ethernet bonding: are my twon NICs on the same or different busses

I found this quote in the newsgroup while I was reading up on ethernet
bonding:

"Most of the servers I am familar with two ethernet ports are designed
to have one used as a backup, as they are both connected to the same
bus. None of them had sufficent bus bandwith to actually drive two
gigabit ethernets anywhere near full capacity. Indeed, depending on
what else was on the bus the performance may be less than one gigabit
interface. " (Frank Stutzman )

What's the way to check if my twin NICs are on the same / different busses?

--
Rahul

Rahul wrote:
> I found this quote in the newsgroup while I was reading up on ethernet
> bonding:
>
> "Most of the servers I am familar with two ethernet ports are designed
> to have one used as a backup, as they are both connected to the same
> bus. None of them had sufficent bus bandwith to actually drive two
> gigabit ethernets anywhere near full capacity. Indeed, depending on
> what else was on the bus the performance may be less than one gigabit
> interface. " (Frank Stutzman )
>
> What's the way to check if my twin NICs are on the same / different busses?
>
You might run

lspci -v

as root. Here is mine for my 3 NIC's.

03:02.0 Ethernet controller: Intel Corporation 82546EB Gigabit Ethernet
Controller (Copper) (rev 01)
Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 50
Memory at fc200000 (64-bit, non-prefetchable) [size=128K]
I/O ports at 3000 [size=64]
Capabilities: [dc] Power Management version 2
Capabilities: [e4] PCI-X non-bridge device
Capabilities: [f0] Message Signalled Interrupts: 64bit+ Queue=0/0
Enable-

03:02.1 Ethernet controller: Intel Corporation 82546EB Gigabit Ethernet
Controller (Copper) (rev 01)
Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 58
Memory at fc220000 (64-bit, non-prefetchable) [size=128K]
I/O ports at 3040 [size=64]
Capabilities: [dc] Power Management version 2
Capabilities: [e4] PCI-X non-bridge device
Capabilities: [f0] Message Signalled Interrupts: 64bit+ Queue=0/0
Enable-

05:01.0 Ethernet controller: Intel Corporation 82557/8/9/0/1 Ethernet Pro
100 (rev 08)
Subsystem: Intel Corporation EtherExpress PRO/100+ Management Adapter
Flags: bus master, medium devsel, latency 66, IRQ 66
Memory at fc400000 (32-bit, non-prefetchable) [size=4K]
I/O ports at 4000 [size=64]
Memory at fc500000 (32-bit, non-prefetchable) [size=1M]
[virtual] Expansion ROM at f8800000 [disabled] [size=1M]
Capabilities: [dc] Power Management version 2

The first two are on PCI-X bus 03 (a two-channel chip) and the last one is
on PCI-X bus 05 (a single card with only one network interface).

--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ PGP-Key: 9A2FC99A Registered Machine 241939.
/( )\ Shrewsbury, New Jersey http://counter.li.org
^^-^^ 12:20:01 up 15 days, 18:26, 4 users, load average: 4.79, 4.63, 4.58

Jean-David Beyer <(E-Mail Removed)> wrote in
news:sQBrk.473$Ro1.459@trnddc04:

>>
>>
> You might run
>
> lspci -v
>
> as root. Here is mine for my 3 NIC's.
>
> The first two are on PCI-X bus 03 (a two-channel chip) and the last
> one is on PCI-X bus 05 (a single card with only one network
> interface).
>

Thanks Jean! I think I'll run that and figure things out.


--
Rahul

In comp.os.linux.networking Rahul <(E-Mail Removed)> wrote:
> I found this quote in the newsgroup while I was reading up on ethernet
> bonding:

> "Most of the servers I am familar with two ethernet ports are designed
> to have one used as a backup, as they are both connected to the same
> bus. None of them had sufficent bus bandwith to actually drive two
> gigabit ethernets anywhere near full capacity. Indeed, depending on
> what else was on the bus the performance may be less than one gigabit
> interface. " (Frank Stutzman )

Someone else has already pointed-out the lspci stuff so instead I'll
say:

At one point that assertion was probably true, however I suspect that
if that assertion has not already aged beyond accuracy for some
definition of "current hardware," it will shortly. Virtually every
dual-port GbE interface with which I am familiar has a bus interface
capable of driving both ports simultaneously at or near to capacity.

Then, perhaps I live in a different context from Frank Stutzman.

rick jones
--
a wide gulf separates "what if" from "if only"
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...