Core Minutes 7/2/2013

• ScienceTools: (Jim) There is just one item: updates from Jeremy to improve TimeRangeError output.

  See Science Tools Development Notes for details.

• FSSC: (Tom S.) Just ongoing internal work; no news for this meeting.

• Pass7 Reprocessing: (Tom G.) Root reprocessing is now up to date (through June 25). The bulk of it took about 2 days + another day and a half for outliers: jobs stalled for no apparent reason and so forth.

  The question of whether or not to redo FITS files is under discussion, to be decided by C & A.

• Hardware: (Richard, Tom G.) The new file server is now online. It has 2 10-Gbit ethernet connections and about 350 Tbytes of space. We're channeling a lot more data through a single machine than has been the case up till now (previously no single server had more than 91 Tbytes), an experiment to be watched.

  (Heather) What about the 2 new mac boxes? Can we get them? (Richard) Yes, we have the money. He asked Yemi for a spec a while ago; no response yet. He will poke again.

• Pass8 (Tracy) GR 20-09-00 finished building this morning. This is our candidate for reprocessing. (Tom G.) is ready to push the button and expects to do so later today.

  (Richard) What are the final timing numbers? (Tracy) The result of Philippe's CalRecon work targeting horizontal tracks is a constraint on radiation length. This saves some time with no bad side-effects; we're within 3 x P7 timing, as hoped for. If we eliminate CR tracking it would be more like 2.5. That discussion is ongoing. For a few weeks at least we'll leave CR tracking in to see if a case for it (as an aid in background rejection) can be made. (Tom G.) Will be able to make estimates of how long the reprocessing will take after a week or so.

• Core talk:

  • GR 20-09-00 (Heather) will install on fermi-ts. Among other things, it contains Joanne's fix, so that all of the "all" solution files (all.sln, alluserAlg.sln, allGleam.sln) are being properly produced.

  • jo proposal (Tracy) would like to make it easier to track down exactly which job options were used for a given run, e.g. maybe attach to the relevant JIRA.

  • New Windows password for glast account (Heather) Coming soon if there are no objections [there aren't any].

  • Release Manager/Jenkins (Tom S.) Tony J. did some work on Mac Jenkins and it's looking good; it may be fixed. He promised to look at Windows Jenkins today. (Richard) lsf9 upgrade is scheduled for Aug. 5 so it's important to get Windows Jenkins going.

  • Memory leaks (Heather) 1 Gbyte after 20,000 events: it's troubling, but doesn't cause problems as long as chunks are kept small.

  • Vacation! (Heather) disappears for 10 days starting Thursday, but she'll be in touch. (Richard) expects many will take off Friday. We should be sure we're covered.

previous

minutes index

next

J. Bogart, Last Modified: