Re: Debian at GSI
Hi Christopher,
That is indeed a great story of success for the Debian project. Thank
you for sharing! I wonder if there is any representative page at gsi.de
which would mirror your statement?
Googling for debian on gsi.de comes back with lots of links, including
such as
http://theory.gsi.de/~tneff/theory/news.html
Upgrade to Debian 2.0/Kernel 2.1.125 Thomas Neff: October 16, 1998
providing pieces of evidence to your description ;) but I hoped there
could be some good summary somewhere.
Cheers,
Yarik
On Mon, 21 May 2012, Christopher Huhn wrote:
> Dear debian-science community,
> I'm sorry for this rather long post, but in the spirit of the "Debian at ESRF" post I'd like to give some information about the Debian-based computing infrastructure at GSI [1], a German non-profit research lab that runs a heavy ion accelarator and is part of the Helmholtz association [2].
> We are running a more-or-less Debian-only Linux "farm" with more than 1000 nodes. We used the Debian distro for our Linux boxes since the last millennium (maybe 1997?).
> We recently put on a 10 000 core compute cluster and 1.5 PB Lustre
> storage with Infiniband-only networking running Debian Squeeze and managed by GridEngine - on top of our existing ~ 5000 CPU core cluster with more than 2 PB Lustre storage.
> Apart from our compute cluster we also provide some hundred desktops,
> servers for data acquisition, central IT services (DNS, DHCP, Radius, MTAs, ...) and collaboration web services (Wikis, Subversion, ...) with Debian.
> Our accelerator controls unfortunately moves from VMS to Red Hat for
> the shiny care-less all-inclusive support - but at least it's Linux ...
> Our Debian installation is actually not particularly "scientific". The big
> packages like ROOT and Cernlib are installed from source as our
> scientists need dozens of customized versions installed at the same
> time and don't really honor (maybe understand?) the merits of proper package
> management.
> Nevertheless we profit a lot from the 2000+ packages we pull from the
> standard Debian repositories for our default desktop installation.
> In the high energy physics community we are quite alien as everybody
> else uses "Scientific Linux" that has been created by CERN and Fermilab
> when Red Hat changed its licensing. At that time we already were happy
> with Debian and promoted it as an alternative but unfortunately the big
> sites were too scared to make that move (different packaging system, no
> commercial support, the usual complaints).
> BTW: We currently have an open position in our HPC group [3]. The job offer is in German only unfortunately m(. Basically the job description covers all our tasks from high-level configuration management with Chef to fiddling with broken hardware in case of an urgent emergency (this became more and more seldom during the years). The new position will probably focus a lot on monitoring, which did not grow fast enough with the growth of the farm during the last years.
> Please contact me if you have any questions and don't hesitate to send any applications in English.
> Cheers,
> Christopher
> [1]http://gsi.de/
> [2] http://www.helmholtz.de/en
> [3] http://gsi.de/informationen/internal/vw/pa/1920-12.50.html
--
Yaroslav O. Halchenko
Postdoctoral Fellow, Department of Psychological and Brain Sciences
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419
WWW: http://www.linkedin.com/in/yarik
Reply to: