The policy `ClusterChunksResizePolicy` is splitting the chunks to the defined `chunkSize` on downgrade.

It was implemented to go back to a chunk size scenario where the flag `featureFlagBalanceAccordingToDataSize` was disabled (the flag has been deleted on 6dad29fa6).

That feature flag was enabled for v6.0.3 on SERVER-66781. So, assuming all downgrades from v7.0 will target a version equal or greater that v6.0.3, we must get rid of this policy to avoid over splitting the chunks on downgrade.