Hi Geoff,
some SuperMicro systems (and yours as well) have an IPMI interface running via the 1st onboard NIC, which means IPMI shares the phys. NIC with the typically LAN configuration on OS
level while the IPMI interface is configured within the Bios.
Anyway IPMI is great for remote tasks and we also use dedicated Intel NICs in
our SuperMicros to free up the onboad NIC for IPMI only tasks.
Maybe this link is interesting to understand the 2 MAC addresses on 1 NIC for
IPMI (post 2):
*
http://serverfault.com/questions/259792/how-does-ipmi-sideband-share-the-ethernet-port-with-the-host
Kind regards, Tim
Geoff Nordli wrote:
On 15-01-15 11:13 PM, Chris Buechler wrote:
On Thu, Jan 15, 2015 at 6:56 PM, Geoff Nordli <[email protected]> wrote:
Hi.
We have a Superserver 5015A-EHF-D525
(http://www.supermicro.com/products/system/1u/5015/sys-5015a-ehf-d525.cfm)
running pfsense 2.1.5-RELEASE (amd64) with 2GB of RAM.
Which has this board in it:
http://www.supermicro.com/products/motherboard/ATOM/ICH9/X7SPE-HF-D525.cfm
In this chassis we also have a 4 port Intel NIC which shows up as igb
interfaces.
We were experiencing substantial packet loss when using the em interfaces,
but since we switched over to the igb interfaces things have been good.
I have both Hardware TCP Segmentation Offloading and Hardware Large Receive
Offloading disabled.
This is not a heavily used firewall.
Anyone else experiencing packet loss on the em interfaces. Are there any
other settings I should look at?
My best guess with those symptoms, if it isn't a hardware problem, is
one or more of the affected NICs ending up on the same IRQ as a USB
controller or something else that's causing issues. Most of the time
that's no big deal, on occasion with certain systems with several NICs
it can cause packet loss or performance issues. If that's the case,
may find a BIOS update that fixes it, or may be able to muck with BIOS
settings to make it go away.
thanks Chris.
Anything visible I can see on the local machine -- without going to bios.
Doing a vmstat -i shows the msi interrupts being used for those controllers.
interrupt total rate
irq18: ehci0 uhci5 2 0
irq19: uhci2 uhci4+ 2649947 6
cpu0: timer 761654301 1992
irq256: igb0:que 0 9765149 25
irq257: igb0:que 1 2747978 7
irq258: igb0:que 2 2562344 6
irq259: igb0:que 3 2387427 6
irq260: igb0:link 2 0
irq261: igb1:que 0 13670253 35
irq262: igb1:que 1 7013567 18
irq263: igb1:que 2 7502120 19
irq264: igb1:que 3 4520889 11
irq265: igb1:link 2 0
irq266: igb2:que 0 17501710 45
irq267: igb2:que 1 8228974 21
irq268: igb2:que 2 6685269 17
irq269: igb2:que 3 5603615 14
irq270: igb2:link 2 0
irq279: em1:rx 0 233940 0
irq280: em1:tx 0 215119 0
irq281: em1:link 15 0
cpu1: timer 761634299 1991
cpu2: timer 761634319 1991
cpu3: timer 761634321 1991
Total 3137845564 8206
I can definitely try a bios update next time I am close to the machine.
Geoff
_______________________________________________
pfSense mailing list
https://lists.pfsense.org/mailman/listinfo/list
Support the project with Gold! https://pfsense.org/gold
_______________________________________________
pfSense mailing list
https://lists.pfsense.org/mailman/listinfo/list
Support the project with Gold! https://pfsense.org/gold
