Clarify retained audit-store exposure (#952)

[sehkone]

Summary

• Clarifies the non-empty audit-store outcome on the bootroot infra up surface in English and Korean.
• Explains that an existing underlying openbao/ remains a bind source after a failed mount until records are relocated onto the reserve.
• Adds regression coverage for both localized rendered outcomes and existing-records manual passages.

Closes #952. Part of #924.

Test plan

• Verify rendered English and Korean non-empty-store outcomes state the infra up refusal, retained bind exposure, no-records-change boundary, separate relocation, and directory-mode opt-out.
• Verify both operations manuals explicitly state the live infra up refusal and contain no relocation procedure.
• Run ./scripts/check-docs.sh.
• Run cargo fmt -- --config group_imports=StdExternalCrate --check and cargo clippy --all-targets -- -D warnings.
• Run scripts/preflight/run-all.sh (including core tests and the Docker E2E matrix).

[sehkone]

[Reviewer Round 1]

No findings. The PR matches Issue #952’s documentation-only scope.

• The shared non-empty-store finding now clearly describes the infra up refusal, retained failed-mount bind exposure, relocation boundary, and directory-mode opt-out in both locales: en.rs, ko.rs.
• Both existing-records manual passages state the live refusal directly and avoid adding a relocation procedure: English manual, Korean manual.
• The new regression tests exercise the rendered NotActivated outcome in English and Korean and protect the manual passages against loss of the required boundary: reserve tests.

PR hygiene is also correct: #953 uses Closes #952, the permitted Part of #924, includes a checked test plan, and has no review-thread issue creation.

[sehkone]

[Review Verdict Round 1: APPROVED]