Sorry for two posts in a row, but I suggest the following to the Mezzmo developers, since they'll probably scratch their heads after the issue will be noticed in the future:
After you guys fix this redundancy/space problem, you will need to keep the existing user data. You probably understand by now that re-compacting a database that has gigabytes in size, will be impossible in the most scenarios, due to space and performance limitations. So here's my most valuable advice for your 4.0 issue-solver code: do not run a VACUUM on a database with gigabytes in size, since it could take like forever. Be smart and create a NEW database file, and copy the existing un-redundant tables and data from the OLD Mezzmo.db file. This will not require the same-as_db-size free space issue.
I could have done this myself in the script I posted, but since I have no idea what tables or columns you might decide to add in a future scheme, it's basically not something I would risk myself to do. Plus, the database might be updated by the Mezzmo process while the script is still copying data, so yeah... it's your responsability to solve this.
Bookmarks