Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: Matlab monkey business

  • Subject: Re: Matlab monkey business
  • From: Roy Schestowitz <newsgroups@schestowitz.com>
  • Date: Tue, 07 Jun 2005 04:56:46 +0100
  • Newsgroups: uk.comp.os.linux
  • References: <slrnda8u32.8oq.frink@homer.cmp.liv.ac.uk>
  • User-agent: KNode/0.7.2
Doctor J. Frink wrote:

> Hullo,
> 
> We have a central server that people run programs on that is having some
> bother with (we think) Matlab. It appears to be a case of something
> using up all available memory. Then when other programs or services try
> to run they encounter VM errors such as
> 
> Jun  5 17:48:25 homer kernel: __alloc_pages: 0-order allocation failed
> (gfp=0x1d2/0)
> Jun  5 17:48:33 homer kernel: VM: killing process gkrellm
> 
> Jun  5 17:58:33 homer kernel: VM: killing process qmgr
> Jun  5 17:58:33 homer kernel: __alloc_pages: 0-order allocation failed
> (gfp=0x1d2/0)
> 
> And so on, which eventually totals the machine as far as connecting to
> it is concerned. This is on 2.4.25, 512M, 800M swap.
> 
> I have two questions:
> 
> 1) Has anyone else seen this sort of thing (particularly with Matlab)
> 
> 2) Is there a way of limiting processes to a certain amount (or
> percentage) of RAM usage?
> 
> Cheers,
>        Frink

To add my humble bit to this discussion, I have once run experimental MATLAB
stuff on our server in order to generate some on-line documentation. My
colleague informed me that MATLAB sessions remained alive (or something to
that effect) and drained memory. So apparently there is something rotten
about it.

Roy

-- 
Roy S. Schestowitz
http://Schestowitz.com

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index