We should have detected the test failure in our CI, we need to fill the gap. Here is the goal:

• We want to show that we have table IDs that can overlap and lead to collision for the HS entries

The test in the BF does not include a drop, but it includes a restart, it seems that the following is needed to reproduce the bug:

• Create a table T1 which automatically gets assigned some table ID
• Insert/updates a range of keys so we generate the corresponding HS entries for that table
• Restart, the table ID should go back to 0
• Create tables so we generate table IDs that overlap with the one that was generated by T1
• Somehow here, we need to trigger verify so it checks the HS entries. Verify should find the ones initially associated with T1 but the table ID should lead to one of the newly created tables which should not contain these keys.

The last step is still unclear.