MathGroup Archive 2009

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Inappropriate kernel limits!

  • To: mathgroup at smc.vnet.net
  • Subject: [mg96304] Re: [mg96268] Inappropriate kernel limits!
  • From: John Fultz <jfultz at wolfram.com>
  • Date: Wed, 11 Feb 2009 05:20:35 -0500 (EST)
  • Reply-to: jfultz at wolfram.com

On Tue, 10 Feb 2009 05:54:39 -0500 (EST), Scott Morrison wrote:
> I'm running a Mathematica program that is crashing badly (more on that
> later); I have to kill the processes (on OSX) by hand using Activity
> Monitor or the terminal, as the frontend becomes completely
> unresponsive.
>
> Afterwards, however, when I try to restart Mathematica (or the old
> copy of v6 I have around), I get the message "The 2-user limit on this
> license has been reached." I've carefully checked through the running
> processes and made sure there are no lurking kernels. Rebooting the
> computer solves the problem, but this is extremely unacceptable!
>
> Can someone tell me how to fix this problem? I'm really annoyed.
>
> thanks,
> Scott Morrison

If the processes are completely terminated, then they should completely free=
 the 
resources used for process counting.  However, if any vestige of the process=
 
remains, such as if the process were to become a zombie process, then that=
 can't 
be guaranteed.  This is my best guess as to what's happening on your=
 system.

If that's the case, then it's probable that using Terminal.app to kill the 
application from the Unix shell is the only option that will work (short of=
 
rebooting, of course).  E.g., something like...

ps -efwww | grep Math
kill -9 pid

where pid is the number in the second column of the ps output.

As to why this might be happening, I couldn't comment without understanding=
 
what's leading to the crashes you're experiencing.  I can say that I haven't=
 
seen any common bugs which leave this kind of process debris for many, many=
 
years now, but there are quite possibly uncommon bugs that I'm unaware of.

Sincerely,
 
John Fultz
jfultz at wolfram.com
User Interface Group
Wolfram Research, Inc.




  • Prev by Date: Re: Data interpolation
  • Next by Date: Re: Manipulating list of functions
  • Previous by thread: Re: Inappropriate kernel limits!
  • Next by thread: Manipulating list of functions