[Sis-SCPS-INTEREST] tx collisions: duplex state resolved, collisions are not.

Thu May 4 18:26:24 EDT 2006

You will never get full-duplex as hubs are not full duplex, collision
are the norm in hubs too, replace the hub with a switch and you will
solve your problem (if half-duplex and collisions are a problem for
you).  As for sniffing traffic you can always use tcpdump or ethereal
so the hub is needless.

-otto

On Thu, May 04, 2006 at 12:18:58PM -0700, Dan Miller wrote:
> <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
> <html>
> <head>
>   <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
> </head>
> <body bgcolor="#c0c0c0" text="#000000">
> <font size="+1"><font face="Times New Roman">Okay, I determined the
> state of the linux eth interface by the simple expedient of unplugging
> and re-plugging the connector!!&nbsp; It displays disconnect/connect
> messages in syslog, and thereby confirmed that it was 100Mbps,
> half-duplex.&nbsp; <br>
> I then went to my Windows machine and changed the eth interface from
> auto-negotiate to 100Mb/half-duplex.&nbsp; <br>
> <br>
> ftp transfers *still* get the same number of collisions.<br>
> <br>
> So now I'm pretty sure duplexity (if that's a word) is not the issue.<br>
> <br>
> What else could cause collision??&nbsp; Again, keep in mind that if I merely
> kill "gateway" and do another transfer, no collisions occur, and the
> transfer is proceeding at a higher data rate (tho nowhere near
> 100Mbit/sec).<br>
> <br>
> Dan Miller<br>
> <br>
> </font></font>
> </body>
> </html>
> 

> _______________________________________________
> Sis-SCPS-INTEREST mailing list
> Sis-SCPS-INTEREST at mailman.ccsds.org
> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-scps-interest