Attendees

Agenda

Minutes

  • Notes from documentation review
    • Release notes:

      • rewrite change details to make it understandable to anyone

      • add list of major corrections

      • complete known issues

      • open discussion about release pipeline (like tagging component repositories)

    • Installation guide:

      • diagram about cluster nodes (controller, worker nodes) with difference about nodes

      • describe cluster functions (controller vs. worker nodes)

    • Testing document

      • more description about executed tests

    • Gerrit Code overview:

      • hyperlinks to gerrit projects

  • We discussed the possibility of a recurring reservation of the UNH community pod to run CI testing of REC on ARM driven by the LF Jenkins
  • Virtual media issue on ThunderX seems to have been solved. We need to identify what the solution is (e.g. was it code that was pushed or instructions or something else)
  • Vacation schedule
    • Paul Carver to be on vacation Nov 28, Dec 19, Dec 26. If someone else wants to host the this meeting, please volunteer and use the Zoom bridge and post meeting minutes. Otherwise I will cancel the meeting on those dates.
  • Build 237 is a REC Release Candidate for Akraino Release 2. There might be an issue on Dell, we aren't sure yet. The CD system is running our reference hardware Open Edge as described in Radio Edge Cloud Validation Lab, so as long as that is passing all tests we plan to declare build 237 as the Release 2. If we can identify a specific bug on Dell and fix it in another RC before the TSC imposed cutoff date of November 30th we will, but if not we will use build 237. We don't have the option of changing the release cutoff date.
  • Much discussion of the RIC. Paul suggested that much of it could be better discussed in the RIC meeting https://wiki.o-ran-sc.org/display/RICP/Project+meetings
  • Discussion of the possibility of running REC on VMs or on a single (non-clustered) bare metal server. It might work, but is not a target of the REC blueprint. Anyone who wants to try it should report back their results. Whether it would work on VMs probably depends on whether there is a hypervisor that can emulate IPMI because the REC/TA cluster deployment automation depends on IPMI to boot cluster nodes off of the #1 controller node.
  • Discussion of the possibility of using the TA blueprint family to run a different RIC, i.e. not the O-RAN Software Community's non-realtime RIC.
  • No labels