Running simple Vacuum after INSERT in a columnar table makes it give random errors

[muntdan]

I have run a migration of 143mil rows into 1 table and 50mil rows into a second one.
After advice given on Discord about running Vacuum after each 5min batch insert for read query to work after,
I run the migration like first time with 2 parallel inserts on the 2 tables fallowed by vacuum after each insert for their respective table.
I now encounter random errors from time to time which I did not on the first migration (I am using same docker and same tool to migrate):
Exception data: Severity: ERROR SqlState: XX000 MessageText: insufficient data for reading boolean array File: columnar_reader.c Line: 1728 Routine: DeserializeBoolArray
or
Exception data: Severity: ERROR SqlState: XX000 MessageText: attempted to modify an unexpected stripe, columnar storage with id=10000000021 does not have stripe with id=1117 File: columnar_metadata.c Line: 1692 Routine: UpdateStripeMetadataRow

Hydra docker is now only using 230MB RAM (it has 60GB available).

[mkaruza]

Fixed with #74