[ale] OT: ethernet port issues

[Steve Litt]

Solomon Peachy via Ale said on Sat, 24 Jul 2021 14:40:31 -0400

>On Sat, Jul 24, 2021 at 02:27:26PM -0400, Boris Borisov via Ale wrote:
>>  PC looses connectivity with Cisco switch. When I go to site there
>> is no LED lit on PC ethernet port. Once I unplug and plug cable back
>> connection is restored.
>> 
>> Possible causes?  
>
>flaky port on switch, flaky port on PC, flaky cable?

Probably even more possible causes, but the preceding is a good start.
Now is the time to start diagnostic tests. IMHO easiest test is to
spray electronics lubricant on both ends of the cable and in the ports
of the switch and PC. Either you'll get lucky and it was just a matter
of contact oxidation, or you'll rule that out.

My next step would be to try different ports on the switch and see if
the symptom changes.

My next step would be to swap the cable with a known good cable ---
that's usually easy, even if instead of running it thru the wall you
just temporarily tape it to the floor.

>
>I don't know what sort of logging the cisco switch has, but if the PC
>is running Linux there might be something useful in the logs when the 
>connection drops.

If the switch doesn't have logging, you can make your own logging by
creating your own daemon, written as a shellscript, to ping -c1, a
known active computer on your LAN every 10 seconds. You write the
shellscript exactly as you'd write it if you were running it from the
command prompt. Runit, s6, OpenRC and systemd can all run a daemon that
doesn't background itself. If you're using sysvinit, either run/respawn
runit or s6 from /etc/inittab, or do it with cron, but I think you'd
be limited to 1 measurement per minute (please correct me if I'm
wrong). The runit init run script would look something like this:

=====================================
#!/bin/sh
exec /path/to/my/shellscript
=====================================

The shellscript itself would look like the following:

=====================================
#!/bin/sh
while /bin/true; do
   ping -c1  192.168.0.14 > /dev/null \
   || date  "+%Y-%m-%d@%H:%M:%S" > /var/log/mylog.log
done
=====================================

>
>Oh, other possiblities: adminsistrative action/policies on either 
>side, plus the ever possible spectre of plain ol' software bugs.

I'd run it overnight running a live CD with a distro as different from
the installed distro as possible, running the daemon from a command
prompt. In other words, swap operating systems.

Also, run it from the installed OS, freshly rebooted, with no
software running, overnight. That would tend to rule in or out software
you run on top of your installed OS.

SteveT

Steve Litt 
Spring 2021 featured book: Troubleshooting Techniques of the Successful
Technologist http://www.troubleshooters.com/techniques