Uploaded image for project: 'Red Hat OpenStack Services on OpenShift'
  1. Red Hat OpenStack Services on OpenShift
  2. OSPRH-25831

cyborg virtualized nvme managemet dev env

XMLWordPrintable

    • Icon: Spike Spike
    • Resolution: Done
    • Icon: Normal Normal
    • None
    • None
    • openstack-cyborg
    • None
    • 5
    • False
    • Hide

      None

      Show
      None
    • False
    • rhos-workloads-evolution
    • Workload Evolution Sprint 18
    • 1

      The scope of this spike is to create a libvirt XML definition that can be used
      to emulate a compute node with multiple NVMe devices that can be managed and
      partitioned. To enable this functionality, the domain should be defined with
      a vIOMMU and utilize the q35 machine type to ensure proper PCIe bus handling.

      This involves adding multiple virtual NVMe controllers or emulated NVMe
      drives to the guest topology. The goal is to create a configuration that
      emulates a host environment where storage resources can be assigned,
      grouped, or isolated at the hardware level.

      The important part of this spike is to create a working domain XML that is
      capable of nested NVMe passthrough or assignment, documenting the setup,
      and testing it with a nested libvirt VM or DevStack if time permits.

      The intent is to enable the initial development work into extending Cyborg to
      support NVMe management and orchestration in the next sprint. This should be
      time-boxed to 4–5 days at most.

        1. nvme.xml
          10 kB
          Chandan Kumar

              rhn-engineering-chkumar Chandan Kumar
              smooney@redhat.com Sean Mooney
              rhos-workloads-evolution
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: