• Created by Unknown User (jalt), last modified on Sep 30, 2016

Date

Attendees

Discussions

Hardware Acquisition

QServ

  • (Jason) Any discussion/decision on the dbdev* hardware replacement(s)
  • (Fritz) Consider that qserv developers want a 'playground' environment (ssd, nvme). 
  • (Jason) Can we wait until Skylake? Would like to give you Apache Pass + 6 memory channels to play with.

PDAC

Stripe82 Transfers (Igor Gaponenko)

    • all coadds (FITS files) have been restored from IN2P3's iRODS and copied over to NCSA at: /datasets/gapon/data/DC_2013/coadd/deepCoadd
    • ~55% of calexpes have been transfered from IN2P3 to NCSA so far. They're found at: /datasets/gapon/data/DC_2013/calexps/sci-results/

Note, this is a GPFS file system mounted on some nodes of PDAC.

    • The stripe82 catalogs have been (some file are still being transferred from Nebula onto /datasets) partitioned and are ready to be loaded into PDAQ Qserv as soon as it will be available.

SUI

  • Can we do builds on the machines (Xiuqin)
  • Yes, just remember no DR is provided (Jason)

Qserv / Containerization

  • Fabrice Jammes able to launch on 30 nights
    • Working through env config issues with NCSA (quick turn around)
  • Brian Van Klaveren working on containerized DAX services
    • Working on that on the next 2 days. Waiting on work on dax01.
    • Need docker (may yum install) will send request to NCSA

Calibration of single epoch exposures

  • Necessary software/module/libraries not currently available on monolithic database, but could be easily deployed

Qserv / Dax services available next week to SUI team (very tentative)

    • Qserv up, Brian hopeful on services
    • Exposures to become available as they are transferred into /datasets
    • Some Calexp exposures missing (~20%), Igor Gaponenko to research and report on what is missing.
    • (gpdf) what happens when a calexp is accessed that is missing? Brian Van Klaveren imgsrv should return an error