Description of problem:

In OCPBUGS-65674  we found that the operator updates vsphere-csi-config-secret with snapshot configuration on every sync, just reordering items in it:

 [Snapshot]
+granular-max-snapshots-per-block-volume-vvol = 15
 global-max-snapshots-per-block-volume = 5
 granular-max-snapshots-per-block-volume-vsan = 10
-granular-max-snapshots-per-block-volume-vvol = 15 

and then back:

[Snapshot]
-granular-max-snapshots-per-block-volume-vvol = 15
 global-max-snapshots-per-block-volume = 5
 granular-max-snapshots-per-block-volume-vsan = 10
+granular-max-snapshots-per-block-volume-vvol = 15 

As a consequence, the driver Deployment and DaemonSet are re-deployed to re-read the Secret. They may be re-deployed many times, generating a lot of noise.

Version-Release number of selected component (if applicable):

4.21.0-0.nightly-2026-01-19-212205    

How reproducible:

always

Steps to Reproduce:

1. Run e2e vSphere snapshot configuration tests, e.g.:

./openshift-tests run-test "[sig-storage][FeatureGate:VSphereDriverConfiguration][Serial][apigroup:operator.openshift.io] vSphere CSI Driver Configuration snapshot options in clusterCSIDriver should allow all limits to be set at once [Suite:openshift/conformance/serial]"    

Actual results: The `generation` field of the driver Deployment / DaemonSet grows more than by 2. It's often 4, but I've seen more that 10.

Expected results: The `generation` field of the driver Deployment / DaemonSet grows by 2, as each test updates the ClusterCSIDriver twice and the Secret should be updated twice too.