Meeting 2019-01-07


  • Brian
  • Derek
  • Edgar
  • Shawn
  • John


Action items

  • Derek/John/MarianZ - Documentation for pS collectors administration
    • Collectors running on host under UNL T2 puppet
    • Hosts are maintained by Marian Z.
    • Get documentation into OSG or SAND folders
    • How to restart docker containers? Repo locations?

Existing items

  • Derek - RSV replacement
    • Now need to test and compare results
    • Shawn’s comments:
      • Start testing by polling sites with capable hardware: USCMS or ATLAS
      • Concerned about overloading pS instances on marginal hardware
      • Validation tools don’t exist yet, done manually with Kibana queries
      • Compare overlapping time ranges to confirm record counts match
      • OK to remove SOCKS support (no longer needed, as VM hosts are now on LHCONE)
  • Ilya and Nebraska - Meeting topics
    • Improve UNL ElasticSearch monitoring, alert when data rate is less than expected
    • Derek: Could data be transformed and inserted with Logstash, rather than directly with ElasticSearch?
    • Shawn: Wants alerts when pS instance stops reporting
  • Derek - Put together weekly summary email of Condor transfer data
  • Brian - Research on how to include TCP flow statistics for XRootD
  • Derek - Start email about acknowledgements for ps-collectors!
  • Monitoring
    • The pipelines and the contents of the pipelines
    • Will work with Ilya to monitor UNL ElasticSearch rates
    • SAND monitoring draft outlining Nagios probes for project services (both Nebraska and other sites), mostly based on architecture document
    • Brian emphasized knowing the alert destinations (email addresses, etc.)
    • Brian created alerts at group for notifications
    • Will follow up with Chicago and Michigan to fill in gaps
    • Nebraska will be responsible for the “plumbing” monitoring?
      • Confirmation of flow with test messages
  • Shawn: RabbitMQ authentication
    • Shawn discussed the possibility of a plugin functionality with the pS developers, and they seemed open to the idea.
    • Keeps them out of the business of storing credentials
    • pS team agrees it’s a reasonable feature request
    • Lead developer for this psconfig request is out of the office until second week in January
  • Expanding web presence
    • Shawn: Create github PR and Derek will review
    • Logo, both text and graphical
    • Pointers to existing OSG network docs
  • Complete documentation for each service component in architecture document
  • How to send check_mk status info from Nebraska to OSG perfSonar ETF?
    • LiveStatus API?