Mato, Fabrizio Furano, Gabriele
Cosmo, Marco Cattaneo, Vincenzo Innocente, Antonino Sergi, Andrea
Valassi, Predrag Buncic, John
Harvey, Rene Brun, Fons Rademakers, Marco
Gowdy, Liz Sexton-Kennedy, Peter
Line Management Issues (John)
Response to ATLAS Savannah feedback
Report from the LIM meeting (Benedikt)
News from projects
version 1 12/4/2011 @ 11:00)
Management Issues (John)
- John informed us of the task force on licenses for open software
at CERN. This is a limited duration task force to recommend a CERN
specific license. The PH department, to whom he is representing, has
some specificity due to the handling of the copyright in case of
collaborations. Any feedback message from the experiments can be given
- There has been a call for Technical Students. John is collecting
the requests from the SFT projects.
- Better document the actual contents of a
- Database has been fixed. Every release will get a
pointer to releases notes (with an anchor). Action can be closed.
- Install the gcc 4.3.x
compiler and use it in
one of the nightly slots
Provide an answer to the
Savannah feedback document produced by ATLAS
- See next point in the minutes
Response to ATLAS Savannah feedback
- We has been discussing in the AF the future of Savannah triggered
by the ATLAS feedback given to us last November. An answer has been
produced and distributed to the AF (also attached to SPI-33 ) we are
currently waiting for ATLAS to agree or not.
- Few weeks ago IT started a discussion about trackers. They
launched a survey to the users of Savannah and Trac. Currently they are
preparing a better questionnaire.
- It was mentioned that we need to find an entry point to the
experiments that is not limited to offline software since users of
these kind of services are ranging from operations, online and
- Vincenzo suggested that when giving wishes, experiments should
quote which another tool(s) provides the wished feature to avoid
inventing non-implementable or conflicting requirements.
- Savannah was designed as a bug tracking tool, but complicated
flows has been added-on. This would make difficult the migration to
another tool that is simply a bug tracker.
- As soon as ATLAS agrees with the feedback we will be adding the
proposed actions in the first part of the feedback in the to-do list.
Planning Preparation (Pere)
- Pere reported that he has not received any feedback from the stories
under discussion from the experiments. It was agreed that the
deadline for comments or feedback is until next AF meeting (April 14).
After this time all these stories will be considered as discussed.
Report from the LIM meeting
- Savannah and its future was discussed.
- Compiler version and platforms:
- 4.3.5 update requested by ATLAS. Perhaps is not solving their
issues since they asked for 4.4, which is buggy.
- ATLAS is currently testing gcc 4.5 to see if we could move to
- Quickly discussed gcc 4.6. According to Vincenzo there is no a
major regression, the risk is rather low but is perhaps too early.
Proposed to go to latest gcc 4.5.x instead.
- CernVM-FS discussion
- At next LIM meeting there will be a discussion on the proposed
- ROOT is changing its tagging model:
- Fons explained that one month before a major release they will
make a branch in which they will tag 'candidate releases' (5.30.00-rc1,
5.30.00-rc2, etc.). Once everything has been tested and validated by
the experiments they will make the final tag and then continue in the
same branch with the patches (5.30.01, 5.30.02, etc.)
- This new way allows to continue the development in the main
'trunk' without any disruption. The LLVM project is doing similarly.
- The nightlies will need to re-reconfigured to use this branch.
News from Projects
- New CernVM version 2.3.0 will be announce later today.
- It includes the compliance to the new FS namespace. The CMS
repository migration to the new structure is already done. ATLAS is
starting the migration. Then the rest of the repositories will be done
- It includes also the contextualization for cloud deployment.
- CernVM-FS replicas at RAL and BNL are prepared. The CERN one in
under construction. At the GDB meeting next week there will be a report.
- Predrag reported about the Google Summer of Code since CernVM
has been accepted as one of the software project organizations. More
than 30 students have contacted the team. Next Monday is the deadline
for proposals from the students. Expected to get a few of them.
- There has been a request from DIRAC to create a top level
repository in CernVM-FS. What has been proposed is to create a top
level projects.cern.ch as common root and develop a portal that will
allow projects to 'publish' their software installations easily. This
could be interesting for other groups such as TH.
- Vincenzo has been contacted by Super-B and CMS to express their
- It has been proposed to organize a follow-up workshop for
multi-core and virtualization. This can be the opportunity to make an
update of the status on adopting multi-core by the experiments and
computer centers (whole-node task force) and to present the
strategies for virtualization and cloud computing. We need to choose
the time an announce it. --> Action Pere.
- Working on some changes in Oracle access to protect against
network glitches. Complicated business.
- The Geant4 development release of March is ready. It
contains fixes to the physics setup for anti-ions to be used by ALICE
- Antonino asked what to do with the AF problems?
- Benedikt has been reporting them to the IT C5 meetings.
- A kernel upgrade and a client upgrade did happen
- Some more discussion was raised since it is not clear if this
is an AFS problem or something else problem. Perhaps the detailed
description and perception of the problem is not transmitted correctly
to the IT AFS service managers.
- Pere will be meeting Tim Bell and Helge Meinhard next week to
discuss IT representative to AF.
- Pere informed that there is preparation a NeuroBayes license to
be signed by CERN on behalf of all the experimental groups to simply
the paper work with respect the current agreements. The proposal will
circulated for discussion.