Does anyone have an idea what might be wrong with my network? The output below is from my little server, Debian Squeeze, kernel 3.2.0-4-amd64. The network does have issues every second day, the server sometimes not responding to anything after boot. Only shutting it down and disconnecting mains will make it work again. NIC is an onboard Nvidia:
00:0a.0 Ethernet controller: NVIDIA Corporation MCP79 Ethernet (rev b1)
accessed by the forcedeth driver.
Today the server started rather fine and works OK. Seconds after booting this was the output of
(note the RX line)
eth0 Link encap:Ethernet Hardware Adresse 00:01:2e:27:14:15
inet Adresse:192.168.25.1 Bcast:192.168.25.255 Maske:255.255.255.0
inet6-Adresse: fe80::201:2eff:fe27:1415/64 Gültigkeitsbereich:Verbindung
inet6-Adresse: 2a01:198:2d2:0:1:2eff:fe27:1415/64 Gültigkeitsbereich:Global
UP BROADCAST RUNNING MULTICAST MTU:1500 Metrik:1
RX packets:2243 errors:4506 dropped:0 overruns:0 frame:4506
TX packets:1551 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:335013 (327.1 KiB) TX bytes:724603 (707.6 KiB)
Since that point the errors didn't grow any further despite processing some 2 GB. I noticed this in /proc/interrupts:
22: 1961755 0 IO-APIC-fasteoi ohci_hcd:usb3, eth0
I already experimented with module settings of the forcedeth driver, to no avail. The settings don't seem to be applied in the first place.
Other days the overruns explode without network load. Am i having hardware issues? The box is connected to 8-port Cisco GBit switch. The installation is 2 weeks old, Debian Squeeze OOTB with a few network services installed and configured (isc-dhcp-server, bind9, radvd, nfs-kernel-server, lio-utils)
Dumping the old system 2 weeks ago and installing from scratch didn't help a bit. I was running a self- configured kernel then, but now it's debian default.
Any hints would be greatly appreciated.