This one isn’t a long one, but I found an interesting case of active vowel disharmony in Nora England’s grammar of Mam, a Mayan language. Most cases I know where vowels actively dissimilate are raising of adjacent vowels in hiatus, e.g. in many Basque dialects the article -a triggers raising of an immediately preceding vowel, as in seme-a “son-the” = semia. But this is not the case in Mam, where a number of causative suffixes have vowels of the opposite backness to the vowel of the preceding syllable in a non-hiatus environment.

In England’s survey, the most frequency of these suffixes was –pii (after o, u) vs –puu (after i, e, a), with 3/42 exceptions that failed to show disharmony.

Excluding vowels in hiatus, vowel disharmony still seems like a relatively rare process which is generally restricted to specific morphemes when it does occur, but it’s still interesting to discover examples.