Check for misordered small clusters as well
The check for misordered clusters was only performed for large clusters. This leads to a rare behaviour where corrupted data is still misordered but gets stored anyway. Fixes follow-up of Rec#512 (closed) and discussions in Moore!2678 (merged).
Additionally, we will now skip both misordered clusters, as there is no clear way to check which are corrupted.
Edited by Michel De Cian