← Back to team overview

openstack team mailing list archive

Re: Open vSwitch Errors on When Launching Instance


Thanks Akihiro!

Jun 26 21:35:07|00103|netdev_linux|WARN|ethtool command ETHTOOL_GDRVINFO on
network device tap86944445-96 failed: No such device
Jun 26 21:35:07|00104|netdev_linux|WARN|ethtool command ETHTOOL_GSET on
network device tap86944445-96 failed: No such device

root@test1:~# ovs-vsctl list-ifaces br-int
tap86944445-96  <--------------

root@test1:~# ip netns exec qdhcp-357165e7-a848-4a10-bd86-f121752bbe8c ip
link show
15: tap86944445-96: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc
noqueue state UNKNOWN  <--------------
    link/ether fa:16:3e:c6:da:f3 brd ff:ff:ff:ff:ff:ff
16: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00

Looks like it working ok then.   I am not sure if I should be happy or sad.
 I was hoping this would lead me to the cause of my bigger problem:  The
First VM launches fine, when I try launching a second one I get these

/var/log/quantum/openvswitch-agent.log.1:2013-06-26 21:36:21    ERROR
[quantum.agent.linux.ovs_lib] Unable to execute ['ovs-vsctl',
'--timeout=2', 'get', 'Interface', 'qvoc978ce9d-d4', 'external_ids'].
/var/log/quantum/openvswitch-agent.log.1:Stderr: 'ovs-vsctl: no row
"qvoc978ce9d-d4" in table Interface\n'
/var/log/quantum/server.log.1:2013-06-26 21:38:38  WARNING
[quantum.db.agentschedulers_db] Fail scheduling network {'status':
u'ACTIVE', 'subnets': [u'd11f8cc2-1b10-4837-aee9-0d203e1f6b4e'], 'name':
u'demo-net', 'provider:physical_network': None, 'admin_state_up': True,
'tenant_id': u'acca139f15944782abf83e16ee6cfe8a', 'provider:network_type':
u'gre', 'router:external': False, 'shared': False, 'id':
u'357165e7-a848-4a10-bd86-f121752bbe8c', 'provider:segmentation_id': 2L}

Thanks for the help!


On Wed, Jun 26, 2013 at 11:52 PM, Akihiro MOTOKI <amotoki@xxxxxxxxx> wrote:

> Hi Samuel,
> Open vSwitch plugin creates a network device and then moves it to a
> network namespace.
> After the device is moved to a network namespace, we cannot see the device
> in the root namespace.
> This is the reason that OVS log message reports "no such device".
> If ovs-vsctl show the device and you can find the device in one of network
> namespaces,
> it looks no problem in most cases though it is a little confusing.
> Thanks,
> Akihiro
> 2013/6/27 Samuel Winchenbach <swinchen@xxxxxxxxx>
>> Hello
>> I have taken a step back and am trying to get an "all-in-one" node up and
>> running.  Everything works wonderful (a few warnings in the log about using
>> default MAC addresses) until I launch an instance.
>> here is the output of openvswitch-vswitchd.log:
>> http://pastie.org/pastes/8084556/text
>> Does anyone have any clue what could be causing this?  It says it creates
>> the tap device and then it can't find it?
>> P.S. I know those are warnings are in the logs, not errors... but based
>> on the severity of the problem it seems as though they should be errors!
>> If you need any more information please ask!
>> Thanks,
>> Sam
>> _______________________________________________
>> Mailing list: https://launchpad.net/~openstack
>> Post to     : openstack@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~openstack
>> More help   : https://help.launchpad.net/ListHelp
> --
> Akihiro MOTOKI <amotoki@xxxxxxxxx>