Comment 4 for bug 578357

Deleting both ib_logfile0 and ib_logfile1 in ~/.local/share/akonadi/db_data/, and then running 'akonadictl start' fixed the problem for me. I don't know how the log file got corrupted, but the system should be more robust than that. I mean, deleting a (corrupted) log file does not need to be an unrecoverable error!

It's completely reasonable in such a situation to delete it and proceed with startup. The database should still be in a consistent state, and the worst that could happen is that the last few transactions are missing. Or at the very least notify the user what the real problem is. As it is, akonadi server's output is utterly unhelpful, and it took me forever to find mysql.err, where the problem was actually clearly identified.