On Thu, 14 Sep 2017 09:11:36 -0400, grg wrote: > On Wed, Sep 13, 2017 at 10:36:40PM -0400, Richard Pieri wrote: >> On 9/13/2017 10:13 PM, Robert Krawitz wrote: >> > This is 1000Base-T, with standard cat 5e cable. scp isn't much slower. >> >> You're using full-duplex with Cat 5e? You're off spec. And now I'm >> wondering if the data corruption problems you were having a few weeks >> ago were a consequence of it.
No, I'm quite certain they aren't. These problems exist with only one machine (including using the same cable and NIC port as on another machine that works fine), and I've seen them with loopback operation also. The nature of the failures -- aligned relative to a 64-byte boundary -- are also not what I would expect to see in the case of bad ethernet operation. I would also expect TCP checksumming to catch errors of this type. > Which spec are you referring to? Please cite your source. > > FWIW (some, but never the definitive answer) Wikipedia disagrees with you: > "Each 1000BASE-T network segment can be a maximum length of 100 meters (330 > feet), and must use Category 5 cable or better (including Cat 5e and Cat 6)." > https://en.wikipedia.org/wiki/Gigabit_Ethernet#1000BASE-T > So with 5e he's actually a grade above the minimum cat5. > (Note the top of that article says that full duplex is used exclusively, so > they're not talking about half duplex operation over cat5 or 5e or cat6.) -- Robert Krawitz <r...@alum.mit.edu> *** MIT Engineers A Proud Tradition http://mitathletics.com *** Member of the League for Programming Freedom -- http://ProgFree.org Project lead for Gutenprint -- http://gimp-print.sourceforge.net "Linux doesn't dictate how I work, I dictate how Linux works." --Eric Crampton _______________________________________________ Discuss mailing list Discuss@blu.org http://lists.blu.org/mailman/listinfo/discuss