Bug 684322

Since RHEL 6.1 External Beta has begun, and this bug remains
unresolved, it has been rejected as it is not proposed as
exception or blocker.

Red Hat invites you to ask your support representative to
propose this request, if appropriate and relevant, in the
next release of Red Hat Enterprise Linux.

Since RHEL 6.2 External Beta has begun, and this bug remains
unresolved, it has been rejected as it is not proposed as
exception or blocker.

Red Hat invites you to ask your support representative to
propose this request, if appropriate and relevant, in the
next release of Red Hat Enterprise Linux.

Interfaces taken out of the control of NM should be unmanaged by NM by writing NM_CONTROLLED=no to an ifcfg file that contains the HWADDR of the interface, by whatever tool takes control of that interface.

Keep in mind that this bug was filed in the context of RHEL6, where NM is supposed to ignore all bridge, bond, and vlan devices *and* the ethernet devices attached to bridges/bonds. Nobody has any desire to take these devices out of the control of NM, but this particular version of NM is unable to control them (and has no method of informing us of this so that we could selectively decide if we needed to set some variable).

The problem described here is that NM believes that it isn't managing the device, while the shell scripts behind /sbin/ifdown think that it is, the result being that neither path in the code (either the "do it with nmcli" or the "do it with ifconfig" form) is taken. (I just checked on my RHEL6 machine and this problem still exists).

Note that neither netcf nor libvirt has ever wanted to explicitly exclude NetworkManager from managing the interfaces it creates; they do nothing that  would preclude NM management of the device once that was supported, so don't want to close that door. It doesn't make sense for us to explicitly add a setting saying "don't touch this device", when NM can already understand from the rest of the config that it is unable to deal with this device anyway.

I had thought that at one time NM would notice when an ethernet device has "BRIDGE=xxx" in its ifcfg, and from then on ignore that device just as if it had NM_CONTROLLED=no (and actually I remember it working this way - perhaps it was in a Fedora release?)