On Feb 8, 2011, at 11:19 AM, Peter Kjellström wrote: > On Tuesday, February 08, 2011 04:52:27 pm Johnny H wrote: >> Dear all, >> Thank you very much for your answers; am a little overwhelmed as I >> did >> not expect so many so fast. I will take a look at your suggestions >> and >> sites. >> >> The cluster is medium performance for something called >> bioinformatics; >> we are looking at next generation sequence data. >> >> I know it had Centos installed and was using Torque/PBS for running >> cluster jobs. This was set up by a company and we want to install >> fresh. > > What you describe here is quite clearly a compute cluster, HPC, or > whatever > you want to call it. > > Maybe you should go have a look at a packaged solution like rocks > (www.rocksclusters.org). If you don't have much experience I'd > recommend > against building something yourself. With rocks you'll also get the > benefit of > prepackaged bio-informatics stuff (rocks has a large community). > > /Peter > >> If I get anywhere without a sys admin, I will post (maybe another >> question). >> >> Cheers, > <signature.asc><ATT00001.txt> I second the suggestion of ROCKS. It centralized much of the administration of the cluster including setting up jobs resource managers, has prebuilt software including some for bioinformatics (and it's built on CentOS). There an active and helpful mailing list. Tony Schreiner