Hi John,
> though. It will be a while before Jeremy and the T2 coords get round to
> checking VO support at all sites, although I will be happy to be proved
> wrong. More important is the experiment software being installed (and
> tested).
IMO sites can respond only for the environment (LCG+?) installation not
for the experiments software and definitely not for its testing which is
up to the experiments. Even knowing what software tag is published, as in
the case of LHC experiments, what do we (deployment) know about what that
tag is for, what problems it has, if it is too old or if it is yes old but
still used/needed.... and more importantly do we (deployment) have to know
it? I think it is more sensible if experiments report this information not
us.
cheers
alessandra
On Sat, 2 Oct 2004, John Gordon wrote:
> Steve, that bit of my suggestion was trivial for the LHC experiments as
> they are all included in the LCG distribution so it should just happen. I
> still think the experiment people in the UK should take some initiative
>
> John
>
> On Fri, 1 Oct 2004, Steve Lloyd wrote:
>
>> John,
>> I believe that to first order all sites should support all (HEP) VOs
>> and then we use target shares and scheduling to set the right amounts
>> (as per my ~GridPP11 talk). This is much more efficient and allows one
>> experiment to get maximum resource if nothing else is going on. e.g.
>> this summer we had mostly LHCb jobs running at QM even though we are
>> only in ATLAS. We would have given priority to ATLAS if their jobs came
>> our way but they rarely did for some reason.
>> Cheers Steve
>> --
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>> + Steve Lloyd Queen Mary, University of London +
>> + E-mail: [log in to unmask] Physics Department +
>> + Phone: +44-(0)20-7882-5057 Mile End Road +
>> + Fax: +44-(0)20-8981-9465 London E1 4NS, UK +
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>
>
|