Core Minutes 3/4/2014

• ScienceTools: (Jim) No new items in Development Notes since last time. He's been talking to Philippe about issues related to the Pass8 energy dispersion discussion.

• FSSC: (Joe) No news for this meeting.

• Reprocessing: (Tom G.) The fixed diffuse response is being incorporated into P7 recon. What was expected to be a simple 1-day affair turned out to be much more complicated, apparently due to changes in the behavior of the batch farm. In particular, jobs which were supposed to be "trickled in" were bunched back up by the batch system, leading to unsupportable contention on afs03. The overly-popular file has been moved to a newer, faster server. The bunching has also put a strain on the Data Catalog. There is a theory that the new problems have to do with the MPI queue, which tends to release machines in bunches. Tom is working on techniques to force trickling, but it's a struggle.

• Servers: (Tom G.) Three more servers have been retired. Only two remain: glastlnx07 and glastlnx20.

• Pass8: (Leon) Several fixes have been put in place to get ACD histograms for MC and data to agree. Success has been declared!

  That fact that energy dispersion was not included in source analysis seems to account for energy differences between MC and data, so declare success there as well.

  No new round of recon is requested just yet.

• Core talk:

  • GR (Heather) Carmelo is asking for a new GR to include the ACD updates.

  • Fred on Linux (Heather) has built Ruby 1.8.6 for Redhat 5 and Ruby 1.8.7 for Redhat 6, to be made part of our distribution for Fred. This should take care of the incompatibility with the default version of Ruby now installed on those OSes.

  • Mountain Lion (Heather for Tom S.) It's all ready to go except for getting Jenkins going. The Computer Center is going to try resetting the glastrm password in hopes that this will take care of the login issue. Meanwhile, we'll get a local account for Tony J.

previous

minutes index

next

J. Bogart, Last Modified: