Meeting 2019-07-01


  • Brian
  • Derek
  • James Deaton
  • John
  • Shawn

Condor monitoring

  • On every submit host, we’re parsing the condor transfer logs
    • The transfers that Condor manages: input and output files
    • Not for transfer with the jobs (such as internal transfers in CMS code)
  • The regional networks are implementing more metrics (James)
    • Better metrics are giving more opportunities to discover improvements
    • NetSage global science registry (Shawn):
    • Regional traceroute:
    • Who are the top three campuses to engage? (Brian)
      • Most interesting by retransmissions: Site in Mexico, Syracuse, UMD
      • Interesting metric: retransmissions per GB
    • Action item for Derek: Start Google doc to track outreach progress
      • Share with mailing list and James

Upcoming SAND presentation

  • To SIG-PMV, July 2, 2019 (Shawn)
    • GÉANT SIG:Performance Monitoring and Verification
  • Review and feedback to Shawn

Face-to-face followups

  • Derek and Shawn met with Mark
    • Complete configuration for RabbitMQ archiver
    • New changes in 4.2 that are needed to make it work
      • pScheduler and other updates for new features
    • Ready to demonstrate that it works
    • Next step is to contact pwa developer to add queue information to psConfig

Older items

perfSONAR collector

  • Rename the host to remove historical “RSV” (Derek)
    • Repo is now renamed
    • Email Shawn to update hostnames/CNAMEs (Derek/Shawn)
  • Add ASN to collected data
    • Follow up with Marian B

check_mk authentication

  • Migrated to CILogon OIDC
    • Planning integration with COmanage for authZ
      • Awaiting OSG instance of COmanage (UW enrollment)
      • Tested CILogon COmanage test instance (John)

Outstanding action items

  • Pick the four most important services and have them captured by check_mk (John) - 2019-04-01
    1. ps-collectors
      • Need to enable additional checks
      • Is there a rabbitmq client or log file to monitor?
      • Have ingester write out a processed event count?
    2. Topology - HTTPS tests, will improve
    3. Repo - HTTPS tests, will improve
      • Hosted CEs - Derek implemented
      • GRACC
  • Link presentations from the SAND website (Shawn) - 2019-03-25