Hello,
I have got Dell PowerEdge 750 with on-board em(4) network cards and I
use it as a router. The interface em0 stops forwarding traffic from time
to time. Resetting the interface with ifconfig down & up fixes the
problem.
em0@pci1:1:0: class=0x020000 card=0x01651028 chip=0x10758086 rev=0x00
hdr=0x00
vendor = 'Intel Corporation'
device = '82547EI Gigabit Ethernet Controller'
class = network
subclass = ethernet
em0: <Intel(R) PRO/1000 Network Connection Version - 6.2.9> port
0xece0-0xecff mem 0xfe2e0000-0xfe2fffff irq 18 at device 1.0 on pci1
em0: Reserved 0x20000 bytes for rid 0x10 type 3 at 0xfe2e0000
em0: Reserved 0x20 bytes for rid 0x18 type 4 at 0xece0
em0: bpf attached
em0: Ethernet address: 00:14:22:7a:1c:10
ioapic0: routing intpin 18 (PCI IRQ 18) to vector 49
em0: [MPSAFE]
I am running RELENG_6_2, with CARP. PF with traffic normalization is
active (scrub in all fragment reassemble). I have tried turning off
checksum offload. When the interface does not pass traffic it looks
normal in ifconfig(8). ARP does not work and nothing interesting get
written to the logs.
em0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
options=b<RXCSUM,TXCSUM,VLAN_MTU>
inet 192.168.1.8 netmask 0xfffffff0 broadcast 192.168.1.15
ether 00:14:22:7a:1c:10
media: Ethernet autoselect (1000baseTX <full-duplex>)
status: active
The machine is processing couple of megabits worth of traffic all the
time and it takes anywhere from an hour to more than a day to occur. I
had about 10 interface freezes on em0 interface but none on em1. When I
put additional network card into the machine (xl(4)) and use it together
with em1 it works without a problem.
The only difference between em0 and em1 there I can think of is that
there may be some out-of-band management functions in the em0 which
driver does not correctly turn of or something.
I have noticed a small update to the driver in -STABLE but I don't think
this could fix my issues. Due to the nature of the problem and use of
the machine I am afraid I can not do much to help narrow the problem
further down.
Thank you
Michal