Unnecessary CSI migration errors in vSphere Syncer in multi-VC clusterss

[gnufied]

We are noticing vsphere-syncer spamming following log messages, when K8s cluster spans multiple-VCs and have CSI migration disabled:

sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer.initVolumeMigrationService
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/pkg/syncer/util.go:366
sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer.podAdded
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/pkg/syncer/metadatasyncer.go:1852
sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer.InitMetadataSyncer.func7
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/pkg/syncer/metadatasyncer.go:530
k8s.io/client-go/tools/cache.ResourceEventHandlerFuncs.OnAdd
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/controller.go:243
k8s.io/client-go/tools/cache.(*processorListener).run.func1
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/shared_informer.go:973
k8s.io/apimachinery/pkg/util/wait.BackoffUntil.func1
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:226
k8s.io/apimachinery/pkg/util/wait.BackoffUntil
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:227
k8s.io/apimachinery/pkg/util/wait.JitterUntil
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:204
k8s.io/apimachinery/pkg/util/wait.Until
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:161
k8s.io/client-go/tools/cache.(*processorListener).run
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/shared_informer.go:967
k8s.io/apimachinery/pkg/util/wait.(*Group).Start.func1
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/wait.go:72
2024-12-12T20:35:41.93236794Z syncer/metadatasyncer.go:1853 podAdded: failed to get migration service. Err: volume-migration feature is not supported on Multi-vCenter deployment
sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer.podAdded
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/pkg/syncer/metadatasyncer.go:1853
sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer.InitMetadataSyncer.func7
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/pkg/syncer/metadatasyncer.go:530
k8s.io/client-go/tools/cache.ResourceEventHandlerFuncs.OnAdd
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/controller.go:243
k8s.io/client-go/tools/cache.(*processorListener).run.func1
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/shared_informer.go:973
k8s.io/apimachinery/pkg/util/wait.BackoffUntil.func1
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:226
k8s.io/apimachinery/pkg/util/wait.BackoffUntil
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:227
k8s.io/apimachinery/pkg/util/wait.JitterUntil
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:204
k8s.io/apimachinery/pkg/util/wait.Until
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/apimachinery/pkg/util/wait/backoff.go:161
k8s.io/client-go/tools/cache.(*processorListener).run
        /go/src/github.com/kubernetes-sigs/vsphere-csi-driver/vendor/k8s.io/client-go/tools/cache/shared_informer.go:967
k8s.io/apimachinery/pkg/util/wait.(*Group).Start.func1

Do we even need to watch for podAdd events in clusters where CSI migration is disabled?

cc @divyenpatel

[gnufied]

Also, I actually had csi-migration: false set in featurestate configmap. But apparently this feature is hardcoded and can't be turned off.

So, I am thinking of multi-vc are detected, we should simply disable the migration feature.

[gnufied]

cc @xing-yang