PDAC Meeting 2017-06-01

Krzysztof Suberlak

• Wrapping up report. Performed detailed single-object search, confirmed that the data match what is at UW.
• Occasionally get "none" values for PSF flux in Forced Source table. Simon Krughoff: were flags set?
• Did source density comparisons between PDAC and UW catalogs in coadded i-band.
  • In PDAC catalog there are extra sources with near identical coordinates for ~10% of sources
  • Some question about how deblending was done. Suggestion that the UI help the user understand the structure of deblends and how to determine which catalog entries to use for deblended sources.
• Suggestion on how to select columns for display: allow users to select/deselect related groups of columns (e.g., lots of "multishapelet"-related columns)

Fritz Mueller et al.

• Prioritizing work on the DAX v1 Web services, expecting to be much more IVOA-oriented (Brian Van Klaveren and Kenny Lo)
  • Subset of SODA for imgserv
  • Subset of SIAv2 for image-finding part of metaserv
  • RegTAP for metaserv
  • Asynchronous-query support (currently implementing in Qserv itself, then will push up through dbserv)
• Will update planning page with the related epics today
• Working on authentication-token passing
• Concerns about column-type metadata not coming down from Science Pipelines
  • No UCDs
  • Units are mostly OK, but problems with angles
  • Brian Van Klaveren has discussed this with Jim Bosch
  • Closely related to ability to do afw.table round-tripping

Igor Gaponenko (reported by Fritz Mueller)

• Working on the installation of the "4-band cryo" and "3-band cryo" datasets; transferred from IPAC, working on column-metadata loading.
  • See DM-10740 - Getting issue details... STATUS  

• Planning to release tomorrow an initial version of the plan for consolidated problem-reporting
• Have dealt with a variety of filesystem-full and ran-out-of-inodes problems. Once recent case of large-inode-count was both reported and triggered by Igor Gaponenko; Unknown User (pdomagala) asked for details. Plan to discuss this in more detail at the 08 Jun 2017 infrastructure meeting.
• Monitoring:
  • Currently no monitoring on Nebula, need to do something about that. May use that as a chance to compare using Nagios to some newer alternatives (e.g. Grafana, InfluxDB)
  • Unknown User (pdomagala) will arrange for a report on near-term plans for monitoring at next week's infrastructure meeting.
  • Fritz Mueller asks that SUIT post some specifics of what they would like to have monitored in their systems
  • Fritz Mueller wonders if the alternatives being considered can match the real-time latency of Ganglia?
  • Technology decisions to be made by Michelle Butler

Unknown User (xiuqin) et al.

Gregory Dubois-Felsmann