I have a gentoo router with 3 NICs running on 2.6.35-gentoo-r4 kernel. It seems like one of the NICs (Intel PRO/1000GT) for local LAN (eth1) randomly stops working like once in a few days (there is no pattern in downtime, but usually it happens in mornings when more LAN users comes in and in night when backups are running through the network). Other NICs seems to work ok while eth1 is down (i can ping outside internet via eth0, but everything dies on LAN). Sometimes eth1 NIC starts to work after ~20min downtime, sometimes i have to reboot router in order to make it work. I changed this NIC with another (same Intel PRO/1000 chip) but problem still persists. It was working fine for a year until i upgraded Cacti monitoring tool and added bunch of devices to monitor. Any ideas where to dig? Thanks.
Only thing i notice in /var/log/messages is:
------------[ cut here ]------------ WARNING: at net/sched/sch_generic.c:258 dev_watchdog+0xb9/0x10c() Hardware name: NETDEV WATCHDOG: eth1 (e1000): transmit queue 0 timed out Pid: 0, comm: swapper Not tainted 2.6.35-gentoo-r4 #5 Call Trace: [<c102aff6>] warn_slowpath_common+0x60/0x75 [<c12b177f>] ? dev_watchdog+0xb9/0x10c [<c102b06f>] warn_slowpath_fmt+0x26/0x2a [<c12b177f>] dev_watchdog+0xb9/0x10c [<c1030030>] ? adjust_resource+0x63/0x81 [<c103533f>] ? mod_timer+0x6b/0x73 [<c1034be8>] run_timer_softirq+0x15e/0x201 [<c12b16c6>] ? dev_watchdog+0x0/0x10c [<c102f75e>] __do_softirq+0xaa/0x14e [<c102f828>] do_softirq+0x26/0x2b [<c102f925>] irq_exit+0x29/0x5c [<c10148eb>] smp_apic_timer_interrupt+0x67/0x75 [<c133e1e2>] apic_timer_interrupt+0x2a/0x30 [<c10083fe>] ? mwait_idle+0x5c/0x62 A[<c1001a34>] cpu_idle+0x44/0x60 [<c1329f50>] rest_init+0x6c/0x6e [<c14cb884>] start_kernel+0x2b1/0x2b6 [<c14cb0c9>] i386_start_kernel+0xc9/0xd0 ---[ end trace e074c399af7f99c9 ]--- e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX
Output of lspci -v :
01:01.0 Ethernet controller: Intel Corporation 82541PI Gigabit Ethernet Controller (rev 05) Subsystem: Intel Corporation PRO/1000 GT Desktop Adapter Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 22 Memory at ff8e0000 (32-bit, non-prefetchable) [size=128K] Memory at ff8c0000 (32-bit, non-prefetchable) [size=128K] I/O ports at bc00 [size=64] Expansion ROM at ff8a0000 [disabled] [size=128K] Capabilities: [dc] Power Management version 2 Capabilities: [e4] PCI-X non-bridge device Kernel driver in use: e1000