SUMMARY: Ethernet connection problems

From: Kamal Kantawala (
Date: Wed Apr 12 1995 - 09:56:51 CDT

My original question was:

> Subject: Ethernet connection problems
| Sun Managers,
| One of the machines (A) (a Sparc 1 with SunOS 4.1.3_U1)
| in my department had a faulty ethernet port
| and was dropping more that 50 % packets. Thus it
| had a lot of trouble contacting our NIS/NFS servers.
| I took another diskless and dataless Sparc1 (B) and moved the disks from
| A to B. It did not boot. I realized I had to change the
| the appropriate entries in the eeprom to make it boot locally
| from the hard disk. Now it has no problem booting.
| Now the ethernet interface on B is working fine but it still cannot
| contact the NIS/NFS servers. My network is divided into two floors.
| I am able to ping it perfectly from all machines on its floor (2). But
| from the the floor (1) on which my NIS/NFS servers are, I cannot ping
| it at all.
| etherfind on floor 1 sees no packets from B. etherfind on
| floor 2 sees all packets from B. When I ping from B to floor 1, etherfind
| on floor 2 sees replies coming from floor 1 but B does not seem to accept
| them.
| There is a bridge between floor 1 and floor 2. I suspect it has something
| to do with this problem. But I am not sure. I power cycled the bridge
| with no success. I am at my wits end.
| Any help would be appreciated.
| Kamal Kantawala
| System Administrator, MIPL
| SUNY, Stony Brook.
| (
>-- End of excerpt from Kamal Kantawala

I got a lot of replies, all of them with very helpful suggestions.
Most suggested that the bridge may be filtering out packets from
replaced Sparc 1 (B). Some suggested updating the arp cache.
Unfortunately the problem was elsewhere. For some reason, the
replacement Sparc 1 (B) could not support the full length AUI cable
which connected it to the multi-port transceiver. That was the
reason it could not talk to any machines, except the ones on its
own floor (2). The signal got too attenuated. For the same reason
it could not see any packets coming from the other floor (1).
By running etherfind on B and machines on the floor 1 and 2 and
comparing the results, I suspected that the length of the cable
was creating problems. I cut the slack from the cable which shortened
it by about 25 feet. The problem was solved.

Thanks to all who replied. A partial list follows: (Phil Gaskell)
 James Buszard-Welcher <jwelcher@systems.DHL.COM>
 grevemes@VTC.TACOM.Army.Mil (Steven Grevemeyer)
 gibian%typhoon@stars1.HANSCOM.AF.MIL (Marc Gibian)
 craig@rufus.state.COM.AU (Craig Johnson)

Kamal Kantawala
System Administrator, MIPL
SUNY, Stony Brook.

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:21 CDT