[Users] Creating bridge fails on Centos 6.3 node / vdsm

Wed Sep 5 11:54:05 UTC 2012

On Wed, Sep 05, 2012 at 08:59:33AM +0000, Joern Ott wrote:
> Hello everybody,
> 
> I installed oVirt Engine on a CentOS 6.3 host (also serving repositories via HTTP and NFS, running puppet, razor and svn as well as xinetd/TFTP, NFS, DNS and DHCP). When I set up a new node with a minimal CentOS 6.3 install and then use the OVirt GUI to install vdsm on it, I get an error setting up the bridging.
> 
> As a next step, I set up the node again and installed vdsm and vdsm-cli manually before setting up the node via the GUI. This time, I don't see any errors in the event log of the node in ovirt GUI. However, the bridge is not created and the node marked as "not operational". When I click on the node and click "set up Networks", I can assign the ovirtmgmt network to eth0. When I click on OK, I get "Error while executing action Setup Networks: Internal oVirt Engine Error" in the gui. The engine.log shows the following messages:
> 2012-09-05 10:49:03,833 INFO  [org.ovirt.engine.core.bll.SetupNetworksCommand] (ajp--127.0.0.1-8702-7) [3e1fa402] Running command: SetupNetworksCommand internal: false. Entities affected :  ID: 0c6f61b8-f72b-11e1-add6-1cc1de08c55e Type: VDS
> 2012-09-05 10:49:03,838 INFO  [org.ovirt.engine.core.vdsbroker.vdsbroker.SetupNetworksVDSCommand] (ajp--127.0.0.1-8702-7) [3e1fa402] START, SetupNetworksVDSCommand(vdsId = 0c6f61b8-f72b-11e1-add6-1cc1de08c55e), log id: 50e63074
> 2012-09-05 10:49:03,839 INFO  [org.ovirt.engine.core.vdsbroker.vdsbroker.SetupNetworksVDSCommand] (ajp--127.0.0.1-8702-7) [3e1fa402] FINISH, SetupNetworksVDSCommand, log id: 50e63074
> 2012-09-05 10:49:04,356 ERROR [org.ovirt.engine.core.vdsbroker.vdsbroker.BrokerCommandBase] (ajp--127.0.0.1-8702-7) [3e1fa402] org.ovirt.engine.core.vdsbroker.vdsbroker.VDSNetworkException: VDSGenericException: VDSNetworkException: Specified netmask or gateway but not ip
> 2012-09-05 10:49:04,357 ERROR [org.ovirt.engine.core.vdsbroker.VDSCommandBase] (ajp--127.0.0.1-8702-7) [3e1fa402] Command SetupNetworksVDS execution failed. Exception: RuntimeException: org.ovirt.engine.core.vdsbroker.vdsbroker.VDSNetworkException: VDSGenericException: VDSNetworkException: Specified netmask or gateway but not ip
> 
> In the installation log /tmp/vdsm-bootstrap-phase1-20120905092647.log on the node, I see these errors
> 
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 386 Bridge ovirtmgmt not found, need to create it.
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 438 getAddress Entry. url=http://mgmt0.be.tng:80/Components/vds/
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 460 getAddress return. address=mgmt0.be.tng port=80
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 859 makeBridge begin.
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 343 _getMGTIface: read host name: mgmt0.be.tng
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 350 _getMGTIface: using host name mgmt0.be.tng strIP= 10.4.116.9
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 356 _getMGTIface IP=10.4.116.9 strIface=eth0
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 897 makeBridge found the following bridge paramaters: ['BOOTPROTO=dhcp', 'IPV6INIT=yes', 'MTU=1500', 'ONBOOT=yes', 'UUID=79df5c4c-8df0-474b-9e98-22f3fb9b54c9']
> Wed, 05 Sep 2012 09:27:19 DEBUG    deployUtil 122 ['/usr/share/vdsm/addNetwork', 'ovirtmgmt', '', '', 'eth0', 'BOOTPROTO=dhcp', 'IPV6INIT=yes', 'MTU=1500', 'ONBOOT=yes', 'UUID=79df5c4c-8df0-474b-9e98-22f3fb9b54c9', 'blockingdhcp=true']
> Wed, 05 Sep 2012 09:27:20 DEBUG    deployUtil 126 Specified netmask or gateway but not ip
> 
> Wed, 05 Sep 2012 09:27:20 DEBUG    deployUtil 127 WARNING:Storage.LVM:Cannot create env file [Errno 2] No such file or directory: '/var/run/vdsm/lvm.env'
> WARNING:root:options BOOTPROTO is deprecated. Use bootproto instead
> WARNING:root:options ONBOOT is deprecated. Use onboot instead
> Traceback (most recent call last):
>   File "/usr/share/vdsm/configNetwork.py", line 1382, in <module>
>     main()
>   File "/usr/share/vdsm/configNetwork.py", line 1351, in main
>     addNetwork(bridge, **kwargs)
>   File "/usr/share/vdsm/configNetwork.py", line 835, in addNetwork
>     bridged=bridged, **options)
>   File "/usr/share/vdsm/configNetwork.py", line 771, in _addNetworkValidation
>     "Specified netmask or gateway but not ip")
> ConfigNetworkError: (22, 'Specified netmask or gateway but not ip')
> 
> Wed, 05 Sep 2012 09:27:20 DEBUG    deployUtil 930 makeBridge Failed to add ovirtmgmt bridge out=Specified netmask or gateway but not ip
> 
> err=WARNING:Storage.LVM:Cannot create env file [Errno 2] No such file or directory: '/var/run/vdsm/lvm.env'
> WARNING:root:options BOOTPROTO is deprecated. Use bootproto instead
> WARNING:root:options ONBOOT is deprecated. Use onboot instead
> Traceback (most recent call last):
>   File "/usr/share/vdsm/configNetwork.py", line 1382, in <module>
>     main()
>   File "/usr/share/vdsm/configNetwork.py", line 1351, in main
>     addNetwork(bridge, **kwargs)
>   File "/usr/share/vdsm/configNetwork.py", line 835, in addNetwork
>     bridged=bridged, **options)
>   File "/usr/share/vdsm/configNetwork.py", line 771, in _addNetworkValidation
>     "Specified netmask or gateway but not ip")
> ConfigNetworkError: (22, 'Specified netmask or gateway but not ip')
> 
> ret=22
> Wed, 05 Sep 2012 09:27:20 ERROR    deployUtil 947 makeBridge errored:  out=Specified netmask or gateway but not ip
> 
> The node gets its IP on eth0 (backend network) via dhcp. The other nic, eth1, is currently not connected but will be connected to the frontend network and also get its IP via DHCP.
> 
> Any help how to resolve the error 'Specified netmask or gateway but not ip' is appreciated.

I cannot see why the script thinks that you've specified netmask or
gateway. Would you do a little research?

Which version of vdsm-bootstrap do you have installed on Engine? Which
version of vdsm is to be installed on the host?

Would you share the content of your
/etc/sysconf/network-scripts/ifcfg-eth0?

Would anything work better if you comment out the IPV6INIT and UUID
lines just before installing the host? I'm suspecting that this is
somehow related to the root problem of http://gerrit.ovirt.org/#/c/7411/ :
we try to copy whatever in ifcfg-eth0 into the ifcfg-ovirtmgmt, which
does not always makes sense.

Regards,
Dan.