My guess is that it's not resuming the VACUUM, it's trying to check/rescue data from your corrupt database using both the database file and any journal files. Since a lot of new data was written, and it all has to be read, the rescue takes a lot of time. Make sure you have at least twice as much free space on the volume as the database file takes up. Use any SQLite program to open the corrupted database and let the rescue procedure take as long as it takes. Then allow the program to quit normally. You would not lose any data if this is allowed to proceed. Do not use any more VACUUM commands until opening the database takes the normal amount of time, and you have a lot of free space on that volume.