Problem
A new standby fast truncate design will be adopted to enable fast truncate. The new design means that we will track uncommitted and committed truncates on the layered dhandle. If the sweep server sweeps the layered dhandle with these truncate entries still existing in the dhandle, we have effectively caused a data corruption.

What needs to be done
The ticket will focus on investigating and creating a python test that re-enacts this scenario. If the investigation reproduces this problem, we must disallow sweep server from sweeping layered dhandle.