Hi Jeff,
My I put forward the issue from the experiments point of view?
Our problem is not if there are 1, 2 or 10 different tags, but
to know what is the minimum "system" software that we can expect to be
installed on the site.
For each of these tags we would need a WN installed in a
"minimalist" approach were we can check if our applications run. If an
application running on that system fails at some site publishing the
same TAG due to missing libraries, executables,... the site should
either align with the reference or change the tag.
I find it great if we can agree on a reduced number of these
tags, but the work is not final until their "meaning" clearly
established and sites adhere to it.
Of course the immediate question is who defines these
references?
Regards
Ricardo
========================================================================
========
Ricardo Graciani Diaz
Dept. Estructura i Constituents de la Materia
Facultat de Fisica Tel: +34 93 403 9183
Universitat de Barcelona Fax: +34 93 402 1198
Diagonal, 647
E-08028 Barcelona
========================================================================
========
> -----Mensaje original-----
> De: LHC Computer Grid - Rollout [mailto:[log in to unmask]]
En
> nombre de Jeff Templon
> Enviado el: miércoles, 23 de febrero de 2005 14:30
> Para: [log in to unmask]
> Asunto: Re: [LCG-ROLLOUT] OS tag proposal (was unambig CE tags)
>
> Hi,
>
> OK, so I already did it for the new part of NIKHEF (lcg 2.3) -- take a
> look at the GRIS on tbn20.nikhef.nl
>
> JT
>
> On Feb 23, 2005, at 13:39, Kostas Koumantaros wrote:
>
> > Sound Great,
> > and if We automate Dist ID and Release It would be perfect
> > :)
> >
> > Kostas
> >
> >
> > Jeff Templon wrote:
> >
> >> Hi,
> >>
> >> the /usr/bin/lsb_release -i -r idea from Kostas seems to have
> >> potential. The only thing that is missing is some generic tag,
since
> >> SL and CentOS are really two derivatives of RHEL. Perhaps we could
do
> >> the following:
> >>
> >> name: EL for enterprise Linux, RHC red hat classic numbering
scheme,
> >> etc.
> >> version: the "distributor ID" as provided by /usr/bin/lsb_release
-i
> >> release: the release number as provided by /usr/bin/lsb_release -i
-r
> >>
> >> Here is the output I got from four systems, along with what the
Glue
> >> Schema would contain if we go with the above proposal:
> >>
> >> lxplus.cern.ch
> >> Distributor ID: ScientificCERN
> >> Release: 3.0.3
> >>
> >> Name: EL
> >> Version: ScientificCERN
> >> Release: 3.0.3
> >> ==================================
> >>
> >> tbn20.nikhef.nl (our new CE)
> >> Distributor ID: CentOS
> >> Release: 3.3
> >>
> >> Name: EL
> >> Version: CentOS
> >> Release: 3.3
> >> ===================================
> >>
> >> juno.nikhef.nl (my desktop)
> >> Distributor ID: RedHat
> >> Release: 7.3
> >>
> >> Name: RHC
> >> Version: RedHat
> >> Release: 7.3
> >> ===================================
> >>
> >> my laptop
> >> dhcp-122:~ templon$ /usr/bin/lsb_release -i -r
> >> -bash: /usr/bin/lsb_release: No such file or directory
> >>
> >> [ oh well ]
> >>
> >> dhcp-122:~ templon$ uname -a
> >> Darwin dhcp-122.nikhef.nl 7.8.0 Darwin Kernel Version 7.8.0: Wed
Dec
> >> 22
> >> 14:26:17 PST 2004; root:xnu/xnu-517.11.1.obj~1/RELEASE_PPC Power
> >> Macintosh powerpc
> >>
> >> Name: BSD
> >> Version: Apple
> >> Release: 10.3.8
> >>
> >> What do people think??? Thanks Kostas for an interesting idea.
> >>
> >> JT
> >>
> >> On Feb 23, 2005, at 12:36, Kostas Georgiou wrote:
> >>
> >>> $ /usr/bin/lsb_release -a
> >>> LSB Version: 1.2.0
> >>> Distributor ID: RedHat
> >>> Description: Red Hat Linux release 7.3 (Valhalla)
> >>> Release: 7.3
> >>> Codename: Valhalla
> >>>
> >>> Kostas
> >>
> > <kkoum.vcf>
|