I've just setup a new machine, in an unusual way (to avoid having to reinstall hundreds of packages and configure all the mail and samba and print etc. etc. again), by doing something risky:
1) Install a barebones Ubuntu 9.10 from CD
2) Copy *all* files from old PC also running 9.10 to the new
(yes, even all of / including /etc!)
3) Fixed /etc/fstab and grub config and /etc/hosts to use the
new hard drive arrangements and local IP addresses
4) Eventually got ethernet working, after realising eth0's name
had changed to eth1 on the new PC
Then things get weird. I read all the NetworkManager ubuntu documentation, but there was precious little in there to actually work out what it's doing under the hood and therefore diagnose what's wrong. After a lot of effort I could get NM to try to connect and after about 40 seconds, fail. It seemed to take the time that it took dhclient to fail if I ran it manually from the command line, like so::
Code:
# dhclient
Internet Systems Consortium DHCP Client V3.1.2
Copyright 2004-2008 Internet Systems Consortium.
All rights reserved.
For info, please visit http://www.isc.org/sw/dhcp/
Listening on LPF/eth1/e0:cb:4e:51:ff:23
Sending on LPF/eth1/e0:cb:4e:51:ff:23
Sending on Socket/fallback
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 6
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 7
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 8
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 12
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 13
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 8
DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 7
No DHCPOFFERS received.
No working leases in persistent database - sleeping.
Note that the network would work fine if I ran these commands as root:
Code:
# ifconfig eth1 inet 192.168.1.102 netmask 255.255.255.0 broadcast 192.168.1.255
# route add default gw 192.168.1.1
I purged my old NM packages, and reinstalled the ones from the PPA ppa:network-manager/trunk. What happens for me now is that NM tries to connect, and just sits there forever trying, according to the status bar.
On the plus side, if I run dhclient manually now, it instantly gets a correct lease and the network is up.
Because I can't find a way to discover what NM is actually doing when it tries to connect (if I try running it from the command line with debug enabled, it just complains that NM is already running), I've decided the package jut isn't worth the trouble, and will just change /etc/network/interfaces to run dhcpc itself, and tell NM that it should manage nothing.
I've done lots of googling to try to fix the problem, and NM just seems too much a black box, not diagnosable, so I'm scrapping it.
From the logs on my firewall/router, it looks like NM links to it on port 137 (so I assume it's asking for a DHCP lease), but it never seems to get one. Whereas now if I run dhclient manually, I do get one.
For what it's worth, here are my /etc/network/interfaces file before I got rid of NM:
Code:
# This file describes the network interfaces available on your system
# and how to activate them. For more information, see interfaces(5).
# The loopback network interface
auto lo
iface lo inet loopback
# Try adding this:
# The primary network interface
auto eth1
/etc/NetworkManager/nm-system-settings.conf before I changed the managed to false:
Code:
# This file is installed into /etc/NetworkManager, and is loaded by
# NetworkManager by default. To override, specify: '--config file'
# during NM startup. This can be done by appending to DAEMON_OPTS in
# the file:
#
# /etc/default/NetworkManager
#
[main]
plugins=ifupdown,keyfile
[ifupdown]
managed=true
luke
Bookmarks