Uploaded image for project: 'Red Hat OpenStack Services on OpenShift'
  1. Red Hat OpenStack Services on OpenShift
  2. OSPRH-1419

RHOSP 17.1 to RHOSO 18.0 Upgrade Using Adoption Mechanism with Live Workloads and Compute Node Rotation using Optimize Service (Watcher)

XMLWordPrintable

    • Icon: Epic Epic
    • Resolution: Unresolved
    • Icon: Critical Critical
    • rhos-18.0.14 FR 4
    • None
    • None
    • None
    • RHOSP 17.1 to 18 Adoption in High Density node situation (+250 compute nodes)
    • False
    • Hide

      None

      Show
      None
    • False
    • Committed
    • Proposed
    • To Do
    • Committed
    • No impact
    • 0% To Do, 100% In Progress, 0% Done

      Summary
      Test the end-to-end upgrade process from RHOSP 17.1 to RHOSO 18.0 using the Adoption mechanism, while maintaining workload continuity.
      The test includes:

      • Using a buffer pool of unused compute nodes.
      • Migrating workloads via the Optimize Service (Watcher) by emptying nodes and migrating to unused compute nodes.
      • Updating empty compute nodes from RHEL 9.2 to RHEL 9.4

      The goal is to validate this process and gather data for field enablement and customer guidance.

      Detailed test plan described here

      Goal
      Validate a repeatable and non-disruptive upgrade process that:

      • Maintains active workloads without downtime
      • Leverages Optimize Service to evacuate compute nodes
      • Supports rolling OS updates of compute nodes
      • Produces documentation and performance metrics to guide future adoption scenarios

      Deliverables / Outcomes

      Operational Documentation:

      • Step-by-step guide for RHOSP 17.1 to RHOSO 18.0 Adoption with workloads
      • Instructions for using Optimize Service (Watcher) to migrate workloads off compute nodes after the upgrade/adoption has completed successfully
      • Guide for upgrading compute nodes from RHEL 9.2 to RHEL 9.4 after evacuation done by Watcher
      • Notes on failure handling, automation opportunities, and known caveats

      Timing Metrics:

      • Time for Control Plane Service Deployment during Adoption
      • Time for Database Copy during Adoption
      • Time for Control Plane Adoption
      • Time for Data Plane Adoption
      • Time to migrate workloads from M compute nodes to N buffer nodes using Optimize Service
      • Time to update compute nodes from RHEL 9.2 to RHEL 9.4 (for M compute nodes per cycle)

      Documentation:

      • Central location aggregating documentation and timing results
      • Lessons learned and recommendations
      • Links to automation scripts or tooling used

              rhn-support-jraju Jaison Raju
              pnavarro@redhat.com Pedro Navarro Perez
              rhos-dfg-perfscale
              Votes:
              0 Vote for this issue
              Watchers:
              15 Start watching this issue

                Created:
                Updated: