Prototype Data Access Center (PDAC) meetings take place every other Thursday at 9:00 Project Time on the BlueJeans infrastructure-meeting channel: https://bluejeans.com/383721668

Date

Attendees

Goals

Discussion items

TimeItemWhoNotes

PDAC hardware status

  • QServ czar upgrade, see
Unknown User (pdomagala)

Last week this was still stuck in NCSA procurement. Current status?

  • Donald Petravick signed off on this yesterday. It has been through PMO already & Wil signed off.  Next step is for NCSA to issue the PO. Unknown User (pdomagala) actively tracking this.
  • Gregory Dubois-Felsmann, Fritz Mueller: Note that the schedule contingency for the loading of the WISE and NEOWISE single-epoch photometry (LSST "Source"-like data) during this cycle has now largely been consumed (the czar upgrade request has been pending since June). This is a high-level milestone visible to our funding agencies.

Firefly security screening

  • See

Last week Unknown User (pdomagala) discussed this with Unknown User (awithers), resulted in delegating this to Christopher Clausen. Awaiting a status update on the ticket.

  • Christopher Clausen requested a network configuration change at NCSA on to enable access from a scanning appliance. No news on that request yet.
  • He will try to do some other analysis work on this while waiting.
Walk-ons


Status of

  • The need for improving the current capabilities of NCSA for exchanging data with CC-IN2P3 is still present, to allow CC-IN2P3 to validate and demonstrate its capability to import and export LSST data at the level required for the operations phase.
  • Unknown User (pdomagala) suggested to try using SLURM-managed compute nodes for the data exchange exercises. Fabio Hernandez stated although using compute node as ephemeral data transfer nodes could be technically possible provided that the NCSA firewall configuration allows for this (see the ticket for details), it would not be very convenient for rapid cycles of testing.

Appropriate forum for addressing Kubernetes deployments

Monitoring of container services

  • Considering using https://prometheus.io to monitor the "middle level" of container services in addition to the lower level hardware (e.g., CPU load, memory commitment) monitoring already in place.

Action items

Attachments