Hi All,
the issue of a single or multiple CEs is one we are facing also at ucl-
hep. I think some of the arguments in favor of a separate CE are the
ability to provide separate VO software areas. We have currently
penned down the option of a separate (test) CE for SL5 WNs, which will
imply a switchover at a later date. Obviously, a single CE with
appropriate sub-clusters defined is an attractive alternative in order
to streamline the transition. We would also be interested in the RALPP
files as example.
cheers,
Gianfranco
On 15 Sep 2009, at 09:09, Stephen Jones wrote:
> Chris, Stephen,
>
> We have used cpu scaling factors up to now, to allow
> our 1cpu32bit and 8cpu64bit nodes to co-exist. This will end when we
> go to SL5, as the 1cpu32bit nodes must remain on glite 3.1. We
> aren't too intimidated by the missing
> yaim, but we are curious about why LHCb request that SL4/32bit and
> SL5/64bit resources
> are not made available to them off the same CE. That is exactly what
> we ARE thinking of doing (!)
>
> It would be nice to have your RALPP files as examples - could you
> attach them?
>
> Cheers,
>
> Stephen Jones
>
> Brew, Chris (STFC,RAL,PPD) wrote:
>> As the original spark for this I guess I should chip in.
>>
>
> --
> Steve Jones [log in to unmask]
> System Administrator office: 220
> High Energy Physics Division tel (int): 42334
> Oliver Lodge Laboratory tel (ext): +44 (0)151 794 2334
> University of Liverpool http://www.liv.ac.uk/physics/hep/
--
Dr. Gianfranco Sciacca Tel: +44 (0)20 7679 3044
Dept of Physics and Astronomy Internal: 33044
University College London D15 - Physics Building
London WC1E 6BT
|