[nos-bbs] The KPC3 TNC

[George [ham] VerDuin]

I believe I may have just seen the Achilles heel of the TNC.

The K8RRA node is a Fedora Linux host with jnos2.0f.  It drives a Kam 
KPC3P TNC and a 125Watt GE Delta 2M rig.

The station has been trouble-free for years.  A few days ago I chose to 
move jnos from a workstation host next to my remaining ham gear, to a 
server host next to other LAN gear with a UPS.  The target host is an 
IBM PL300 -- an old slow but under utilized http & file server.  Jnos 
has been operating there for some hours with a known RS-232 problem I 
was working on but had not yet fixed.

While I was reading email, I heard my station working a link.  The long 
packet from K8RRA was immediately answered by a remote short packet.  I 
thought it was pretty normal until the pattern became boring and 
non-changing.  When I tried to investigate by attaching to jnos I found 
the session had crashed -- a very unusual thing.  When I tried to attach 
to the TNC with minicom, I found the TNC to be blocking my attempt.  Now 
this is strange -- the TNC was still driving the same pattern of 
transmission without any link to a supervising computer.  The way I 
caused the TNC to stop was by killing the power.

Cycling power off/on did stop and reset the TNC from KISS to TERMINAL 
interface mode.  Yet this action gives me serious hesitation.  Presently 
I have little information about why jnos crashed, and exactly what the 
TNC was repeatedly transmitting.  The RS-232 fault may be the cause.

Has anyone had their own TNC act on it's own before?  Was power-off your 
only way to stop the TNC?  Are there any known ways to initiate a 
never-ending sequence of transmissions without the computer being able 
to stop it?

Thanks & 73!
Skip