Strange intermittant network failure...

bulliver · 10-25-2005, 07:47 PM

Hello all...

I have recently (2 weeks or so) noticed my net connection has been going down at random times. I generally have to cycle the power on my cable modem to get it back up. My network consists of a linux router/gateway (hostname: vyvyan) and five other natted clients, running linux/solaris/OSX.

Now I got my ISP techs to come down here, and after some extensive line and modem tests they told me that the problem is with my network, and not their lines. This seemed to be confirmed when testing the connection directly with my laptop (ie: bypassing vyvyan)

So, I am positive I have not changed any networking settings since before this started happening...and I should note that when the connection goes down, all the clients can still connect with each other and with vyvyan no problem.

I am hoping this is simply a hardware failure with vyvyan's eth0, since this is the only interface that seems to be screwing up. I don't have a spare nic right now to test this though...

Here is output of ifconfig on vyvyan (ip scrubbed):

Code:

eth0      Link encap:Ethernet  HWaddr 00:40:F4:39:9D:90
          inet addr:24.X.X.X  Bcast:24.255.255.255  Mask:255.255.254.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:18597889 errors:61915 dropped:0 overruns:0 frame:0
          TX packets:14998757 errors:0 dropped:0 overruns:0 carrier:0
          collisions:66893 txqueuelen:1000
          RX bytes:2140821997 (2041.6 Mb)  TX bytes:2615270966 (2494.1 Mb)
          Interrupt:10 Base address:0xd000

Note the huge number of RX errors. I am also getting errors like this in my logs:

Code:

Oct 25 17:14:37 vyvyan eth0: link down
Oct 25 17:14:38 vyvyan eth0: link up, 10Mbps, half-duplex, lpa 0x0000
Oct 25 17:14:38 vyvyan eth0: link down
Oct 25 17:14:38 vyvyan eth0: link up, 10Mbps, half-duplex, lpa 0x0000
Oct 25 17:16:39 vyvyan eth0: link up, 10Mbps, half-duplex, lpa 0x0000

I have not been able to confirm these errors are getting written in response to my problem, but I suspect they are.

So, should I just go get a new nic already, or is there more troubleshooting I should be doing? Has anyone seen a situation like this where it wasn't hardware failure? Is there any software tool that can check if the nic is indeed malfunctioning?

Any insight appreciated. If you need more info or output of a command please just let me know.

Thanks.

jailbait · 10-26-2005, 01:17 PM

"So, should I just go get a new nic already, or is there more troubleshooting I should be doing?"

In the dim dark distant past I saw a similar problem where a wire was broken within a cable and thus intermittently not making contact. Swap the cable on your nic and see if the problem goes away.

-------------------------
Steve Stites

bulliver · 11-01-2005, 02:53 PM

Hey.

So I swapped the cable, but the problem remains.

I have ordered a new NIC, and I am hopeful that this will take care of it but I am still unsure. When I reset the interface, the problem still exists. A power cycle on the modem is the _only_ way I have found to get it going again.

Going to take a week or so to get the new NIC, so I am still open to ideas in the interim...