<div dir="ltr"><div><div><div>Hi Xavier,<br><br></div> If you are running metagenomics with Qiime version 1.9 adds a new script called "start parallel jobs with slurm". I'm not familiar with its inner workings since I'm still using the default Qiime 1.8 in Bio-linux 8, but it looks promising. <br><br><a href="http://qiime.org/scripts/start_parallel_jobs_slurm.html">http://qiime.org/scripts/start_parallel_jobs_slurm.html</a><br><br></div>Best Regards,<br></div>- James<br><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Feb 5, 2015 at 3:18 AM, Tim Booth <span dir="ltr"><<a href="mailto:tbooth@ceh.ac.uk" target="_blank">tbooth@ceh.ac.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Xavi,<br>
<br>
To be honest I don't really know. Hopefully someone on the list will<br>
have some practical experience to share. We use Torqe on a cluster here<br>
at CEH but it's not actually running Bio-Linux and I had no part in<br>
setting it up so I don't know how easy it is to configure.<br>
<br>
I have experience with Condor and I don't think it is what you need at<br>
all. SGE used to be good but Oracle have not looked after it well.<br>
<br>
My guess would be that Slurm is the best answer for you. The latest<br>
version of QIIME recommends it, which is a good sign, and I'm told that<br>
set-up for a small cluster is "very simple". You might well have issues<br>
with running a cluster of Bio-Linux 7 and 8 machines, though, because<br>
the packaged versions of Slurm and of the bioinformatics software will<br>
be different on the different nodes. I'd strongly recommend upgrading<br>
everything to the same version.<br>
<br>
Also, you really should set up a fast shared NFS drive that all the<br>
machines can see. Running a cluster without a shared filesystem is<br>
possible but for most bioinformatics tasks it's highly impractical.<br>
<br>
Cheers,<br>
<br>
TIM<br>
<div class="HOEnZb"><div class="h5"><br>
<br>
<br>
<br>
On Thu, 2015-02-05 at 07:16 +0000, Xavier Martínez Serrano wrote:<br>
> Hi,<br>
><br>
> We have 4 computers with ~20CPU each in a hospital research group and<br>
> we would like to set a cluster with them. Which is the best suitable<br>
> job scheduler for us if we are running Biolinux7 in three of them and<br>
> Biolinux8 in the other?<br>
><br>
> We have been looking at common job schedulers like: Slurm, SGE,<br>
> Condor, Torque, ... but we wonder which best suits for our small<br>
> resources, to make it as easy as possible.<br>
><br>
> What can you recommend us?<br>
><br>
> Many thanks in advance,<br>
> Xavi<br>
><br>
<br>
</div></div><span class="HOEnZb"><font color="#888888">--<br>
Tim Booth <<a href="mailto:tbooth@ceh.ac.uk">tbooth@ceh.ac.uk</a>><br>
NERC Environmental Bioinformatics Centre<br>
<br>
Centre for Ecology and Hydrology<br>
Maclean Bldg, Benson Lane<br>
Crowmarsh Gifford<br>
Wallingford, England<br>
OX10 8BB<br>
<br>
<a href="http://environmentalomics.org/bio-linux" target="_blank">http://environmentalomics.org/bio-linux</a><br>
<a href="tel:%2B44%201491%2069%202297" value="+441491692297">+44 1491 69 2297</a><br>
</font></span><div class="HOEnZb"><div class="h5"><br>
_______________________________________________<br>
Bio-Linux mailing list<br>
<a href="mailto:Bio-Linux@nebclists.nerc.ac.uk">Bio-Linux@nebclists.nerc.ac.uk</a><br>
<a href="http://nebclists.nerc.ac.uk/mailman/listinfo/bio-linux" target="_blank">http://nebclists.nerc.ac.uk/mailman/listinfo/bio-linux</a><br>
</div></div></blockquote></div><br><br clear="all"><br></div></div>