HP OpenVMS Systems

TCP/IP (UCX) packet fragmentation, connection rese

» close window

The Question is:

 
We have currently migrated to UCX from Multinet.  Since that time we have
been
having troubles with ftp file transfers from the Alphas to
our other machines.  We are getting "connection reset by peer" on the
server side with no errors on the VMS side.  When we catch this as it
is happening we find that the file transmission just stops.
It finaly fails after the client does a time out.  When we put
a sniffer in, it is as though the client is trying to ack packets and
the Alpah is not seeing the ack and just keeps transmitting
the same ip packet over and over again.
 
We are having no other FTP issues anywhere else in our network. We
have installed many different patches on the machines that the files
are being transfered too.
 
We are running Digital TCP/IP services OpenVMS Alpha Version V4.2 - ECO 1 on
a Alpha Server 8400 Model
 
The one major difference that we found is that the UCX stack seems to
fragment the packets where the multinet did not.
 
It should be noted that we are running from FDDI to ethernet to FDDI.
 
Any ideas?
 
Andy
We do have machines running Multinet and we do not find this issue.
It should also be noted that
 
 

The Answer is :

 
  Please contact the customer support center for assistance.
 
  It is not clear if you are encountering problems with the MultiNet
  package or with the TCP/IP Services for OpenVMS (UCX) package.  If
  you are running UCX V4.2 and having problems with it, then please
  load the most current ECO kit -- at present, this would likely be
  ECO 2 or later.  The customer support center can work with you to
  resolve this problem.
 
  If you are having problems with MultiNet, please contact MultiNet
  customer support organization directly.

  

     

     
answer written or last revised on ( 10-NOV-1998 )
     
» close window