Re: NM refuses to manage a VLAN



On Fri, 2014-07-11 at 08:23 -0600, Pete Zaitcev wrote:
On Thu, 10 Jul 2014 12:58:20 -0500
Dan Williams <dcbw redhat com> wrote:

[root elanor network-scripts]# nmcli c up id vlan-ethgray
Error: Connection activation failed: Device not managed by NetworkManager or unavailable

I tried to reproduce this with that RPM version from F20 (and even on
i686!) but cannot.  But in any case, could you:

systemctl mask NetworkManager
systemctl stop NetworkManager
/usr/sbin/NetworkManager --no-daemon --log-level=debug

and then try to reproduce the issue?  Then we'll see if the logs
elucidate anything.

Full log is here:
 http://www.zaitcev.us/things/messages.nm.01

Could this be the issue?

Jul 11 07:46:01 elanor NetworkManager[21005]: ifcfg-rh:
parsing /etc/sysconfig/network-scripts/ifcfg-ethmain.4 ...
Jul 11 07:46:01 elanor NetworkManager[21005]: ifcfg-rh:     read
connection 'System ethmain.4'
Jul 11 07:46:01 elanor NetworkManager[21005]: ifcfg-rh: Ignoring
connection 'System ethmain.4' / device 'ethmain.4' due to
NM_CONTROLLED/BRIDGE/VLAN.

Do you have NM_CONTROLLED=no
in /etc/sysconfig/network-scripts/ifcfg-ethmain.4?

Also, " Ignoring connection 'System ethmain' / device
'00:23:54:9f:be:a5' due to NM_CONTROLLED/BRIDGE/VLAN." implies that the
VLAN interface will also be ignored, because it'll have the same MAC as
the parent.

Dan

The segment at the moment of the problem (nmcli c up id vlan-ethgray)
looks like this:

Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.479459] [platform/nm-platform.c:956] 
nm_platform_vlan_add(): link: adding vlan 'ethmain.4' parent 2 vlanid 4 vlanflags 0
Jul 11 07:49:03 elanor systemd-sysctl: Overwriting earlier assignment of kernel/sysrq in file 
'/etc/sysctl.d/99-sysctl.conf'.
Jul 11 07:49:03 elanor radvd[4223]: attempting to reread config file
Jul 11 07:49:03 elanor radvd[4223]: no linklocal address configured for ethmain.4
Jul 11 07:49:03 elanor radvd[4223]: resuming normal operation
Jul 11 07:49:03 elanor radvd[4223]: attempting to reread config file
Jul 11 07:49:03 elanor radvd[4223]: no linklocal address configured for ethmain.4
Jul 11 07:49:03 elanor radvd[4223]: resuming normal operation
Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.492995] 
[platform/nm-linux-platform.c:253] get_kernel_object(): get_kernel_object for link: ethmain.4 (13, family 0)
Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.493109] [platform/nm-platform.c:1969] 
log_link(): signal: link added: ethmain.4 (13)
Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.494238] [devices/nm-device.c:7086] 
nm_device_update_hw_address(): (ethmain.4): hardware address is 00:23:54:9F:BE:A5
Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.494415] [devices/nm-device-vlan.c:92] 
update_initial_hw_address(): (ethmain.4): read initial MAC address 00:23:54:9F:BE:A5
Jul 11 07:49:03 elanor NetworkManager[21005]: <info> (ethmain.4): carrier is OFF
Jul 11 07:49:03 elanor NetworkManager[21005]: <debug> [1405086543.494720] [devices/nm-device-vlan.c:532] 
constructed(): (ethmain.4): kernel ifindex 13
Jul 11 07:49:03 elanor NetworkManager[21005]: <info> (ethmain.4): VLAN ID 4 with parent ethmain
Jul 11 07:49:03 elanor NetworkManager[21005]: <info> (ethmain.4): new VLAN device (driver: '8021q' ifindex: 
13)
Jul 11 07:49:03 elanor NetworkManager[21005]: <info> (ethmain.4): exported as 
/org/freedesktop/NetworkManager/Devices/7
Jul 11 07:49:03 elanor NetworkManager[21005]: <info> (ethmain.4): No existing connection detected.

So... nmcli knows that connection exists, and NM daemon brings up the
right interface, but when the info circles back to NM daemon from kernel,
it cannot find the connection?

I tried this:
 nmcli c mod id vlan-ethgray 802-3-ethernet.mac-address 00:23:54:9F:BE:A5

But it made no difference.

-- Pete




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]