Core Minutes 5/6/2014

• News of the day: (Tom S.) will be taking a tenure-track faculty position at Brigham Young University as of June. He'll be winding down Fermi Infrastructure work (more on that below in Core Talk) but will continue to be a Fermi Collaborator.

• ScienceTools: (Jim) No new development news.

• FSSC: (Joe) No news this week.

• Downtime: (Richard) We're in the midst of it. Affected hardware includes Oracle servers (upgrade), file servers (patch) and fermilnx boxes (patch). According to schedule it should all be put back together by 7 PM tonight.

  (Tom G.) In addition 7 machines are being physically moved during this outage. Next Monday there will be another outage to move the new Oracle servers.

• Reprocessing: (Tom G.)

  • There was a task to create new merit files for the new event classification. That is now complete (that is, we're up to date through June 2013).
  • There is a catch-up request for Pass8 reprocessing. Some changes are being made to this task to improve efficiency in interactions with Data Catalog.
  • In the wings: new FITS files for P7 reprocessing. It's essentially ready to go. He's waiting for decision about final release date.

• New hardware plans: (Richard) The first of two meetings (one for budget proposal; the second for final agreement) is coming up on June 6. We'll ask for an Oracle development server, another fermilnx machine as back-up, and some relief for the glast ground web server situation. Current configuration involves a load balancer and two Windows web servers running a deprecated OS version.

• Pass8: (Leon) needs to create overlay livetime files. In theory this is a routine procedure, but it's been a while since the last time and there has been significant code rot. This needs to get done before Saturday, when he leaves on vacation for two weeks.

• Core talk:

  • Python (Heather) Our new release was temporarily hung up when I discovered Anaconda (Python distribution kit) cannot install without hard links. Our externals are installed in AFS, where hard links are not possible. For this release I'll install by hand (in progress). Anaconda developers have assured me that their next version will support copying, soft links, etc.

  • Future of RM (Heather) has summarized in Confluence (Tom S.) Follow link to my documentation to do list to monitor progress in that area. RM itself is at the maintenance-only stage except for the following:

    • Get Jenkins to do rhel5 64-bit and rhel6 64-bit builds. It's already handling all the others.
    • Add Jenkins mechanism to clean up old builds.
    • Would like to get rid of our last little dependency on lsf

    Should we quit supporting ST on Windows? (Tom S.) Yes! It hasn't built successfully for a year and half, and no one seems to care. FSSC doesn't distribute Windows builds either. (Heather) What about Mac? Difficult to support if developers won't keep an eye on the builds. (Jim) is willing as long as he has a convenient way to log on to the platform.

  • Checking up on calibrations (Joanne) refurbished an old program to do just that in response to request from Leon. The program checks for gaps or overlaps for calibrations of a specified flavor. It hadn't been touched since we moved to SCons builds so needed some work, but is now in running condition. (Leon) asked for this as part of a larger plan to test correctness of new entries before they're unleashed for production use.

previous

minutes index

next

J. Bogart, Last Modified: