I forward this to TB support in case not all are on hepsysman.
-----Original Message-----
From: HEP System Managers mailing list
[mailto:[log in to unmask]] On Behalf Of Winnie Lacesso
Sent: 07 April 2006 13:48
To: [log in to unmask]
Subject: Using external (non-SL cluster) with LCG grid SE/CE/MON?
Greetings,
Bristol has a small PP LCG Grid (CE,SE,MON,UI & few WN, all
ScientificLinux
3.0.5) & may soon have access to a large University cluster (>1000
nodes) which would be managed by University Info Systems admins.
Using that cluster (portion anyway) as PP Grid WN is said to be
possible, ie
PP LCG CE/SE/MON -> uni.cluster.PBS.headnode => many.cluster.WN
Many Uni groups will use the Uni.Cluster.
What constraints are there re: configuration &/or OS of the
cluster.PBS.headnode & cluster.WN to be 'the farm' of the PP LCG
CE/SE/MON?
Does it matter if the remote cluster is Linux of some species (probably
not Scientific Linux), or not?
Does there need to be _any_ LCG-specific software or config of the
cluster.PBS.Headnode or cluster.WN? What & how much?
There might be restrictions on having ports open on the cluster.WN to
external machines (ie Bristol's PP LCG CE/SE/MON). Would that cause
problems?
Does anyone have practical experience with or done this, or can name
names of LCG site contacts who have ?
Some LCG info online seems to indicate there would be output/ssh issues
between the cluster.WN & the LCG CE requiring cluster.WN config, which
may be problematic (or may not) & that all cluster nodes (head+WN) would
require some specific software &/or config to report back to PP LCG
CE/SE/MON.
Some proposals for the uni.cluster include things like 'WN cannot run an
alternative OS', 'OS will be SuSE 9.3', & security issues that may or
may not (unknown yet) allow ssh & open ports from cluster.WN to external
PP LCG nodes.
So, it's needed to be known what hard constraints there are on using
external cluster as the farm for an LCG grid site, before we agree.
Thanks very much for any advice on this.
Winnie L / Bristol
Winnie Lacesso - Bristol University Particle Physics Computing Systems
HH Wills Physics Laboratory, Tyndall Avenue, Bristol, BS8 1TL, UK
Email: Winnie.Lacesso at Bristol.ac.uk Phone:0117 928 8738 Fax: 925
5624
|