IPP Progress Report for the week 2011.08.08 - 2011.08.12

(Up to IPP Progress Reports)

Eugene Magnier

I have spent more time on the extended source analysis work than expected. I discovered an error in the flux integration for the Sersic models. This was the source of inconsistent results between the generation of the model and the fitted flux, and was misleading me on the total flux. This investigation was driven by the concern raised by Nigel that we are not getting the Kron mags right for bright galaxies. I think I have identified the source of that problem as well: the code was returning the PSF-scale aperture rather than the full extended aperture in the output. Finally, I am suspicious that the same integration error may have been responsible for the FWHM dependence of the psf photometry using the PS1_V1 model. The scale is about right to account for the ~1-2% variation with typical seeing.

I also put together a plan for the MHPCC hardware upgrades and started the large-scale rsyncs needed. These finished Monday, and the first set of hardware upgrades was done by Cindy and Bill.

Serge Chastel

  • IPP czar
  • MOPS czar
  • Crash logs analysis
  • nfs
  • IPP-PSPS testing

Heather Flewelling

  • vacation day monday and out sick thursday & friday
  • restarted addstar
    • invstigated timeouts - these were all database actions, so I think it's likely they occurred during some of the db downtime last week. No harm done.
    • restarted 3 pi dvodb - stopped on friday, we are running out of space
    • restarted/checked SAS dvodbs
  • summitcopy
    • found one of the odd "crash" summitcopy errors that refuses to revert - sent to Bill and Gene for investigation
    • worked with haydn to sort out isp download problems. isp downloads now, and is all caught up.

Roy Henderson

  • 4-day week - vacation on Thursday
  • ippToPsps coding:
    • finally got to work on ODM-polling code in order to monitor progress of batches loaded
    • added new method to datastore class that removes a range of batches
    • more config options: survey and PSPS survey (as PSPS needs a special 3-char version)
    • improved metrics code to report on full status of all batches for a given epoch/DVO database
    • logging now configured from ippToPsps config rather than special logging config
    • logging now and appends existing log file, so we have one log per DVO database regardless of how many times ippToPsps and started/stopped
  • czar on tues/wed
    • summitcopy issues on tuesday morning - Bill and Heather to the rescue
    • restarted stdscience as it was slow, ippc11 down tried to restart; reverted warps; killed chip_imfile.pl processes....
    • reworked 'burntool' code to poll gpc1 for registered exposures rather than using automate_stacks.pl script as before
    • jumps made a return to czar plots. I'd broken something during my last change. Fixed now.
  • PSPS loading
    • rsynced all SAS DVO databases to a different host then begain loading. All are now loaded to the datastore

Mark Huber

  • Czar Monday
  • MD10.V3 reprocessing finished warp-stack diffims for distribution and publishing to MOPS. Holding back stack-stack diffims for large sample testing ppSub convolution direction selection. Started learning clean-up process for MD10 in order to start MD01 reprocessing. Writing up details of MD10 reprocessing.
  • MD01.V3 tessellation setup. MD01.V3 y-band reference stack made.

Bill Sweeney

  • Worked with M31 team at MPG to fine tune the new M31 tessellation. Queued 222 exposures with best image quality to build the reference

stacks for the two filters that are regularly observed (i and r) .

  • Created and tested instructions for installing the subset of the IPP required to build postage stamp requests and use the data store tools (for retrieving results)
  • Helped debug various processing problems.
  • worked on design for the database tables that will be needed for the video photometry stage.
  • Ill part of the week.

Chris Waters

  • LAP: finished debugging new code to skip quickstacks. Launched first set of 16hr+ runs.
  • ROC: Identified further failure cases and tested script to fully correct files. Still needs further testing before it can be set to run automatically. Investigated simply re-creating ROC files to skip these validation and repair steps, and concluded that that would require almost the same level of work to be able to repair all the missing files.
  • Misc: Rereduced and bundled various datasets for MOPS.
  • Out on friday.