NetworkManager runs DHCP clients on CORE-generated interfaces

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Start a CORE scenario on Ubuntu 14.04, CentOS 7.0, etc.
2. Messages may pop up (CentOS 7.0) "Connection Failed: Activation of network 
connection failed"
3. All virtual interfaces appear in nm-applet drop-down
4. tcpdump within CORE shows DHCP BOOTP messages and clients make requests; 
potentially dozens of dhclient processes appear on the host
5. If you have the DHCP server service running, host interfaces (e.g. n3.0.234) 
start receiving IP addresses (normally they have no addresses and are joined to 
a bridge)

What is the expected output? What do you see instead?

We want NetworkManager to ignore CORE-generated interfaces, since CORE will 
take care of assigning IP addresses.

Original issue reported on code.google.com by ahrenh...@gmail.com on 1 Aug 2014 at 2:49

[GoogleCodeExporter]

'nmcli dev status' lists interfaces and their state within NetworkManager

The /etc/NetworkManager/NetworkManager.conf file can be updated with these 
lines:

[keyfile]
unmanaged-devices=mac:00:13:e0:54:86:12;mac:00:14:e4:b3:22:18

and 'sudo service network-manager reload' to re-read the configuration file.

When the devices have the name 'veth*' then they are automatically ignored.

Original comment by ahrenh...@gmail.com on 4 Aug 2014 at 3:18

[GoogleCodeExporter]

using device names of "vethA.B.SS" is ignored under Ubuntu 14.04 but not for 
CentOS 7.0 with updated NetworkManager package.

Ubuntu has a patch to ignore veth:
http://bazaar.launchpad.net/~network-manager/network-manager/ubuntu/view/head:/d
ebian/patches/add-veth-support.diff

Original comment by ahrenh...@gmail.com on 4 Aug 2014 at 5:13

• Changed state: Started

[GoogleCodeExporter]

fixed for Ubuntu in r553

Original comment by ahrenh...@gmail.com on 4 Aug 2014 at 5:29