[nSLUG] Re: Ubuntu crash with scp.

Mike Spencer mspencer at tallships.ca
Thu Nov 8 14:42:12 AST 2007

Donald Teed wrote:

> I'm not sure if you are talking about an actual lock up.
> The kernel is still working if you are able to login remotely.
> I've seen cases that appear to be frozen which eventually
> release, given 10 minutes or so.

I think you may be conflating David Payne's original post and my
tenuously related follow-up.  Since David has identified a hardware
problem, I'll carry on as if you were addressing my console lockup

> If you are killing X remotely and the keyboard still isn't
> responsive that is odd.

I'm not doing that.

    Under some unspecified conditions a program widely known to have
    bugs (Netscape 4.76) appears to provoke a lockup.  No screen activity
    and no response to mouse & keyboard.

    A login to the apparently frozen machine via telnet proceeds
    normally.  Doing ps shows that X is *already* missing from the
    process list.  So is Netscape.  I infer that Netscape 4.76 and
    X have both died, provoked by some buggy action of Netscape.

> Is it a PS/2 or USB keyboard?


> Is there anything consuming CPU cycles when it happens?

IIRC, nothing out of the ho-hum ordinary, no.  Because such an episode
begins with a drastic slowdown of mouse/cursor response, I had assumed
that someting was totally maxing out the CPU.  That proves not to have
been the case.  Something loaded up the CPU and then died, leaving the
console system wedged.

> Do you use suspend to disk/RAM?  


> Do you use proprietary modules or wrapper modules like ndiswrapper?

No. (Former posts to the effect that I use ndiswrapper relate to
     another machine.)

David has found a suspect hardware element.  I'm reasonably  sure my
problem is software or perhaps inadequacy (not defects) in hardware
because the same problem has arisen on a Pentium, PII and PIII CPUs in
different machines.

My question is not how to avoid this (upgrade browser inter alia,
etc. etc.)  but how to recover control of the console(s) gracefully
without a reboot when it happens.

Since it only happens rarely in association (I infer) with HTML markup
that is too complex (or possibly broken) for the browser to handle,
I'll have to wait for the next occurrence before I can futz with it

- Mike

Michael Spencer                  Nova Scotia, Canada       .~. 
mspencer at tallships.ca                                     /( )\
http://home.tallships.ca/mspencer/                        ^^-^^

More information about the nSLUG mailing list