After a  storm of  CRC errors  from UHUPVM1 (RSCS  V2.3), I  removed that
server from the  backbone a few days  ago to see whether  the problem was
actually there (in which case errors should stop coming in) or elsewhere.
The errors first stopped flowing in, but  today a new storm of even worse
proportions originated  from TAMVM1 (RSCS V1!),  so I have removed  it as
well to see what happens.
 
This  is probably  due  to the  usual RSCS  bugs  in processing  multiple
dataset header  files. There are different  flavours of bugs for  RSCS V1
and V2.3, there are probably fixes  but I wouldn't know the APAR numbers.
RSCS V1 tends to  insert garbage in the file, while  buggy levels of V2.3
latter happily decide that the  transfer has successfully completed after
writing a couple  lines to the punch,  and move on to the  next file. For
instance, two  BITNET nodes have  been chronically unable to  receive any
file destined to both nodes for about  one and a half year now (both have
been  informed several  times  and neither  has time  to  deal with  this
problem).  As long  as this  only affects  end nodes  it really  isn't my
problem, but I have to make sure that the backbone works.
 
My plans are to  wait a few days and then try  putting UHUPVM1 back under
the assumption that the problem was at TAMVM1.
 
  Eric