Skip to content

OSG Technology Area Meeting, 30 August 2022

Announcements

  • Carl out Wednesdays
  • Ziyang's last day is today

Triage Duty

  • This week: Mat
  • Next week: TimT
  • 11 (+4) open FreshDesk tickets
  • 0 (+0) open GGUS ticket

Jira (as of Monday)

# of tickets Δ State
191 +1 Open
44 +0 Selected for Dev
26 +2 In Progress
12 -2 Dev Complete
2 -3 Ready for Testing
0 +0 Ready for Release

OSG Software Team

  • AI (Mat): OSDF origin chart
  • AI (Mat): Assist Fabio with setting up the University of Hawaii origin; we are currently waiting on them for Resource Group information, after which Mat will help with Data Federation information -- the format of that information is new
  • AI (Carl): XRootD 5.5.0 release; contact Mat for assistance as needed
  • AI (Carl): Review various PRs for the Tiger Kubernetes cluster
  • AI (BrianL): Review and reprioritize Software Team JIRA tickets
  • AI (BrianL): Review PATh metrics for the monthly report before the end of the month

Discussion

  • Marco: GlideinWMS 3.9.6 expected by the end of the week; fixed setup issues noticed by Mat and others, and added a token generator

Support Update

  • USC (BrianL): helped solve issues with backfill containers; issues caused by tokens and using an outdated image. Admins were surprised that the pilots exited because they didn't get any jobs; this event should be communicated more clearly, perhaps in the container logs
  • Virgo (BrianL): helped Jason resolve issues with Virgo proxy generation due to upstream VOMS server cert update
  • LIGO (Carl): assisted Peter Couvares with getting the HTCondor version on a CE
  • LIGO (Carl): received formal request from LIGO for first-class SIF file support on the OS Pool; redirected to Jason but Mats should also be added -- for now this is a question of OS Pool policy
  • OS Pool (Derek): user wanted to have a very large file accessible on /cvmfs/stash.osgstorage.org. Derek increased the max file size from 26 GB to 500 GB. This change should not affect anything except large files, but staff should keep an eye out for issues
  • MIT (Derek): Credential for lightweight issuer, hopefully can be resolved with a few more back and forths.
  • WTAMU (Derek): observed a difference between pilot and payload hours -- looks to be due to the site having huge slots (64 cores) with Glideins pilots retiring (finishing old jobs but no longer accepting new jobs) but with a handful of long, small jobs keeping the pilot alive. Mats suggested increasing the pilot lifetime
  • OS Pool (Derek): cvmfs-singularity-sync started deleting containers on CVMFS yesterday (8/29). The problem was fixed and the containers restored; the issue was due to code changes in order to support tag wildcards on hub.opensciencegrid.org. Derek is writing a full incident report.

OSG Release Team

  • TimT: new condor releases (9.11.1)

  • Ready for Testing

    • gratia-probe
    • stashcp

Discussion

  • TimT: note prce2 more strict about character classes in regex (hyphen must go at the end [abc-])
  • BrianL: need to find victims to update from stable series to 9.12
  • AI (TimT): reach out to John Thiltges about feasibility of updating to 10.0? (TimT: 9.12.0 in a couple weeks)