Re: network problems on OSF1

From: Nelson H. F. Beebe <beebe_at_math.utah.edu>
Date: Wed, 1 Feb 95 13:10:45 MST

The output of "netstat -is" on alab02.math.utah.edu (uptime: 29 days)
is attached to the end of this message; it shows nothing like your
massive "65535 receive failures" problems (in fact, since 65535 == 2^16 - 1,
I suspect this is a 16-bit unsigned integer counter pegged at its maximum).

The one piece of history that may be useful is that a couple of years
ago, Physics had extremely sluggish network response between the RS/6000
workstations, far slower than comparable systems in Chemistry, CS, and
USI were seeing. The last time I spoke to Gene Loh about this (many months
ago), I recommended that Physics bring in some hardware monitoring equipment
to try to isolate the problems on the net; I never heard back about whether
they did this or not. You might talk to Gene Loh, Richard Price, and
Carleton Detar. That you are now seeing network problems with your brand
new DEC Alpha system certainly makes one suspect that the problems are
still there on the net, and are due to a bad connection, or faulty network
hardware, somewhere in the Physics Department.

netstat -is
 
ln0 Ethernet counters at Wed Feb 1 13:02:29 1995
 
        7528 seconds since last zeroed
   352466419 bytes received
   155107383 bytes sent
     1272538 data blocks received
      957599 data blocks sent
    48329556 multicast bytes received
      183150 multicast blocks received
      254705 multicast bytes sent
        4412 multicast blocks sent
       16886 blocks sent, initially deferred
         860 blocks sent, single collision
         984 blocks sent, multiple collisions
           5 send failures, reasons include:
                Excessive collisions
                Carrier check failed
           4 collision detect check failure
           0 receive failures
           0 unrecognized frame destination
           0 data overruns
           0 system buffer unavailable
           0 user buffer unavailable

Our DEC Alpha systems have no pfopen.c file in the file system;
/usr/examples/packetfilter contains only one file:

-rw-r--r-- 1 bin bin 4284 Aug 8 22:34 pfsamp.c

========================================================================
Nelson H. F. Beebe Tel: +1 801 581 5254
Center for Scientific Computing FAX: +1 801 581 4148
Department of Mathematics, 105 JWB Internet: beebe_at_math.utah.edu
University of Utah
Salt Lake City, UT 84112, USA
========================================================================
Received on Wed Feb 01 1995 - 15:11:30 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:45 NZDT