Present: Barry Blumenfeld, Alastair Dewhurst, Alessandro Di Girolamo, Dave Dykstra, Luis Linares, Stefan Roiser
Questions dealt with in this meeting:
  1. Should the GOCDB/OIM squid entry include anything about the application type (that is, Frontier or CVMFS) or not? (This is question #3 on our Questions page). Agreed: don't distinguish the application type in GOCDB/OIM. Perhaps in the future we will take advantage of per-service attributes in GOCDB/OIM to distinguish application type.
  2. Should the GOCDB/OIM squid entry include the squid service port or not? This has no impact on monitoring, so technically it is outside the scope of this task force. However, we are defining the GOCDB/OIM entries now so we need to decide. It is natural to include a port number in the OIM service url, and some services in GOCDB also use a url field that includes a port number. So the OIM entry will include the port number, and for GOCDB if the service is not using the standard port 3128, the url field should be filled in with that port number. Note that means that for GOCDB the service name and url name might be different; the url then will have to have the DNS name, and the service name won't have to. If the url field isn't filled in, the service name has to be the DNS name.
  3. On the monitoring server, should the Monitoring View be stored only as differences from the Information System View or as the entire Monitoring view? This is related to the questions of whether or not the Monitoring View should appear in AGIS, and if so, whether it should be the ultimate source of the information for ATLAS or just an import of the data from the monitoring server. Agreed: Initially only the differences will be stored in a file that is hand-maintained by VO monitoring personnel on the monitoring server, and that will be used to generate another intermediate-format file that is combined with the squid information from GOCDB/OIM which then feeds in to the MRTG configurator. The reason for the second file is that ATLAS may instead later generate it from AGIS so it may be updated by non-expert shifters.
  4. What should be in the initial squid service SAM test(s)? Agreed: it will be based only on the MRTG monitor. The squid monitoring server will not report that a squid is down until it has missed a few probes in a row (because of the unreliability of the UDP probes). It will push out the up/down information to a message broker in the SAM system, with a percentage of the squids in each service that are up.

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2013-02-04 - DaveDykstra
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback