Re: Strange things on GBit / 1000->100 / net.inet.tcp.inflight.*

Raphael H. Becker wrote:
> On Fri, Sep 17, 2004 at 11:30:31PM +0200, Andre Oppermann wrote:
> 
> rebooted the "source"-machine to get a defined state. 
> Did reboot before the last test, too.
> 
> 
>>And could you do this on the 5.3 machine:
>>  # sysctl net.inet.tcp.rfc3042=0
> 
> 
> net.inet.tcp.rfc3042: 1 -> 0
> 
> wget on target:
> Length: 1,160,773,632 (unauthoritative)
> 
>  1% [          ] 12,967,976   325.72K/s    ETA 57:21^C

I've somewhat rebuild the test setup with my workstation dual Opteron
running 6-current and a 4.10 VIA C3-800 on the other side.  A Cisco 2950
in between.  On the GigE side I have a bge interface and fxp on the C3.

When doing an ftp of the second FreeBSD 5.3-BETA4-i386-disc2.iso (~280MB)
from the Opteron to the C3 I get a nice transfer speed of about 10.3MB/s.
Which is about as much as one can get considering the little C3 and it's
poor IDE disk.

I've tried ftp the 'active' and 'passive' mode way to rule out that the
direction of TCP connection setup could play a role.  It doesn't.  Same
speed both times.

Second try with a Netgear FSM726S.  Full speed again.  No problems.

I'm unable to reproduce your problem.

Usually in these cases there is an ethernet duplex mismatch somewhere.
Can set the network cards and the switches to fixed 1000 or 100 full-duplex?

-- 
Andre