Hi,
On Fri, Feb 26, 2016 at 10:55 AM, sjones <[log in to unmask]> wrote:
> Gridsites move through the Linux versions like a herd. No-one will go until
> someone goes, and once someone goes, everybody goes. Since no-one has gone
> with 7 (Centos or SL) we are all watching to see what happens.
>
The problem here is that no lcg bundle was available for the glibc
version on SL7. If that would be available, the installation would
worked I guess.
> As a general rule, it's only safe to assume that any particular software
> will only work on the current OS baseline generally adopted at the
> Gridsites, presently SL6. We should state that somewhere, although as I say,
> it is widely assumed to be true.
>
Right. But I think the general problem is that people from new VOs
like LSST, EUCLID, and others, are not familiar with what is adopted
at Gridsites currently and without the manpower to have a
representative at different sites - but these people could benefit
from the dirac service since they don't have a large software/data
management infrastructure behind them.
These VOs are mostly only researchers trying to get their data on the
grid and run jobs without knowing about the underlying structure,
while the number of people in these VOs is much smaller than for a LHC
experiment. Also, the computing infrastructure used for software
development is most likely just their desktop/laptop, which is also
very different to the LHC experiments.
Maybe it's a more general question to decide for whom this Wiki
section should be - for GridPP sites or for VO members in general. If
it's mostly for VO members, then we should also keep in mind the
members of the non-LHC experiments and their general knowledge about
the infrastructure.
Cheers,
Marcus
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
|