I think when I spoke to Gabriel about it (arbitrary clusters), he mentioned it would be a fundamental change to the way clusters work and would require some sizable effort.
I agree on the point about zero page length and a race condition… that wouldn’t be good!
cheers
ripe
