2 nics on 2 systems keep dropping packets, overrun and frame errors

[psycroptic]

I have 2 Archlinux boxes on my home network which all runs through an 8-port Intellinet gigabit switch. 1 box is a NAT iptables router system with 2 network interfaces, one DHCP-assigned for my cable modem (comcast) and another interface with a static IP 172.16.0.1. The 2nd box is a file & network server, running samba, bind, dhcpd, and openntpd for the local network, with static address 172.16.0.2. Both the internal-facing interface in the router and the interface on the fileserver are both onboard gigabit adapters on nvidia-based motherboards and they both use the same "forcedeth" driver (version 0.64). The fileserver has my old gaming mobo "ASUS M2N32-SLI Deluxe" from circa 2005, and the router has a newly-bought ASUS "M2N68-AM Plus" board. The interface connected to the cable modem is an add-in PCI express card Intel EXPI9301CTBLK, using the e1000e driver.

Everything seems to be working fine as far as data transfer goes, but i've noticed this recently on both the nvidia nics, the first item being ifconfig output from the router and the second from the fileserver:

ROUTER (eth0 is intel, eth1 nvidia)
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500  metric 1
        inet 76.22.129.200  netmask 255.255.254.0  broadcast 255.255.255.255
        ether 00:1b:21:c8:3d:34  txqueuelen 16  (Ethernet)
        RX packets 36016021  bytes 30878687681 (28.7 GiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 12947356  bytes 2313527854 (2.1 GiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
        device interrupt 19  memory 0xdffe0000-e0000000

eth1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500  metric 1
        inet 172.16.0.1  netmask 255.255.0.0  broadcast 172.16.255.255
        ether bc:ae:c5:c9:89:35  txqueuelen 16  (Ethernet)
        RX packets 13022109  bytes 2298983072 (2.1 GiB)
        RX errors 99027  dropped 953  overruns 99026  frame 954
        TX packets 22751031  bytes 30083421724 (28.0 GiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
        device interrupt 46

FILESERVER
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500  metric 1
        inet 172.16.0.2  netmask 255.255.0.0  broadcast 172.16.255.255
        ether 00:17:31:ce:e4:85  txqueuelen 1000  (Ethernet)
        RX packets 4931  bytes 3149770 (3.0 MiB)
        RX errors 5  dropped 0  overruns 5  frame 2
        TX packets 4318  bytes 613602 (599.2 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
        device interrupt 43  base 0x4000

As you can see, there are many RX errors on both systems (the fileservers numbers are typically similar to the router, i had just rebooted it recently.) It's only within the local network; the external intel interface in the router doesn't have the problem.

That fileserver's mobo has 2 network jacks, both on the NVIDIA MCP55 chipset, and i've tried swapping the active interface between both jacks with no result. All of the cabling is new (mid 2011 to now), and the gigabit switch is ~2009/2010.

Should I be concerned about this? I use the fileserver all the time as my main media storage, so it transmits tens of gigabytes a day. I'm suspecting that switch might be going bad?

Any thoughts? TIA

[Rubberman]

Since both are having problems, I'd advise looking at the switch. Try swapping it out with another and see if the problems go away. I was having similar problems with a gigabit switch I had when it started overheating.