[nos-bbs] Ubuntu 10.4 Server and JNOS anyone?

[George [ham] VerDuin]

On 06/21/2010 08:41 PM, Jim Smith wrote:
>
>
>   How long after the path gets broken does it occur? 
Many hours / my node runs 24/7 very unattended so I have no way to hang 
a time on it and I have not found any tracks left behind in trace or 
logs.  It is also rare with an easy fix so I remain disinterested in 
reproducing the fault and killing the bug [if there is one].   On this 
node, AX.25 mail forwarding is defined but the server is not started 
[usually], light traffic area, email over tcp/ip is the major use.  The 
software is very reliable as long as there is no fault in the hardware.

> I have nothing attached to the serial port yet. I'm wondering if that 
> is the problem. I would hope that things would be written such that if 
> a port became non-responsive the system would not chew up CPU.
Never ran w/o I/O here, attachments have always been between 1 and 3.  
I've never pursued the cause either, the suspect for me has been piled 
up TCP frames that have no place to go -- but remember suspicion has 
very little value.   The gdb debugger might be your friend, interupt and 
backtrace a number of times and maybe a pattern will show up?

>
> I wish I knew C better, I'd go trawling through the code to see...
You don't need c to isolate the high CPU code, just a good bit of 
diligence and luck.  I can't speak for the jnos community, but it MAY be 
interested in your findings because it represents unusual behavior.  
Fixing it is another matter...

>
> Cheers!
>
> Jim N8AVX

73
Skip