meeting 12th May
Cosmo, Benedikt Hegner, John Harvey,
Fabrizio Furano, Sebastian Binet [REMOTE], Rene Brun, Pere
Mato, Marco Cattaneo, Predrag Buncic,Vincenzo
Innocente , Fons Rademakers, David Rousseau, Peter
Clemencic, Stephen Gowdy, Liz Sexton-Kennedy
Line Management Issues (John)
NeuroBayes license (Pere)
Report from the LIM meeting (Benedikt)
News from projects
0 (18/5/2011 @ 11:00)
Management Issues (John)
- Install the gcc 4.3.x
compiler and use it in
one of the nightly slots
open issue of distcc not working with gcc 4.3.x is now solved. Fabrizio
reported that it took a bit long and several
interactions. Before finally closing the ticket it is required to
document what was done to be used for other future occasions when the
compiler version will be changed.
mentioned the concept of 'power users' in the new CERN ticketing system
(SNOW) which is already taken advantage by ATLAS and CMS. It was felt
that the SPI
team should also be included in there.
- SPI-33 Provide
an answer to the
Savannah feedback document produced by ATLAS
to the next iteration has been entered to the open JIRA issue. The
- Start a
negotiation with NeuroBayes for a new clause
granting access to the source code
under certain conditions (Pere)
next item in the minutes.
Last time it was agreed that Pere will contract the company to
start a negotiation of a new clause granting access to the source code
under certain conditions. This is what has been done:
- The signature process was
out on hold.
- Company contacted and their
answer has been distributed to the AF mailing list.
- At the same time we have
re-contracted to legal service in order
to get advice and eventually to get an example text of similar
conditions that have been applied to other licenses. They suggested to
get in touch with the KTT (knowledge and Technology Transfer) service.
They have experience on formulating agreements for technology developed
at CERN to be used outside, but not vice versa. They offer their help
re-direct us to the specialist on software license buyer/negotiator.
- Only one experiment (LHCb)
is actually using the software and currently needs it. Marco said that
the LHCb management has been informed and the worries for the long-term
availability of the software are not important. They are ready sign the
- The text of the agreement
will need to be changed to reflect that is only LHCb. Perhaps an ACL
would be required to ensure that only LHCb can access
the Teacher libraries.
- There was a general
agreement to close this chapter at this moment.
Planning Preparation (Pere)
Report from the LIM meeting
See the minutes
of the last LIM meeting for details.
- There was the confirmation from LHCb that Frontier for the
Windows platform is not really necessary.
- Benedikt informed us of the incidence of last weekend.
There was a total service interruption due to account
problems. The lcgnight service account was considered orphan and got an
- A graduate student from TU Munich is working on analysis of
usages of Savannah and other issue tracking tools. She is a guest of
the SFT group for few months.
- The Savannah user survey was sent around and now experiments will
be approached. The results will be presented at the next ITUM.
- The migration from LCGCVS to CVS still not fully
satisfactory. Stephen reported that for CMS has been suffering
from a huge performance penalty (factor 10 slower). It seems that they
were running from a single server due to some kerberos problems of the
load balancing mechanism. This is now solved and the performance is
News from Projects
- ROOT (Fons)
- Patch 5.28.00d was released in urgency during the weekend under
the request of ALICE. The changes are
documented at the patch
- CernVM (Predrag)
- Released new development
version for LXcloud
- This new CernVM image can be
contextualized for any experiment, 'Pilot jobs' can be started, and it
contains the EOS disk cache software to read and write data.
- Unfortunately there was a glitch on
the Fermilab server and they needed to
reboot some nodes. Predrag mentioned that FNAL do not follow the
deployment model that we
recommend and this causes problems to debug it. The person responsible
is at CERN
and a discussion has been scheduled.
- There was another incident
yesterday that effected LHCb. ATLAS started a bulk update of their 300
GB repository to the new structure all at
once. This is large operation affected LHCb when trying to publish a
new software release. The solution was to suspend the ATLAS update and
the LHCb did their installation. Production was not affected because
they were re-directed to stratum 1 servers.
- Multi-core (Vincenzo)
- Paolo Carlini (gcc standard libraries) is visiting.
Discussions with Geant4 and ROOT has been scheduled. The Computing
Seminar is tomorrow.
- PF (Andrea)
- Planning for this project is still missing.
- Clarification on the desire from LHCb to drop POOL. This
operation will take
a time scale of one year. It will have implications for ATLAS.
The solution to optimize buffer size, unnecessary seeks and speed was
already adopted by ATLAS with their use of Tree and Branches.
Sebastien suggested that the
implementation should be perhaps in GAUDI such
that ATLAS can profit. This needs more discussion and Andrea will
follow this up
- Following up some problems with
ATLAS to do with Kerberos and Oracle,
- Marco asked about software licensing policies at
CERN. John said that task force dealing with will meet again next
Tuesday. He will report on the outcome of the task force at due
- Starting to validate for
good the configuration LCG_60b. So far so
- Doing some cleanup on
the Savannah projects.
- Liz wanted to talk about
persistency problems. CMS is using ROOT directly
since many years and they are interested in making use of the ROOT
schema evolution mechanism presented at Vancouver's CHEP. Their people
what they have is just vaporware. She wants to organize a meeting next
week. An evaluation of what is the effort needed to make it usable is
- Continuing with some patching
of ROOT 5.27.06.
- ALICE suffered from specific performance
problems in the analysis. They requested a urgent patch release of ROOT
to be ready for Quark Matter in