[nos-bbs] bizarre - power spike, tnc kiss frame corruption, jnos crashes

[Kerry Smith]

Hey Maiko,

Now that the hair is not straight up like your hanging onto 
a static ball, here is something else about the problem you
found.

I found someone else having the same problem as me after
digging into the past messages from this list.

My system was running great.  Moved everything to my new
place and in the process, could not move all the antennas.
Thus, the 4 RF ports on the computer got kicked back to
one RF port with the other three still there.  I just
unplugged the TNC's from the ports that were not being used
and ran the system with one RF port.  What I came to find was
you could telnet into the system from another computer,
and do almost anything you wanted.  BUT, if you attempted
to gain access locally, it would happen, but it would take
you forever in a day to get logged in.

What I found was the CPU running at Max.  When I did a TOP,
I found jnos sitting on top of the stack, using up 98% of the
CPU.  A quick reboot, and all is fine.  Give it a couple
hours, and it will go back to the top.

The solution was to detach the unused ports after the
system starts.  I found and so did the person having the
same problem before me, that jnos is not very happy when
it attempts to send data out it's serial port, and can't
talk to the TNC at the other end.  I used the Detach idea
but just having the tnc at the other end turned on, with
the radio turned off would have done the same thing in 
correcting the problem.

I'm saying this becuase of the problem you found with the
header corruption.

Kerry


--- On Fri, 9/10/10, Maiko Langelaar <maiko at pcs.mb.ca> wrote:

> From: Maiko Langelaar <maiko at pcs.mb.ca>
> Subject: [nos-bbs] bizarre - power spike, tnc kiss frame corruption, jnos crashes
> To: "NOS BBS Mailing List" <nos-bbs at tapr.org>
> Date: Friday, September 10, 2010, 11:20 PM
> This is just too bizarre, but rather
> interesting in the end ...
> 
> JNOS is humming along nicely, KISS tnc working fine.
> 
> I'm doing some rewiring in the basement, accidently touch
> BLACK to
> WHITE and BANG, big flash, breaker on panel trips like it
> should,
> my kid is hanging on the ceiling tiles with white knuckles
> :)
> 
> Looking at my JNOS logs a bit later, and I'm noticing JNOS
> crashes
> each time something comes in on the KISS tnc - most bizarre
> thing.
> 
> After some debugging and head scratching, decide to power
> off the tnc, then power it on again (reset it). Everything
> is back to normal
> all of a sudden. My JNOS system is on a different breaker
> than the
> one that blew but the same panel of course.
> 
> After looking at the code, it would appear that if the
> *PORT* of
> the incoming KISS frame is *corrupt* or not valid, JNOS
> does not
> deal very well with that, resulting in constant crashing. I
> WILL
> fix that, but I thought I would share this with others.
> 
> Guess the SPIKE or SURGE or whatever brought on by my
> little wiring blunder did a temporary number on the TNC,
> corrupting the header of
> the kiss frames it was sending to JNOS. Thankfully
> resetting the TNC
> has fixed the problem (this time anyways).
> 
> Just wierd !
> 
> Maiko Langelaar / VE4KLM
> 
> 
> _______________________________________________
> nos-bbs mailing list
> nos-bbs at tapr.org
> https://www.tapr.org/cgi-bin/mailman/listinfo/nos-bbs
>