What causes overruns, and how to troubleshoot/fix?

AlucardZero · 09-13-2007, 10:17 AM

Code:

alucard@organa:~$ sudo ifconfig
eth0      Link encap:Ethernet  HWaddr 00:01:02:77:9E:21
          inet addr:24.147.72.Z  Bcast:255.255.255.255  Mask:255.255.248.0
          inet6 addr: fe80::201:2ff:fe77:9e21/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:24113931 errors:0 dropped:0 overruns:336 frame:0
          TX packets:9640256 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:434177027 (414.0 MiB)  TX bytes:1872923977 (1.7 GiB)
          Interrupt:11 Base address:0x2000

eth1      Link encap:Ethernet  HWaddr 00:08:C7:CA:34:92
          inet addr:192.168.0.1  Bcast:192.168.0.255  Mask:255.255.255.0
          inet6 addr: fe80::208:c7ff:feca:3492/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2594358 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3281753 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:665139121 (634.3 MiB)  TX bytes:3106174744 (2.8 GiB)

Code:

alucard@organa:~$ lspci | grep -i eth
00:12.0 Ethernet controller: 3Com Corporation 3c905C-TX/TX-M [Tornado] (rev 74)
00:13.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 05)

I don't know how to tell which card is eth0 and which eth1 without having physical access to the box, which I don't at the moment.

This is on a Debian Etch server, 2.6.18 kernel.

The box has been up for 11 days, and has 336 overruns. This is bad, right? What are some things I can do to fix this?

Thanks for your time

osor · 09-13-2007, 01:06 PM

Quote:

Originally Posted by AlucardZero

I don't know how to tell which card is eth0 and which eth1 without having physical access to the box, which I don't at the moment.

You can figure out which device is which by using sysfs (assuming it is enabled in your kernel). E.g., you might get

Code:

$ ls "/sys/bus/pci/devices/0000:00:12.0" | grep net
net:eth0
$ ls "/sys/bus/pci/devices/0000:00:13.0" | grep net
net:eth1

(or vice versa).

Quote:

Originally Posted by AlucardZero

The box has been up for 11 days, and has 336 overruns. This is bad, right? What are some things I can do to fix this?

I am not sure. These are overruns on the receiving end, so it could mean a flaky device connected on the other end of the line. It could also mean your NIC itself is flaky. You might diagnose the problem further by compiling the module for the affected device with whatever debugging options enabled. Then look peruse the kernel logs to see when the overruns happen.

soroccoheaven · 09-13-2007, 11:43 PM

Quote:

RX packets:24113931 errors:0 dropped:0 overruns:336 frame:0
TX packets:9640256 errors:0 dropped:0 overruns:0 carrier:0

Receiver overruns usually occur when packets come in faster than the kernel can service the last interrupt (probably because of low memory).

Quote:

I don't know how to tell which card is eth0 and which eth1 without having physical access to the box, which I don't at the moment.

AFAIK you can know which card is eth0 or eth1 or so on as you can see interrupts by ifconfig -a
the card which gets first interrupt will be the eth0 and next will be eth1 and so on ..i cant see your eth1 interrupt but i am sure it should be more than 11..(because your eth0 already having interrupts 11)