After upgrade 1.3.x->1.8.x ownerReference points to obsoleted ReplicaSet

[evgkrsk]

What steps did you take and what happened:
I have upgraded lvm-localpv to current latest release (1.8.0).

All pods is fine (running + ready), but there is endless errors in controller log:

[12.01.2026, 11:13:48,819 PM] lvm-provisioner-lvm-localpv-controller-768f47467f-bqj2q csi-provisioner E0112 16:10:31.951647       1 capacity.go:551] "Unhandled Error" err="update CSIStorageCapacity for {segment:0xc000c00168 storageClassName:redis-data}: CSIStorageCapacity.storage.k8s.io \"csisc-cq6n5\" is invalid: metadata.ownerReferences: Invalid value: []v1.OwnerReference{v1.OwnerReference{APIVersion:\"apps/v1\", Kind:\"ReplicaSet\", Name:\"lvm-provisioner-lvm-localpv-controller-74d8dd6565\", UID:\"019fb1f3-bf2f-453a-9c24-e531aae9b657\", Controller:(*bool)(0xc033f801d9), BlockOwnerDeletion:(*bool)(nil)}, v1.OwnerReference{APIVersion:\"apps/v1\", Kind:\"ReplicaSet\", Name:\"lvm-provisioner-lvm-localpv-controller-768f47467f\", UID:\"d4b48962-2cea-4db0-9255-ca209b09a3ee\", Controller:(*bool)(0xc033f801da), BlockOwnerDeletion:(*bool)(nil)}}: Only one reference can have Controller set to true. Found \"true\" in references for ReplicaSet/lvm-provisioner-lvm-localpv-controller-74d8dd6565 and ReplicaSet/lvm-provisioner-lvm-localpv-controller-768f47467f" logger="UnhandledError"
[12.01.2026, 11:13:48,819 PM] lvm-provisioner-lvm-localpv-controller-768f47467f-bqj2q csi-provisioner W0112 16:10:31.952776       1 capacity.go:552] Retrying capacity.workItem{segment:(*topology.Segment)(0xc000c00168), storageClassName:"redis-data"} after 10 failures

for all nodes in cluster.

What did you expect to happen:

Clean logs.

Anything else you would like to add:

Workaround is to regenerate all CSIStorageCapacity from scratch with:

kubectl delete CSIStorageCapacity --all

Environment:

• LVM Driver version: 1.8.0
• Kubernetes version (use kubectl version): v1.34.1
• Kubernetes installer & version: kubeadm, v1.33.7
• Cloud provider or hardware configuration: Bare metal
• OS (e.g. from /etc/os-release): Ubuntu 22.04

[abhilashshetty04]

@evgkrsk , I think its because old pod/replicaset is not scaled down. So, now there are 2 pods trying to update CSIStorageCapacity.

How did you upgrade the localpv-lvm to latest?
cc: @niladrih , Do we handle these in upgrade logic?

[evgkrsk]

We use qbec to in-place patch previous version of helm chart render to new version. As result, default deployment controller in KCM create new replicaset and step-by-step scale down old RS as new RS scaled up.

Forget to note that these messages in log appers after successful rollout of new RS, no old pods present.