AWS EFS CSI Driver Operator inadvertently upgraded to OCP 4.18 version on OCP 4.12-4.17 clusters

Solution Verified - Updated

Environment

  • Red Hat OpenShift on AWS (ROSA)
    • < 4.18

Issue

The version of the AWS EFS CSI Driver Operator was updated to 4.18.x on an older (4.12-4.17) OpenShift Container Platform (OCP) cluster.

Resolution

This article specifically addresses how to:

  1. Verify if your AWS EFS CSI Driver was affected by this specific window of misconfiguration.
  2. Roll back or recover the driver to the version appropriate for your OCP release.

1. Identify the Installed Operator Version:

To determine if your cluster is running the incorrect 4.18 version of the AWS EFS CSI Driver Operator, list the ClusterServiceVersions (CSV) in the relevant namespace:

oc -n openshift-cluster-csi-drivers get csv

Example Output:
If the cluster is impacted, the VERSION column will display 4.18.x despite the cluster being on a lower version (e.g., 4.12–4.17):

NAME                                               DISPLAY                       VERSION               REPLACES                                          PHASE
aws-efs-csi-driver-operator.v4.18.0-202602040643   AWS EFS CSI Driver Operator   4.18.0-202602040643   aws-efs-csi-driver-operator.4.12.0-202602021306   Succeeded

Note: If the operator version matches your OpenShift Container Platform (OCP) version (e.g., an OCP 4.12 cluster running Operator 4.12.x), no further action is required.

2. Downgrade the AWS EFS CSI Driver Operator

Follow these steps to remove the incorrect version and trigger a fresh installation of the supported version.

a. Delete the incorrect CSV

Replace <CSV_NAME> with the actual name found in the diagnostic step (e.g., aws-efs-csi-driver-operator.v4.18.0-202602040643).

oc -n openshift-cluster-csi-drivers delete csv CSV_NAME

Example Output:

clusterserviceversion.operators.coreos.com "CSV_NAME" deleted

b. Remove the existing Subscription

oc -n openshift-cluster-csi-drivers delete subscription aws-efs-csi-driver-operator

Example Output:

subscription.operators.coreos.com "aws-efs-csi-driver-operator" deleted

c. Recreate the Subscription

Apply the following manifest to recreate the subscription.

oc apply -f - <<EOF
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
 name: aws-efs-csi-driver-operator
 namespace: openshift-cluster-csi-drivers
spec:
 channel: stable
 installPlanApproval: Manual
 name: aws-efs-csi-driver-operator
 source: redhat-operators
 sourceNamespace: openshift-marketplace
EOF

Example Output:

subscription.operators.coreos.com/aws-efs-csi-driver-operator created

d. Approve the InstallPlan

Since the subscription is set to Manual, you must approve the generated InstallPlan to complete the installation.

First, find the InstallPlan name:

oc -n openshift-cluster-csi-drivers get installplan

Example Output:

NAME            CSV                                               APPROVAL   APPROVED
install-p9v28   aws-efs-csi-driver-operator.v4.12.0-202602021306  Manual     false

IMPORTANT: Before proceeding, verify that the version listed under the CSV column in the output above matches your cluster's OCP version (e.g., if you are on OCP 4.12, the CSV should start with aws-efs-csi-driver-operator.v4.12).

Patch the InstallPlan to approve it (replace INSTALL_PLAN_NAME with your result):

oc -n openshift-cluster-csi-drivers patch installplan <INSTALL_PLAN_NAME> --type='merge' -p '{"spec":{"approved":true}}'

Example Output:

installplan.operators.coreos.com/install-p9v28 patched

Verification

Confirm that the new CSV matches your cluster's OCP version. For example, on an OCP 4.12 cluster, the version should now start with 4.12.

oc -n openshift-cluster-csi-drivers get csv

Root Cause

Between 2026-02-03 21:16 UTC and 2026-02-04 05:18 UTC, a catalog release error caused OCP 4.18 operator content to be pushed to OCP 4.12–4.17 clusters.

Clusters with subscriptions set to Automatic approval may have inadvertently upgraded the AWS EFS CSI Driver Operator to a version intended for OCP 4.18.

For more comprehensive details on the global catalog incident, please refer to the primary advisory: Red Hat Knowledgebase Solution 7137887.

SBR
Product(s)
Components
Category
Tags

This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.