Core Minutes 5/10/2005  

Present:    Anders Borgland, David Chamont, Richard Dubois, Matt Langston, Toby Burnett, James Chiang, Igor Pavlin, Tom Glanzman, Zachary Fewtrell, Tony Johnson, Michael Kuss, Anders Borgland, Leon Rochester, Berrie Giebels, Chuck Patterson, Tracy Usher, Dirk Petry, James Peachey, Seth Digel, Navid Golpayegani, Dan Flath
 

 Pre-meeting chat:

• Toby notes that his USB microphone does not work completely with RAT - no mute. Will complain to vrvs folks.
• Richard announces SLAC I&T Tea meeting in Orange Room Thursday 3:30-5:00pm. There will be four GLAST software presentations. Everyone is invited! See http://www2.slac.stanford.edu/ITnTEA/info/2005_05_12.htm for details.
• Tom gets VRVS sound working on Linux, but at some point it boots him out of the meeting (which explains an early gap in the meeting notes - due to shared url going into his main vrvs window and usurping it!).

Meeting:
 

• Data handling: (Tony)
  • Tony J. reports the Windows installer is being beta tested by others, and seems to work. Work continues on DB and pipeline projects. There are still a few details to work out and/or document and/or clarify, e.g. installing CMT, and location of the root library (listed as internal and external). Richard reports success installing GLAST Release.
  • Event peeler progress is being made, by me (Tom). This is ROOT code to select a events from a user-supplied list and write them out into a new .root file. I now have a good data sample to work with - which is exhibiting some interesting "features". The code becomes more complex as more checking is found to be necessary. I'm hoping for a zeroth version in a few days (which is the same as my estimate one week ago).
• Pipeline: (Dan)
  • Dan reports on pipeline data archive, Richard reports the available disk space is now half full. Dan is working on DB issues; ready to connect with Navid's code. System is being tested. Space is being consumed at around 5% per day - so data archive is needed soon.
    Manual archive to tape is a possible backup strategy. There is also a directory permissions issue: directories are being created which are world-writable, so *anyone* can currently write...or delete. This was done on purpose early on to allow Warren et al to easily clean things up on the pipeline disks, and will be expunged from the pipeline scripts.
    
    Parallel processing for Recon jobs being worked on by Warren and Richard. The idea is to submit a batch of batch jobs, then collect and combine their results. The code to handle the threading is done.
• Release Manager: (Navid)
  • there have been problems with determining the return codes from test apps. Some errors seem legitimate and may be due to G4 not cleaning up properly at termination. But others seem almost random. The rc is fine if the test app is run by hand; not (always) if run through Navid's perl harness.
  • Toby noted that RM seemed to have stopped RM'ing. Navid had a path error that is now fixed (post meeting)
  • Richard mentioned that his tests with v6r8. Within the installer, pulsar complains about version of astro. Toby considers such a complaint to be an error but CMT forges ahead. Is this something to be fixed?
     
• System Tests: (Julie)
  • Julie reports progress is being made on system tests and its automation. RM now creates config files; each release tries it out and finds new problems. Issue raised about using the HEAD build. Julie is thinking about asking to have the systests interface modified as done for RM to allow optional hiding of builds.
• CalRecon: (Tracy)
  • Tracy reports Cal recon upgrade work is described on web pages and reported on last Thursday. "Cal val" tool is planned for inclusion once the exact form is agreed upon; schedule is nominally "a few months", but could be expedited if necessary. There was some interest in expediting.
     
• Sundry: (Richard)
  • Riccardo and Heather will be at SLAC next week. Riccardo arrives Monday night, available Tuesday morning.
  • SLAC just upgraded Win XP to SP2, including running anti-spyware software (SpySweeper) - which cannot be turned off. The sweeps are scheduled by SCS to run at noon Tuesdays and Thursdays. Leon reports a slowdown in performance. Toby reports that SpySweeper can disable activation licenses. Richard has been running this software for 6 months without that kind of problem.
  • Richard, et al to visit Udine at end of May; David Chamont and maybe Philipple Bruel to join them.
    
    end 08:53 (respectfully submitted in lieu of Joanne - Tom, with a little help from Richard)
     

previous

minutes index

next

T. Glanzman, Last Modified: 01-Jun-2010 15:47:23 -0700