From 4d23ac58bb02bfb1407f8c822230005d407f622b Mon Sep 17 00:00:00 2001 From: akri3i Date: Thu, 23 Jul 2026 15:23:01 +0530 Subject: [PATCH] update dual stream stp for cnv-5.0 Signed-off-by: akri3i --- .../dual-stream-cluster-rhcos9-rhcos10/stp.md | 241 ++++++++++-------- 1 file changed, 131 insertions(+), 110 deletions(-) diff --git a/stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/stp.md b/stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/stp.md index 986494d0..d43668b3 100644 --- a/stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/stp.md +++ b/stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/stp.md @@ -1,21 +1,27 @@ # Openshift-virtualization-tests Test plan -## **Dual-Stream RHCOS Support (RHCOS9.8 + RHCOS10.2 Worker Nodes) — Quality Engineering Plan** +## **Dual-Stream RHCOS Support (RHCOS9.x + RHCOS10.x Worker Nodes) — Quality Engineering Plan** ### **Metadata & Tracking** -- **Enhancement(s):** https://redhat.atlassian.net/browse/VIRTSTRAT-83 +- **Enhancement(s):** https://issues.redhat.com/browse/VIRTSTRAT-83 - **Feature Tracking:** No separate VEP/HLD; tracked via VIRTSTRAT-83 -- **Epic Tracking:** https://redhat.atlassian.net/browse/CNV-49964 -- **QE Owner(s):** Akriti Gupta +- **Epic Tracking:** https://issues.redhat.com/browse/CNV-84749 +- **QE Owner(s):** Akriti Gupta (@akri3i) - **Readiness Tracking:** [RHCOS 10 Readiness Tracking](https://docs.google.com/spreadsheets/d/1mD1mVkiIqrdyaDIqB0fhRgmywWp4sI9OEy0IVBD6pDc/edit?gid=0#gid=0) - **Owning SIG:** sig-virt - **Participating SIGs:** sig-virt, sig-network, sig-storage, sig-iuo, sig-infra +- **Child STPs:** + - [sig-network](./network.md) — QE Owner: Asia Khromov (@azhivovk) + - [sig-storage](./storage.md) — QE Owner: Kate Shvaika (@kshvaika) + - sig-iuo — child STP not yet created + - sig-infra — child STP not yet created + - sig-virt — covered by this parent STP (no separate child STP) - **Target Release(s):** - DP: N/A - - TP: v4.22 — RHCOS 10.2 worker node support is Tech Preview - - GA: v5.0 — RHCOS 10.2 support GA + - TP: v4.22 — RHCOS 10.x worker node support is Tech Preview + - GA: v5.0 — RHCOS 10.x support GA > **This is a Parent STP.** It defines the overall dual-stream RHCOS testing strategy and cross-cutting > requirements. Each participating SIG is expected to create a child STP that extends this document with @@ -24,15 +30,15 @@ **Document Conventions:** -- **RHCOS9.8:** Red Hat CoreOS 9.8 worker nodes (GA, default for OCP 4.22). -- **RHCOS10.2:** Red Hat CoreOS 10.2 worker nodes (Tech Preview in 4.22, GA in 5.0). -- **Dual-stream cluster:** An OCP cluster running both RHCOS9.8 and RHCOS10.2 worker nodes simultaneously. +- **RHCOS9.x:** Red Hat CoreOS 9.x worker nodes (GA, default for OCP 4.22). +- **RHCOS10.x:** Red Hat CoreOS 10.x worker nodes (Tech Preview in 4.22, GA in 5.0). +- **Dual-stream cluster:** An OCP cluster running both RHCOS9.x and RHCOS10.x worker nodes simultaneously. ### **Feature Overview** -Starting with OCP 4.22, OpenShift Virtualization supports dual-stream clusters running both RHCOS 9.8 and RHCOS 10.2 worker nodes. +Starting with OCP 4.22, OpenShift Virtualization supports dual-stream clusters running both RHCOS 9.x and RHCOS 10.x worker nodes. This enables customers to gradually migrate their worker nodes to RHCOS 10 while maintaining VM workload availability through live migration across node types. -RHCOS 9.8 remains the default for OCP 4.22. +RHCOS 9.x remains the default for OCP 4.22. --- @@ -42,32 +48,36 @@ RHCOS 9.8 remains the default for OCP 4.22. - [x] **Review Requirements** - *Key requirements reviewed:* - - CNV 4.22: RHCOS9.8 is GA/default; RHCOS10.2 is Tech Preview. + - CNV 4.22: RHCOS9.x is GA/default; RHCOS10.x is Tech Preview. + - CNV 4.23: RHCOS9.x is GA/default; RHCOS10.x is Tech Preview. + - CNV 5.0: RHCOS10.x is GA/default; RHCOS9.x is also supported through CNV 5.0, CNV 5.1 and CNV 5.2 - OCPSTRAT-1150: Support dual-stream clusters to allow customers to run mixed hardware. VM live migration must succeed across - RHCOS9.8 ↔ RHCOS10.2 worker nodes. + RHCOS9.x ↔ RHCOS10.x worker nodes. - [x] **Understand Value and Customer Use Cases** - *Feature value to customers:* Customers can run mixed hardware in the same cluster — newer hardware that requires RHEL 10 and older hardware that only supports RHEL 9. - *Customer use cases:* - - As a cluster administrator, I want to add RHCOS10.2 worker nodes to my existing + - As a cluster administrator, I want to add RHCOS10.x worker nodes to my existing cluster. - - As a VM operator, I want to live-migrate VMs from RHCOS9.8 worker nodes to RHCOS10.2 worker + - As a VM operator, I want to live-migrate VMs from RHCOS9.x worker nodes to RHCOS10.x worker nodes. - - As a platform team, I want to validate that CNV behaves correctly on both RHCOS9.8 and RHCOS10.2 + - As a platform team, I want to validate that CNV behaves correctly on both RHCOS9.x and RHCOS10.x nodes. - [x] **Acceptance Criteria** - *Acceptance criteria:* - - All CNV features available on RHCOS 9.8 clusters work identically on RHCOS 10.2 clusters. - - All CNV features work identically on dual-stream clusters. - - A VM running on an RHCOS9.8 worker node can be live-migrated to an RHCOS10.2 worker node and - back without disruption. - - A VM running on an RHCOS10.2 worker node can be live-migrated to an RHCOS9.8 worker node and - back without disruption. - - CNV components deploy and report ready status on RHCOS 10.2 worker nodes with the same - behavior as on RHCOS 9.8 nodes. + - CNV features on dual-stream clusters behave as on single-OS clusters — validated by existing + Tier 1/2/3 suites; SIG-specific coverage documented in child STPs. + - A VM on an RHCOS9.x worker can be live-migrated to an RHCOS10.x worker and back: + VM remains in Running phase throughout migration, a guest workload started before + migration continues running after migration without restart, and the VM is accessible + on the target node — validated by CNV-81251 scenarios in Section III. + - A VM on an RHCOS10.x worker can be live-migrated to an RHCOS9.x worker and back: + VM remains in Running phase throughout migration, a guest workload started before + migration continues running after migration without restart, and the VM is accessible + on the target node — validated by CNV-81251 scenarios in Section III. - [x] **Non-Functional Requirements (NFRs)** - *Applicable NFRs:* @@ -76,8 +86,8 @@ RHCOS 9.8 remains the default for OCP 4.22. - **Monitoring/Observability:** N/A — no new alerts or metrics introduced; existing CNV monitoring applies unchanged. - **Scalability:** N/A — no new scale requirements; existing cluster-level live migration parallelism limits apply. - **UI:** N/A — no UI code changes introduced; UI testing adds no customer value for this feature. - - **Documentation:** Release notes must document RHCOS10.2 Tech Preview status in 4.22 and GA status timeline. - - **Compatibility:** N/A — validated via Tier 1/2/3 runs on both RHCOS9.8 and RHCOS10.2. + - **Documentation:** Release notes must document RHCOS10.x Tech Preview status in 4.22 and GA status timeline. + - **Compatibility:** N/A — validated via Tier 1/2/3 runs on both RHCOS9.x and RHCOS10.x. #### **2. Technology and Design Review** @@ -90,15 +100,17 @@ RHCOS 9.8 remains the default for OCP 4.22. - [x] **Technology Challenges** - *Identified challenges:* - - **CNV compatibility on RHCOS 10.2:** RHCOS 10.2 is a new platform configuration that may + - **CNV compatibility on RHCOS 10.x:** RHCOS 10.x is a new platform configuration that may surface unexpected failures. Tier 1, Tier 2, and Tier 3 testing is the primary mechanism for finding these issues. - - **Dual-stream cluster provisioning:** Clusters with mixed RHCOS9.8 and RHCOS10.2 worker - nodes require specific provisioning tooling. DevOps QE to provide this capability. + - **Dual-stream cluster provisioning:** Clusters with mixed RHCOS9.x and RHCOS10.x worker + nodes require specific provisioning tooling. QE DevOps has provided this capability + and it is validated and available for test use. - *Impact on testing approach:* - - Tier 1, Tier 2, and Tier 3 testing on an RHCOS10.2-only cluster is mandatory for all component - teams — this is the primary testing vehicle. + - Tier 1, Tier 2, and Tier 3 testing on an RHCOS10.x-only cluster is default in CNV-5.0. + - Since RHCOS9.x is still supported in 5.0, testing should also cover RHCOS9.x-only clusters; + component teams decide which tests to run. - Each participating SIG decides which tests to run on dual-stream clusters based on their feature area requirements. - Live migration across RHCOS versions must be tested explicitly. @@ -112,19 +124,22 @@ RHCOS 9.8 remains the default for OCP 4.22. - [x] **Topology Considerations** - *Topology requirements:* - - FOR RHCOS 10.2: + - FOR RHCOS 10.x: - Either high-availability (HA) or a compact cluster - - Requires both control plane and worker nodes to be on RHCOS 10.2 + - Requires both control plane and worker nodes to be on RHCOS 10.x + - FOR RHCOS 9.x: + - Either high-availability (HA) or a compact cluster + - Requires both control plane and worker nodes to be on RHCOS 9.x - For OCPSTRAT-1150: - Dual-stream testing requires only a high-availability (HA) - cluster — at least 1 worker running RHCOS10.2 alongside RHCOS9.8 workers + cluster — at least 1 worker running RHCOS10.x alongside RHCOS9.x workers within the same OCP cluster. --- #### **3. Known Limitations** -- **CNV on RHCOS 10.2 worker nodes uses the same software stack as on RHCOS 9.8 +- **CNV on RHCOS 10.x worker nodes uses the same software stack as on RHCOS 9.x through OCP 5.2.** A fully RHCOS 10-native CNV stack is planned for OCP 5.3 and is out of scope for this STP. - *Sign-off:* Martin Tessun / 2026-05-13 @@ -137,14 +152,13 @@ RHCOS 9.8 remains the default for OCP 4.22. **Testing Goals** -- **[P0]** As a cluster admin, I can run all supported VM workloads on an RHCOS 10.2-only cluster - with the same behavior as on RHCOS 9.8. -- **[P1]** As a cluster admin, I can operate a dual-stream cluster (RHCOS 9.8 + RHCOS 10.2 worker +- **[P1]** As a cluster admin, I can operate a dual-stream cluster (RHCOS 9.x + RHCOS 10.x worker nodes) and all supported VM workloads function correctly on both node types. -- **[P1]** As a VM operator, I can live-migrate a VM from an RHCOS 9.8 worker node to an RHCOS 10.2 - worker node and back without disruption, and vice versa. -- **[P1]** As a cluster admin, CNV components deploy and report ready on RHCOS 10.2 worker nodes - with the same behavior as on RHCOS 9.8 nodes. +- **[P1]** As a VM operator, I can live-migrate a VM from an RHCOS 9.x worker node to an RHCOS 10.x + worker node and back (and vice versa): the VM remains in Running phase throughout migration, + a guest workload started before migration continues running after migration without restart, + and the VM is accessible on the target node. + **Out of Scope (Testing Scope Exclusions)** - **Fully RHCOS 10-native CNV stack** @@ -155,10 +169,8 @@ RHCOS 9.8 remains the default for OCP 4.22. **Test Limitations** -- **Dual-stream cluster provisioning depends on QE DevOps tooling.** The ability to deploy - a cluster with mixed RHCOS9.8 and RHCOS10.2 worker nodes relies on tooling provided by the - QE DevOps team. If this tooling is unavailable or unstable, the dual-stream migration - scenarios cannot be executed. +- None — dual-stream cluster provisioning tooling from QE DevOps is validated and available + (see Dependencies in Section II.2 and Entry Criteria in Section II.4). - *Sign-off:* Martin Tessun / 2026-05-13 @@ -167,29 +179,34 @@ RHCOS 9.8 remains the default for OCP 4.22. **Functional** - [x] **Functional Testing** - - Validates that the full CNV feature set operates correctly on RHCOS10.2-only clusters + - Validates that the full CNV feature set operates correctly on RHCOS10.x-only clusters - *Details:* The primary mechanism is running existing Tier1, Tier 2 and Tier 3 test suites and triaging failures. - All other SIGs (sig-network, sig-storage, sig-iuo, sig-infra) must document their Tier1, Tier 2 and Tier 3 Test results and bugs in their own Jira Stories. - - For OCPSTRAT-1150: Validates that the full CNV feature set operates correctly on dual-stream clusters, with el9.8 userspace. - - VM live migration must succeed across RHCOS9.8 ↔ RHCOS10.2 worker nodes. - - Successful LiveMigration of VM from RHCOS9.8 to RHCOS10.2 worker node and backto RHCOS9.8 - i.e VM created first on RHCOS9.8 Worker Node - - Successful LiveMigration of VM from RHCOS10.2 to RHCOS9.8 worker node and backto RHCOS10.2 - i.e VM created first on RHCOS10.2 Worker Node - -- [ ] **Automation Testing** - - For RHCOS 10.2: No new automation tests needed — existing Tier 1, Tier 2 and Tier 3 suites + - For OCPSTRAT-1150: Validates that the full CNV feature set operates correctly on dual-stream clusters. + - VM live migration must succeed across RHCOS9.x ↔ RHCOS10.x worker nodes. + - Successful LiveMigration of VM from RHCOS9.x to RHCOS10.x worker node and backto RHCOS9.x + i.e VM created first on RHCOS9.x Worker Node + - Successful LiveMigration of VM from RHCOS10.x to RHCOS9.x worker node and backto RHCOS10.x + i.e VM created first on RHCOS10.x Worker Node + +- [x] **Automation Testing** + - For RHCOS 10.x: No new automation tests needed — existing Tier 1, Tier 2, and Tier 3 suites are run as-is. - - For OCPSTRAT-1150 (dual-stream live migration): Manual ad-hoc runs for 4.22 Tech Preview; - node-affinity-based automation in Tier 2 by 5.0 GA. - -- [x] **Regression Testing** — sig-virt and all participating SIGs must run Tier1, Tier 2 and Tier 3 - regression - - On RHCOS10.2-only cluster. - - *Details:* The strategy relies on running existing Tier 1, Tier 2 and Tier 3 test suites against RHCOS 10.2 + - For OCPSTRAT-1150 (dual-stream live migration): Automated tests being added by component teams. + +- [x] **Regression Testing** — sig-virt and all participating SIGs must run Tier 1, Tier 2, and + Tier 3 regression. + - On RHCOS10.x-only cluster: + - *Details:* Run existing Tier 1, Tier 2, and Tier 3 test suites against RHCOS 10.x. + - On RHCOS9.x-only cluster: + - *Details:* Run existing Tier 1, Tier 2, and Tier 3 test suites against RHCOS 9.x. - For OCPSTRAT-1150: See Automation Testing above for the decision and rationale. +- [ ] **Self-Validation Testing** + - *Details:* N/A. RHCOS 10.x-only and dual-stream scenarios require specialized cluster + configurations not available in self-validation environments. + **Non-Functional** - [x] **Performance Testing** — N/A @@ -204,30 +221,31 @@ RHCOS 9.8 remains the default for OCP 4.22. **Integration & Compatibility** -- [x] **Compatibility Testing** — CNV must remain compatible with both RHCOS9.8 and RHCOS10.2 - across the supported version range (4.22 through 5.2). - - *Details:* Compatibility is validated through Tier 1, Tier 2 and Tier 3 runs on both RHCOS9.8 and RHCOS10.2 - clusters. Component teams validate their specific feature areas in child STPs. +- [x] **Compatibility Testing** — CNV must remain compatible with both RHCOS9.x and RHCOS10.x. + - *Details:* Compatibility is validated through Tier 1, Tier 2, and Tier 3 runs on both RHCOS9.x + and RHCOS10.x clusters. This STP covers OCP/CNV 5.0. Component teams validate their specific + feature areas in child STPs. -- [x] **Upgrade Testing** — Out of scope for this 4.22 STP. - - *Details:* Upgrade testing (4.22 → 5.0, EUS-to-EUS, etc.) will be covered in the 5.0 . - Not planned as part of 4.22 testing. +- [ ] **Upgrade Testing** — Out of scope for this STP. + - *Details:* Relevant only from 5.1.0 (see Out of Scope). + - *PM/Lead Agreement:* Martin Tessun / 2026-05-13 -- [x] **Dependencies** — Testing is blocked on specific deliverables from other teams. - - *Details:* QE DevOps team must provide a stable dual-stream cluster provisioning option - (RHCOS9.8 + RHCOS10.2 workers in the same cluster). +- [x] **Dependencies** — Resolved. + - *Details:* QE DevOps team has provided dual-stream cluster provisioning + (RHCOS9.x + RHCOS10.x workers in the same cluster). -- [x] **Cross Integrations** — Other SIGs must create a child STP extending this one to cover adhoc testing for OCPSTRAT-1150. +- [x] **Cross Integrations** — Other SIGs must create a child STP extending this one to cover testing for OCPSTRAT-1150. - *Details:* sig-network, sig-storage, sig-iuo, and sig-infra must each create child STPs - specifying their Tier 1, Tier 2 and Tier 3 test coverage on RHCOS10.2 nodes and dual-stream + specifying their Tier 1, Tier 2 and Tier 3 test coverage on RHCOS10.x nodes and dual-stream clusters. Each SIG is responsible for triaging failures in their area and filing bugs with clear RHCOS-version attribution. **Infrastructure** - [x] **Cloud Testing** - — For 4.22 Tech Preview we must test with FIPS enabled. - - Use cloud setup if it supports FIPS enabled + - *Details:* FIPS is required. Cloud platforms (e.g., PSI, IBM-BM) may be used when + FIPS-enabled RHCOS 10.x is supported. [OCPBUGS-64660](https://issues.redhat.com/browse/OCPBUGS-64660) + is fixed. #### **3. Test Environment** @@ -235,12 +253,14 @@ RHCOS 9.8 remains the default for OCP 4.22. - **Cluster Topology:** - **Dual-stream testing:** High-availability (HA) bare-metal cluster required — - 3-control-plane / 3-worker minimum, with at least 1 worker running RHCOS10.2 alongside - RHCOS9.8 workers. SNO or compact clusters are not supported for dual-stream testing. - - **RHCOS10.2-only testing:** Standard 3-control-plane / 3-worker bare-metal cluster with all - workers on RHCOS10.2. + 3-control-plane / 3-worker minimum, with at least 1 worker running RHCOS10.x alongside + RHCOS9.x workers. SNO or compact clusters are not supported for dual-stream testing. + - **RHCOS10.x-only testing:** Standard 3-control-plane / 3-worker bare-metal cluster with all + workers on RHCOS10.x. + - **RHCOS9.x-only testing:** Standard 3-control-plane / 3-worker bare-metal cluster with all + workers on RHCOS9.x. -- **OCP & OpenShift Virtualization Version(s):** OCP 4.22 with CNV 4.22 +- **OCP & OpenShift Virtualization Version(s):** OCP 5.0 with CNV 5.0 - **CPU Virtualization:** Standard (VT-x / AMD-V enabled) @@ -254,20 +274,22 @@ RHCOS 9.8 remains the default for OCP 4.22. - **Required Operators:** Standard. -- **Platform:** Bare metal (dual-stream cluster provisioned by DevOps QE tooling). +- **Platform:** Bare metal (primary). Cloud platforms with FIPS enabled are also + supported now that [OCPBUGS-64660](https://issues.redhat.com/browse/OCPBUGS-64660) is fixed. + Dual-stream cluster provisioned by QE DevOps tooling. -- **Special Configurations:** Dual-stream cluster. DevOps QE - provides tooling to deploy this configuration. +- **Special Configurations:** Dual-stream cluster with FIPS enabled. QE DevOps dual-stream + provisioning tooling is validated and available to deploy this configuration. #### **3.1. Testing Tools & Frameworks** - **Test Framework:** - - Standard (openshift-virtualization-tests) for RHCOS 10.2 + - Standard (openshift-virtualization-tests) for RHCOS 10.x - For OCPSTRAT-1150 (dual-stream): See Automation Testing in Section II.2 for the decision and rationale. - **CI/CD:** Two cluster configurations are required, both available from CNV 4.22: - - **RHCOS10.2-only cluster:** Existing Tier 1, Tier 2 and Tier 3 jobs run by all component teams. - - **Dual-stream cluster :** Manual ad-hoc live migration runs by component teams. + - **RHCOS10.x-only cluster:** Existing Tier 1, Tier 2 and Tier 3 jobs run by all component teams. + - **Dual-stream cluster:** Automated tests by component teams. - **Other Tools:** N/A @@ -279,8 +301,8 @@ The following conditions must be met before testing can begin: (recommendation doc sign-offs from PM, Engineering, Platform, Product Ops confirmed) - [x] Test environment can be **set up and configured** (dual-stream cluster available via QE DevOps tooling) -- [x] **CNV 4.22 builds based on RHCOS 9.8 are available and validated** - [x] QE DevOps dual-stream cluster provisioning is validated and available +- [x] CNV 5.0 defaults to RHCOS 10.x worker nodes #### **5. Risks** @@ -302,15 +324,10 @@ The following conditions must be met before testing can begin: - **Risk:** - HA resource shortage. - RDU2 to RDU3 migration, bare metal cluster outages. - - Deploying RHCOS 10.2 and Dual-Stream cluster on cloud fails (PSI, IBM-BM or other clouds with FIPS fails). - - QE DevOps dual-stream cluster provisioning tooling may be unavailable or unstable, blocking - Tier 1, Tier 2 and Tier 3 runs. - **Mitigation:** - - Use bare metal cluster with FIPS enabled. - - Engage QE DevOps team early to confirm dual-stream cluster availability timeline. Identify a - fallback of manually provisioning a mixed-node cluster if tooling is delayed. Track - provisioning readiness as an entry criterion. - - *Missing or unavailable environments:* Dual-stream cluster if QE DevOps tooling is not ready. + - Prefer bare metal with FIPS enabled as the primary environment; use FIPS-enabled + cloud (PSI, IBM-BM) as an additional option now that OCPBUGS-64660 is fixed. + - *Missing or unavailable environments:* N/A for FIPS cloud (OCPBUGS-64660 fixed). - *Sign-off:* Martin Tessun / 2026-05-13 **Untestable Aspects** @@ -321,11 +338,9 @@ The following conditions must be met before testing can begin: **Resource Constraints** -- **Risk:** Manual ad-hoc testing for dual-stream live migration scenarios (OCPSTRAT-1150) requires - component team bandwidth through OCP 5.2, until Tier 2 automation (Option 3) is completed. - - **Mitigation:** Each participating SIG allocates time for ad-hoc runs in their test cycle. - Tier 2 automation will be added progressively. See Automation Testing in Section II.2 for the - full options analysis and decision rationale. +- **Risk:** None identified — automation is being implemented for dual-stream live migration + scenarios (OCPSTRAT-1150). + - **Mitigation:** N/A; component teams are adding automated tests for dual-stream clusters. - *Missing resources or infrastructure:* N/A - *Sign-off:* Martin Tessun / 2026-05-13 @@ -345,18 +360,24 @@ The following conditions must be met before testing can begin: ### **III. Test Scenarios & Traceability** -- **[CNV-81251](https://redhat.atlassian.net/browse/CNV-81251)** — As a VM operator, I want to live-migrate VMs between RHCOS9.8 and - RHCOS10.2 worker nodes within the same cluster so my workloads remain available during +- **[CNV-81251](https://issues.redhat.com/browse/CNV-81251)** — As a VM operator, I want to live-migrate VMs between RHCOS9.x and + RHCOS10.x worker nodes within the same cluster so my workloads remain available during node maintenance. - - *Test Scenario:* [Tier 2] **Scenario 1** — VM created on an RHCOS9.8 worker node is live-migrated - to an RHCOS10.2 worker node without disruption; then migrate back to RHCOS9.8 without disruption. + - *Test Scenario:* [Tier 2] **Scenario 1** — VM created on an RHCOS9.x worker node is + live-migrated to an RHCOS10.x worker node, then back to RHCOS9.x. For each migration: + the VM remains in Running phase throughout migration, a guest workload started before + migration continues running after migration without restart, and the VM is accessible + on the target node. - *Priority:* P1 -- **[CNV-81251](https://redhat.atlassian.net/browse/CNV-81251)** — As a VM operator, I want to live-migrate VMs between RHCOS9.8 and - RHCOS10.2 worker nodes within the same cluster so my workloads remain available during +- **[CNV-81251](https://issues.redhat.com/browse/CNV-81251)** — As a VM operator, I want to live-migrate VMs between RHCOS9.x and + RHCOS10.x worker nodes within the same cluster so my workloads remain available during node maintenance. - - *Test Scenario:* [Tier 2] **Scenario 2** — VM created on an RHCOS10.2 worker node is live-migrated - to an RHCOS9.8 worker node without disruption; then migrate back to RHCOS10.2 without disruption. + - *Test Scenario:* [Tier 2] **Scenario 2** — VM created on an RHCOS10.x worker node is + live-migrated to an RHCOS9.x worker node, then back to RHCOS10.x. For each migration: + the VM remains in Running phase throughout migration, a guest workload started before + migration continues running after migration without restart, and the VM is accessible + on the target node. - *Priority:* P1 ---