[Beowulf] Queue Systems

[Tim Cutts]

On 6 Sep 2007, at 10:30 am, andrew holway wrote:

> Hi,
>
> We are trying to work out the differences between these queue systems.
>
> Can anyone shed any light? Pros and Cons...
>
> SGE, Torque (with Maui), PBSPro and LSF

LSF is very good, but costs $$$.  Platform's support is, in my  
experience, extremely good.  The last time I reported a bug they had  
a fix to me inside two hours, which is pretty good going.  What's  
more, although we're a big customer now, my experience of them as a  
small customer in the past, with less than ten nodes, was just as good.

I hear consistently good things about SGE but don't have much  
personal experience of it.  Certainly in the bioclusters area, the  
general consensus seems to be that SGE does the batch scheduling job  
perfectly well, and LSF adds a few bells and whistles over the top  
which are not necessarily strictly related to batch scheduling, but  
if you need them are an absolute winner.  For instance, the LSF 7  
infrastructure is also a high-ability framework which you can use for  
creating failover applications and other services.  The LSF 7 batch  
scheduler is just a client of this more general underlying framework  
(which has the unfortunate acronym of EGO - Enterprise Grid  
Orchestrator).  Of course, you can do that failover stuff with free  
software too, using things like heartbeat.

Tim


-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.