I don't know if it works as designed, but I discovered the following
today: I wanted to shrink my full store from 81GiB cache size to
75GiB. I stopped the node, changed the store size and restarted the
node. In wrapper.log I followed the progress. The node started to scan
for holes in the store prior to shrinking it. I went away, and after
nearly 2 hours I came back and hey: the scan was at block 300.000 of
1.800.000 blocks! This is far to slow and means if I want to shrink my
store it takes around 12 hours!? Can this be faster, please?
