Hey! So an issue myself and Jens have briefly discussed is the decision
process on when to recommend sites to upgrade. Ie do a risk assessment
of the upgrade and a cost/benefit analysis of problems vs. improvements.
Was beginning to look for terms an constraints.
Ie do benefits of upgrade affect multiple/single VOs? Which ones?
Does the upgrade provide an ability which is
necessary/requested/superfluous?
What is the chance the upgrade?
The Tier 1 now has a change control procedure, I though perhaps adapting
this to apply for Storage upgrades at T2 would work?
Comments? ( especially form those at the Tier1 who have been part of the
change control process as to whether this is a sensible idea in the
first place !!)
Regards
Brian
-----Original Message-----
From: GRIDPP2: Deployment and support of SRM and local storage
management [mailto:[log in to unmask]] On Behalf Of Sam
Skipsey
Sent: 11 May 2010 15:22
To: [log in to unmask]
Subject: Potential new golden DPM...
Hello all,
DPM 1.7.3 has been out for more than a month now, and no-one has
complained yet(!).
One of the improvements of 1.7.3 over the previous releases is that it
contains a workaround for brokenness in the lcg_utils library. That
is: 1.7.3 will store checksums requested by lcg_utils functions (like
lcg-cr).
ATLAS would like to be able to rely on checksums they request being
stored by the SE. I would imagine this is also true for other VOs. So,
whilst it isn't Golden yet, I'd like to encourage everyone to move their
DPM to 1.7.3 if possible.
(I've done it just now on svr025, and it seemed to be very
straightforward. svr018 will be moved as soon as I've finished some
drains I'm doing.)
Sam
|