NFS-Root support broken in Hardy

[Torsten Nahm]

Binary package hint: hal

When booting with NFS, the root directory is mounted over NFS. Starting hald (e.g. as part of runlevel 2) break the NFS connection and leaves the system rootless. I found that putting the line

iface eth0 inet manual

into /etc/network/interfaces resolves the problem (when root is mounted over eth0). Still, hald should only report and not erroneously try to auto-configure any network interfaces.

Version:
Ubuntu Hardy Heron Release Candidate
hal 0.5.11~rc2-1ub

[Daniel T Chen (crimsun)]

Is this symptom still reproducible in 8.10 or 9.04?

[Torsten Nahm (tnahm)]

Yes, the problem is still present in Intrepid.
It seems to also to depend on the launch of NetworkManager as part of runlevel 2. Only if hal and NetworkManager are launched is the NFS root connection broken.

If I have "iface eth0 inet manual" set, the system boots successfully, and Network Manager reports (from /var/log/syslog):

Jan 8 21:54:06 sleek NetworkManager: <info> Unmanaged Device found; state CONNECTED forced. (see http://bugs.launchpad.net/bugs/191889)

If the line is missing, the boot process stops after Network Manager is started. The last entries in the syslog are:

an 8 21:36:51 sleek NetworkManager: <info> starting...
Jan 8 21:36:51 sleek NetworkManager: <info> eth0: driver is 'via-rhine'.
Jan 8 21:36:51 sleek NetworkManager: <info> Found new Ethernet device 'eth0'.
Jan 8 21:36:51 sleek NetworkManager: <info> (eth0): exported as /org/freedesktop/Hal/devices/net_00_c0_9f_1d_41_01
Jan 8 21:36:51 sleek NetworkManager: <info> Trying to start the supplicant...
Jan 8 21:36:51 sleek NetworkManager: <info> Trying to start the system settings daemon...
Jan 8 21:36:51 sleek NetworkManager: <info> (eth0): carrier now ON (device state 1)
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: init!
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: update_system_hostname
Jan 8 21:36:51 sleek nm-system-settings: SCPluginIfupdown: guessed connection type (eth0) = 802-3-ethernet
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: update_connection_setting_from_if_block: name:eth0, type:802-3-ethernet, autoconnect:0, id:Ifupdown (eth0), uuid: 681b428f-0000-0000-0000-000039000000
Jan 8 21:36:51 sleek nm-system-settings: SCPluginIfupdown: management mode: unmanaged
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: devices added (udi: /org/freedesktop/Hal/devices/net_00_c0_9f_1d_41_01, iface: eth0)
Jan 8 21:36:51 sleek nm-system-settings: Ifupdown: get unmanaged devices count: 1
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: (157326576) ... get_connections.
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: (157326576) connections count: 1
Jan 8 21:36:51 sleek nm-system-settings: Ifupdown: get unmanaged devices count: 1
Jan 8 21:36:51 sleek nm-system-settings: SCPlugin-Ifupdown: end _init.
Jan 8 21:36:51 sleek nm-system-settings: Loaded plugin ifupdown: (C) 2008 Canonical Ltd. To report bugs please use the NetworkManager mailing list.
Jan 8 21:36:51 sleek nm-system-settings: Loaded plugin keyfile: (c) 2007 - 2008 Red Hat, Inc. To report bugs please use the NetworkManager mailing list.
Jan 8 21:36:51 sleek NetworkManager: <info> (eth0): now unmanaged

[Victor Vargas (kamus)]

Since this report have a long time without activity, could you check (if is possible) in latest Ubuntu Karmic version if this issue is still happening? Thanks in advance.

[ianr (ian-rickards)]

Yes, just checked this and "iface eth0 inet manual" is still needed with Karmic, otherwise init hangs
(ARM platform, as it happens)

[Launchpad Janitor (janitor)]

[Expired for network-manager (Ubuntu) because there has been no activity for 60 days.]