IPP Progress Report for the week 2011.11.07 - 2011.11.12

(Up to IPP Progress Reports)

Eugene Magnier

The team on Maui did an excellent job of installing and setting up the new wave 4 storage machines (13 new 5U nodes w/ 40TB each). We will be putting them into operations as soon as the operating systems are set up.

I spent much of the week working on NSF proposals. On the development side, I merged my branch with updates to the psphotStack analysis and improvements to pantasks into the trunk so we could build a new tag. We set up a new tag for operations: ipp-20111110.

Serge Chastel

Heather Flewelling

  • sick 2.5 days
  • apec conference half a day
  • devising new/better/safer way to build dvodbs without getting corruption
  • answered questions for tom
  • tested new tag (mds through stack, diff, destreak just to verify it works)

Roy Henderson

  • vacation Mon/Tues/Wed
  • monitored loading
  • some tedious work reloading 1500-odd batches that were placed on hold in the ODM. had to be very carefull
  • updated IN batch with 'w' filter in Filter table
  • looked into missing extended properties tables in SAS data
  • caught up with LAP loading after DVO merging was restarted
  • continued to load old 3PI.Now have 9 - 15 hours RA
  • updated PSPS news page with plots of both old and new 3PI.
  • usual mailing list admin
  • created wiki page to bring together all the suggested schema changes
  • added 'deepstacks' server to czartool

Mark Huber

  • czar monday, on-going learning on the processing flow and problem recovery.
  • MD07.V3 refstacks setup, finished, verified and distributed along with sample of nightly stacks with various dates since 4/2009.
  • ppSub/ppStack -- still tracing through process. focused on strange results in convolution to target PSF.
  • psphotStack testing -- tracked down bug causing memory issue, setup for running SAS.footprint, SAS.123 and MD04 staticsky sample.

Bill Sweeney

  • worked on background restore scripts to allow them to operate without a chip or warp BackgroundRun? being defined. This will allow the postage stamp server to use "almost" the same code to implement background uncorrected postage stamps
  • Wrote a script to find and copy all of the magic streaks files to a directory on stare00 for Paul Sydney
  • Helped Heather debug a configuration file problem with the trunk.
  • one day as processing czar. Unfortunately we are still getting corrupt output files. We're not sure if the switch to hard nfs mounts helped or not.

Chris Waters

  • Diskspace: further study of currently used diskspace. Switched to a more complete diskspace measurement based on the actual disk usage. However, counts do not agree well between database entries and disk files, so this requires more work. Updated documentation.
  • Sky background: Various tests and iterations to change which cells receive pattern corrections. It appears that more cells have this correction performed than need it, and this causes the sky background level to be fit poorly.