Core Minutes 12 April 2011

Core Minutes 4/12/2011

Present: Anders Borgland, Toby Burnett, Jim Chiang, Stefano Ciprini, Richard Dubois, Warren Focke, Tom Glanzman, Heather Kelly, David Landriu, Julie McEnery, Chuck Patterson, Leon Rochester, Tom Stephens, Tracy Usher, Eric Winter

ScienceTools: (Jim) 09-23-00 is the current release and will be used for the upcoming P6_V11 data release.
Power outages: (Richard) The first was last weekend; there were some hiccups with xrootd servers. Level 1 experienced an issue with magic 7 data being on both NFS and xroot disk but required all xrootd servers to be present. A list of available xrootd servers is maintained by the ISOC to query directly. It was not updated for the outage. It is not understood why the M7 code was accessing xrootd anyhow since it is said to be just for older file access.

The 'xrdls' command also requires a list of available servers (maintained on the server side) which was not updated for this outage — but will be for the next one.

One more 2-day power outage is on the schedule (weekend of 7-8 May 2010).
Keeping released data up to date: (Tom G.) Released data consists of two parts:
- data products from a reprocessing task (from Aug 2008 up to some cut-off date)
- data products from Level 1 (from the cut-off date to present)
A problem arises when either the reprocessing task or Level 1 reprocesses an old run. While Level 1 reprocessings are automatically propagated to FSSC, there is no alert mechanism for other interested parties. Two possible solutions are suggested:
- Using dataCatalog find command to maintain up-to-date list of files
- Using the download manager
The first option is being investigated for incorporation into an automated alert tool. The second option could be used as an effective 'rsync' between SLAC and remote repositories. Questions on the first option can be directed to Tom G. or Brian, while questions on the download manager should be directed to Tony Johnson.
FT1/FT2 mirror in Italy: (Richard) As part of an effort to keep ASDC in the loop, we've invited Stefano Ciprini to attend software meetings.
Documentation: (Chuck) sends the following:

A ScienceTools Setup page for P6_V11_DIFFUSE has been added and is currently linked in from the LAT Science Analysis page, which has also been restructured and updated. I'm also scheduled to meet with Gino again later today. This will be our second in an ongoing series of meetings to coordinate this section and the Science Tools section of the WB with the advanced pages being developed in Confluence.
GR/Onboard filter: (Tracy) continues work with JJ Russell and Tony Waite who are updating FSW onboard filter to be Linux rehl5/6 compliant and also trying to reduce the number of dependent packages (seven?). There is an issue with making this code work on Windows: FSW team will not do it; do we really need it? Still need to make connection with Gleam to verify all works.

We may not need revised Windows version for P8 development, but Tracy is still seeing how close he can get. Prep for Fermi symposium may soon impinge on this work.
Pass8: (Leon)Validation/development studies are ongoing (with working code!). Current GR version is 18-08-14. Improvements in pattern recognition continue.
Core talk:
- glast-ts and externals (Heather) asked Toby if all was well with the GlastRelease-v18r8p14's Release build. Over the weekend it was found that the OBF external was only installed for VC8debug which prevented one from building in Release mode on glast-ts. The VC8 version OBF is now installed. With Toby's blessing, Heather pushed the button to create the release version of GR-v18r8p14
- Missing binary installs (Heather) On Monday, Michael Kuss reported trouble downloading the binary install of GlastRelease-v18r1p1 (yes v18r1p1). It had disappeared (as had others off u09). Heather asked Kim to reinstate it. That proved difficult. Instead, Heather deleted and rebuilt that version of rhel4_gcc34opt and Michael was ultimately able to install it.
- SCons (Heather) Ongoing work for ST Snowleopard support (f2c); 2 packages are still failing. Fix for problem with 32- vs 64-bit detection is in current HEAD builds.
  
  Ongoing work for GR: new versions of G4 (9.4) and Gaudi (but even newer version of the latter may be needed); lots of clean-up details.
- SCons RM (Tom S.) Nothing to report. Last week's hiccup with windows builds needed a server 'kick'.
AOB (Richard) Instances of MySQL v4 server need to be updated to v5 (found during DOE audit). RM-CMT db on glastlnx01 likely needs an upgrade. Heather will talk to Kim about it.

minutes index

T. Glanzman et. al./jrb, Last Modified: