13 16:51:24 ovs-cent64 kernel: device ovs0 entered promiscuous mode
>> Mar 13 16:51:28 ovs-cent64 kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full
>> Duplex, Flow Control: None
>> Mar 13 16:51:28 ovs-cent64 kernel: ADDRCONF(NETDEV_UP): eth0: link is not
>> ready
>> Mar 13 16:51:28 ovs-cent64 kernel: ADDRCONF(NETDEV_CHANGE): eth0: link
>> becomes ready
>> Mar 13 16:51:28 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --may-exist add-port ovs0 eth0
>> Mar 13 16:51:28 ovs-cent64 kernel: device eth0 entered promiscuous mode
>> Mar 13 16:51:28 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --may-exist add-br ovs0
>> Mar 13 16:51:28 ovs-cent64 ovs-vsctl: 00001|vsctl|INFO|Called as ovs-vsctl
>> -t 10 -- --may-exist add-br ovs0
>>
>> _______________________________________________
>> discuss mailing list
>> discuss at openvswitch.org
>> http://openvswitch.org/mailman/listinfo/discuss
>>



More information about the discuss mailing list