Conference noted::etherworks

1317.0. "Framing errors and runts following 1st Ip address!" by CSC32::A_LICAUSE () Tue Jun 03 1997 12:04

    Field engineer is working with customers new 8400 running V4.0b of
    dunix and 4G RAM.  It has two DE500-XA's both running 100Mbps 1/2 duplex 
    to a Cisco hub.  They are seeing sessions lockup for seconds to minutes.
    Maxusers = 128.   This is an Oracle DB server.  Both XA's at rev 1.2
    
    They are not using NFS.  They are using Advfs.
    
    They are seeing framing errors and runt packets, but only on the 1st IP
    address of 10.1.10.10 (2x255 mask).  The second address 10.1.10.11 is not
    exhibiting the errors.  When they swap the addresses, the errors follow
    the first address.  They are testing from an Alpha station 1000 with a
    DE500-XA.
    
    Testing involves using ftp to copy very large Oracle database files.
    
    TU0 is on slot 11 of bus 0 and Tu1 is on slot 10 of bus 1
    
    The 10BaseT cables are made in house, but are only about 15ft in
    length.
    
    They have tried swapping the devices between busses and slots on busses.
    They do have some kzpsa disk controllers on the same busses which do
    have priority.
    
    Don't know what to tell him other than to try some AA models to see if
    they see any difference. 
    
    Anyone have any ideas or suggestions?