Core Minutes 4/2/2013
Core Minutes 4/2/2013ScienceTools: (Jim) Recent items under Science Tools Development Notes concern a fix for a plotting bug and enhancement to embed_python requested by Matthew.
In other news, Jim has been in touch with Steve Fegan concerning ST refactoring, and Jeremy Perkins will be in town later this week.
FSSC: (Alex) Not much news. Working on lingering 32-bit issues; on track for April 15th release. And introducing Joe Asercion, who is Eric's replacement. [Joe could hear us all right but could not be heard; welcome to SeeVogh.]
Reprocessing (Tom G.) Reprocessing backfill from 11 Dec 2012 - 23 Mar 2013 is mostly complete. All ROOT files generated, but FITS file generation is on hold, awaiting new interstellar emission model and IRFs.
(Richard) People are eager for this reprocessing. They plan to use it to help decide if Fermi's observing strategy should be changed to include more pointed observations.
Hardware (Tom G.) New Fermilnx servers were upgraded from single 1 GB/s network interface to 2x1 GB/s
Pass 8 (Tracy) Situation is similar to last week: main hold-up is better understanding hit errors. There has also been some activity on some JIRAs concerning relatively minor issues. No need for another GR release yet.
(Richard) What is status of overlay generation? (Leon) This is now Stephan Zimmer's court. He had some problem, most likely minor, with catalog access, but has been too busy with higher-priority work to get to it.
Money (Richard) Next IFC meeting is a week from today. He needs to come up with a budget proposal for 2014 which primarily will depend upon what guesses he makes about tape and disk needs. Peter has suggested being conservative; that is, asking for a largish but still plausible amount.
Core talk:
ROOT compression (Heather) made one false start but now knows how to proceed and will try again. She expects the compression will help noticeably.
bulletq (Heather) tried it out with GR 20-08-11. She had to ask unix admin to install the 32-bit version of libm (bulletq is redhat 6 64 bit) but that was the only hang-up. She is in the process of comparing output from rhel 5 and rhel 6 builds.
Jenkins (Heather) Tony tried to fix the long-standing problem of the jenkins macosx build agents not starting unless there is someone logged in as glastrm, by switching from using cron to (mac specific) launchd to start the process. He is optimistic that this will fix the problem.
Tony also continued to investigate the problem with the windows agent. It appears to consistently fail if started as a system service, but works if he logs in and starts it by hand. Tony did a test build yesterday and noticed it took 7 hr and 32min to complete. As Tom explained this is likely due to the error assertion popups that occur when we run some of the unit tests on Windows. (Richard) ...a problem we've had for over 10 years. (Heather) Tom S. hopes to try again to find a solution to this.
Fermi-TS (Heather) has installed latest version of GR, GoGui, calibrations and so forth. Joanne also has an account. Between us we got cvs, ssh, etc. working. Experiments with supersede mode (local development area built against full release build) show it to be probably adequate. (Joanne) believes there is a bug in the creation of the file all.sln which can cause problems under some circumnstances, but
(Heather) Next step will be to make accounts for Tracy and Leon so they can see what it takes to use fermi-ts (and document heretofore undocumented gotchas). (Toby) also volunteered to try out SCons/GoGui on fermi-ts. He's eager to turn off the ancient glast-ts and its more ancient domain controller. fermi-ts is faster and better maintained.
Once the alpha testers are up and running and documentation has been enhanced Joanne can make a field trip to Santa Cruz, preferably in the company of one or both of Tracy and Leon. This would take place next week at the earliest. Exactly when will depend on Bill's schedule.
Proposed new containers (Heather) We need to organize a meeting with FSSC people to discuss plans and implications of the proposed reorganization of low-level packages shared between ST and GR. Can it be SeeVogh? (Alex) People from FSSC interested in attending include Dave Davis, James Peachey, Elizabeth and Joe. He knows of no time when everyone is sure to be available. (Joanne) would like to put off the meeting until we've had a chance to make a first attempt (not official; off in some sandbox) at creating the new containers and building ST and GR against them.
previous
|
minutes index
|
next
 |
J. Bogart, Last Modified: