Skip to content

runcdeprecationcases.md

Latest commit

 

History

History
104 lines (78 loc) · 5.03 KB

runcdeprecationcases.md

File metadata and controls

104 lines (78 loc) · 5.03 KB
Raw

Test Plan: runc Deprecation Cases

Metadata

Field Value
Test file test/extended/node/runcdeprecationcases.go
Package node
Test suite [Jira:Node][sig-node] runc deprecation cases
Feature OCPSTRAT-3154 — runc deprecation warning in OCP 5.0 for clusters upgrading from 4.22
Epic OCPNODE-4013 — Block upgrade from RHCOS 9 to RHCOS 10 when runc is in use
User Story OCPNODE-4567
Assignee Aditi Sahay (asahay@redhat.com)
Component sig-node, MCO, CRI-O
Test type E2E / functional
Ginkgo label [Jira:Node][sig-node]
Disruptive No
Requires reboot No

UseCase3: Fresh Install on RHCOS 9 with crun

Description

A freshly installed OpenShift cluster on RHCOS 9 uses crun as the default container runtime. This is the standard out-of-the-box configuration — no custom ContainerRuntimeConfig should be present, and crun must be the CRI-O default on all worker nodes.

Scope

Item Value
OpenShift versions 5.0, 5.1, 5.2 (fresh install)
OS RHCOS 9 (VARIANT_ID=coreos, VERSION_ID=9.x)
Default runtime crun
ContainerRuntimeConfig None (no custom CRC expected)
Skip conditions MicroShift, Hypershift (external control plane)

Test Function

g.It("RHCOS 9 cluster install should use crun as the default container runtime")

What This Test Validates

This test verifies that a fresh RHCOS 9 cluster has crun configured as the default container runtime with no custom overrides, and that the cluster is fully healthy. It performs the following validations in sequence:

  1. Cluster health — Confirms the ClusterVersion is Available=True and Progressing=False, indicating the cluster is stable before any further checks.

  2. RHCOS 9 OS identity — Selects a worker node and reads /etc/os-release via oc debug to confirm VARIANT_ID=coreos (confirms RHCOS, not plain RHEL) and VERSION_ID=9.x (confirms RHCOS 9 major version).

  3. No custom ContainerRuntimeConfig — Lists all ContainerRuntimeConfig objects cluster-wide and asserts the list is empty. This confirms no administrator-applied runtime override is present, which is the expected state on a fresh install.

  4. CRI-O default runtime is crun — Reads the effective CRI-O configuration on the worker node (crio status config) and asserts default_runtime = "crun". This is the authoritative on-node check that crun is active, independent of the API-level CRC check above.

  5. machine-config ClusterOperator health — Verifies the machine-config ClusterOperator conditions: Available=True, Progressing=False, Degraded=False. A degraded MCO could indicate the runc guard (from MCO PR #5891) has incorrectly fired on a valid crun cluster.

  6. Functional workload execution — Creates a test pod pinned to the inspected worker node using the framework shell image and waits for it to reach Succeeded phase. This confirms the crun runtime can successfully pull and execute containers end-to-end.

This test is read-only and non-disruptive — it does not create or modify any MachineConfig, ContainerRuntimeConfig, or node configuration.

Pass/Fail Criteria

Check Pass condition
ClusterVersion Available AVAILABLE=True, PROGRESSING=False
RHCOS version on worker node VARIANT_ID=coreos and VERSION_ID="9.x"
No ContainerRuntimeConfig List is empty (no custom CRC present)
CRI-O default runtime default_runtime = "crun" in crio status config
machine-config ClusterOperator healthy Available=True, Progressing=False, Degraded=False
Test workload runs successfully Pod reaches Succeeded phase on the inspected node

Related Links