Hi,

I have Ubuntu 6.06 LTS running on an HP DL140 server.

After the machine has been on for about 8 - 10 minutes it suddenly cannot access the internet, it can only access servers on its local lan.

I've ruled out a network/switch/cabling problem as all other machines are able to access the internet fine even when this machine can't.

A reboot appears to fix the problem for a short while but it means I need to reboot the server every 15 minutes to keep network working.

I've disabled ACPI on the server completely, by setting it as per what this chap said: http://neoaddict.wordpress.com/2007/...-acpi-on-boot/

Any one got any ideas.

Thanks

Look into the error log of the router if you use one. And try to restart the network device:
System - Administration - Network

Uncheck the connection, wait, and check it again.

I've seen similar posts quite often on the ubuntu forums and/or here and the cause is often the dns lookup. When it happens, try to ping 66.249.93.104 or enter http://66.249.93.104 in your browser. The latter should give you google; if that's indeed the case or ping works as expected, it's a dns issue. Unfortunately I don't have the solution, but somebody else might know or you can do some more research based on that.

Thanks for your replies.

I've not been able to restart the network router, as it is a Cisco 6506 Layer 3 switch (router) and as all the traffic from other servers and workstations is going through it fine, i don't think theres anything wrong with the router.

As for the DNS issue, this appears to be a lower OSI problem. DNS resolves fine, but if i then try to ping it fails.

For example (DNS is okay, pinging shows its not getting through)
root@xenon:/temp# ping www.bbc.co.uk
PING www.bbc.net.uk (212.58.251.201) 56(84) bytes of data.

--- www.bbc.net.uk ping statistics ---
21 packets transmitted, 0 received, 100% packet loss, time 20002ms

Then after a reboot:
root@xenon:~# ping www.bbc.co.uk
PING www.bbc.net.uk (212.58.251.201) 56(84) bytes of data.
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=1 ttl=249 time=8.83 ms
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=2 ttl=249 time=9.88 ms
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=3 ttl=249 time=11.0 ms
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=4 ttl=249 time=9.15 ms
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=5 ttl=249 time=9.26 ms
64 bytes from www1.telhc.bbc.co.uk (212.58.251.201): icmp_seq=6 ttl=249 time=12.3 ms

--- www.bbc.net.uk ping statistics ---
6 packets transmitted, 6 received, 0% packet loss, time 5102ms
rtt min/avg/max/mdev = 8.836/10.097/12.376/1.250 ms

Working fine again, for about 10 minutes then it stops.
Any log files i can check on to see if it is crashing or something?

Thanks

I didn't think there was anything wrong with the router, but there are a couple of reasons for a router to block you temporarily. That would show up in the error logs.

Restarting the ethernet device (see post above) should do the same as restarting the computer... It reconnects to the router, you get re-assigned by dhcp. If that helps it might be a network card or router issue.

Hi, thanks for your suggestion.

How do you restart networking via the CLI, dont have the GUI installed?

Just to test i left it pinging something off network over night, as you can see it worked all night, just seems to be if it is intermittent traffic that the router appears to ignore it.

64 bytes from 10.11.0.10: icmp_seq=53882 ttl=128 time=0.385 ms
64 bytes from 10.11.0.10: icmp_seq=53883 ttl=128 time=0.365 ms
64 bytes from 10.11.0.10: icmp_seq=53884 ttl=128 time=0.420 ms
64 bytes from 10.11.0.10: icmp_seq=53885 ttl=128 time=0.383 ms

--- 10.11.0.10 ping statistics ---
53885 packets transmitted, 53755 received, 0% packet loss, time 53908314ms
rtt min/avg/max/mdev = 0.333/0.400/13.644/0.232 ms

I also built another server just to test, put Ubuntu 6.06 on it, its an HP ML 110 rather than a DL 140, and guess what it does exactly the same thing.

They both have broadcom network cards, anyone know of an issue with the broadcom NIC drivers and Ubuntu 6.06?

Cheers

sudo /etc/init.d/networking restart

The bcm43xx driver is broken with the 2.6.20.8 kernel: https://launchpad.net/ubuntu/+source...ter/+bug/85099

Okay i've found the cause of the problem appears to be a cisco problem:

Spot the difference layer 2 problem:

Arp -a
When it is not working
? (192.168.197.1) at 00:1B:CD:CC:69:83 [ether] on eth0

When it is working.
? (192.168.197.1) at 00:2A:A2:EC:18:40 [ether] on eth0

.1 being the default gateway.

Now to find out why the mac address of another server is being transposed over the correct mac address of the ip for the router. Very strange but at least i know whats going on now!
The show ip interfaces on the switch shows this (the wrong) mac address to belong to another server on another port!

Thanks for your help.
Tristan

I realize this post is VERY old but I'm having the identical problem.

Were you able to identify the root cause of the switch/router advertising itself as the Mac w/ said IP?