Hi all,

I have a pretty standard Debian 5.0.2 system (2.6.26-2-amd64 kernel), with a 4-core Phenom processor. Everything has been working great for years, and suddenly during a reboot, I got an error message about statd failing.

Indeed, I didn't have any network interface.

I was able to bring up eth0 with:

ifconfig eth0 up

But there was no ipv4 address. I added a static address with:

ifconfig eth0 192.168.0.248

and, then brought up eth0, and restarted networking. [This is the address my router has reserved and forwards to for this server, so that's why I'm not using DHCP]. At least I had an interface again. I can ssh into the machine, and Apache is serving pages again. So it appears that the networking card, cable, and router are all still working well.

For good measure, I restarted portmap.

Now, at least, I get the following:

# ifconfig
eth0 Link encap:Ethernet HWaddr 00:24:8c:c2:03:2c
inet addr:192.168.0.248 Bcast:192.168.0.255 Mask:255.255.255.0
inet6 addr: fe80::224:8cff:fec2:32c/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2131 errors:0 dropped:49777272750 overruns:0 frame:0
TX packets:377 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:151268 (147.7 KiB) TX bytes:61530 (60.0 KiB)
Interrupt:254 Base address:0x2000

Well, when I try to restart nfs-common, I again get a "failed!" message:

# /etc/init.d/nfs-common restart
Stopping NFS common utilities: statd.
Starting NFS common utilities: statd failed!

Also, I still cannot ping from the machine, receiving an unknown host error.

# ping www.ping.com
ping: unknown host www.ping.com

And I cannot ssh off the machine (receiving a "name or service not known error"), but I can ssh into the machine just fine.

Finally, when I reboot, I have to manually add the IP address again. This, despite the fact that the correct info is still in the interfaces file:

# cat /etc/network/interfaces

# --------------
# 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

# The primary network interface
#allow-hotplug eth0
#iface eth0 inet dhcp

# Static address
auto eth0
iface eth0 inet static
address 192.168.0.248
netmask 255.255.255.0
network 192.168.0.0
broadcast 192.168.0.255
gateway 192.168.0.1

I can't think of anything that might have changed to cause this problem. I searched my root history to verify that I hadn't been doing anything.

This did follow, however, a hard shut-down due to a prolonged power failure. I don't think there was a surge. All the drives checked fine on reboot. Just the networking seemed messed up.

Any help or hints you could give me would be greatly appreciated.

-- Madeleine.

It look like power failure make network configuration messed up. You can use NetworkManager tool to reconfigure the network instead of editing network file.

Well, maybe, or maybe not. I haven't been able to determine that for sure. Nor have you provided any information or diagnosis of the problem to indicate that it was due to a power failure. So I think that's probably a premature diagnosis.

I'll keep searching. If you have other ideas, I'd appreciate hearing them.

Regards,
Maddie.

You should check /var/log/syslog and /var/log/messages for better error messages.

If I am not mistaken NFS relies on DNS in order to perform reverse IP resolution. On your machine DNS is not currently working. Did you have bind installed?

Also check in hosts.allow and hosts.deny whether portmapper is allowed to run on all interfaces.

I had this same problem years ago, also on Debian. Unfortunately I don't recall what happened. However, it was not NFS itself which was broken. When the IP stack is not fully functional, or portmapper is not, NFS won't work.

jlinkels

My /var/log/syslog and messages are full of martian source messages. For example:

Jul 31 16:13:14 PeavyServer kernel: [ 1777.185868] martian source 192.168.0.248 from 119.63.196.79, on dev eth0
Jul 31 16:13:14 PeavyServer kernel: [ 1777.185868] ll header: 00:24:8c:c2:03:2c:00:24:7b:20:f1:5c:08:00
Jul 31 16:16:10 PeavyServer kernel: [ 1963.278109] martian source 192.168.0.248 from 208.117.51.112, on dev eth0
Jul 31 16:16:10 PeavyServer kernel: [ 1963.278109] ll header: 00:24:8c:c2:03:2c:00:24:7b:20:f1:5c:08:00

It doesn't appear to me that BIND is installed. I checked:

named -v

and:

man named

but it indicates that is not installed.

My resolv.conf file was automatically updated at some point in the past. I reset it to what was working originally, which is:

domain gateway.2wire.net
search gateway.2wire.net
nameserver 192.168.0.1

Our router is 192.168.0.1. Perhaps there was an automatic firm-ware update on the router that caused this problem??

Both the hosts.allow and hosts.deny don't include anything other than comments.

Thanks for your help,
Maddie.

You can sniffer package when doing ping to www.ping.com and analyze tracing to find out why name resolution is failure.

The messages mentioning martian source are not expected. It is the first time I see those, so for me it is new as well. I looked it up and saw that they indicate traffic from IP addresses which somehow are incorrect with respect to your network settings. You should check on that.

About portmapper, I remember that I had to explicitely allow portmapper in hosts.allow and on all interfaces. Even when this should not be necessary. Given the martian source messages, I would concentrate on that first. Not only indicate the messages a probable source for the problem, those messages could indicate a IP spoofing problem. It doesn t make sense to relax what is allowed in such a case.

jlinkels