Plus, after Monday, we'll have an easy way of putting a cap of the number of nodes a large array job can consume.  Not suggesting we use it, but it's there if the scheduling policy doesn't win as Jan confidently predicts.   (I should be running some - smaller - jobs in parallel with yours from Tuesday, so we'll see if the scheduling policy works in the wild.)

Simon

On 25 Feb 2011, at 14:18, Jan Staunton wrote:

> Go for it, new scheduling policy should win :-)
> 
> Cheers
> 
> Jan
> 
> Sent from my mobile.
> 
> On 25 Feb 2011, at 14:03, David R White <drw@cs.york.ac.uk> wrote:
> 
>> Hi All,
>> 
>> I plan to start running a large number of small jobs.  Each job will take about a minute to execute, but there will be many instances of them - enough to fully occupy the grid for a few days in the absence of other jobs.  I plan to start these jobs today or Tuesday.
>> 
>> The grid scheduling mechanism should cope with this, but please let me know if you have any objections or requests to delay.  I can easily "throttle" grid usage.
>> 
>> I'm aware that the grid is going to be down on Monday.
>> 
>> Cheers,
>> 
>> David
>> -- 
>> Dr David R. White
>> Research Associate
>> Dept. of Computer Science
>> University of York,
>> Deramore Lane, YO10 5GH.
>> http://www.cs.york.ac.uk/~drw
>> 
> 


--
Simon Poulding

Lecturer in Enterprise Systems
Department of Computer Science
University of York
Deramore Lane
York YO10 5GH

Tel: +44 (0)1904 325165
Fax: +44 (0)1904 325599

Email: smp@cs.york.ac.uk
Web: www.cs.york.ac.uk/~smp

EMAIL DISCLAIMER: http://www.york.ac.uk/docs/disclaimer/email.htm