What is the most reliable way to verify Docker named volume restoration using checksums?

When restoring a Docker named volume from a tar backup, the goal is to confirm that the recovered data matches the original exactly, without relying on Docker‑specific verification commands.

Constraints include performing the check on the host system using only standard tools (e.g., md5sum, find, diff) and avoiding extraction of the entire volume into a temporary container if possible. It is unclear whether a simple per‑file checksum comparison is sufficient to detect all forms of corruption, or whether a single recursive hash of the volume’s contents provides equivalent assurance while being faster.

Does comparing individual MD5 sums guarantee detection of any missing or altered files? Is a sorted recursive hash adequate and more efficient? Are there scenarios where changes in file ownership or timestamps affect the verification outcome?