Personally, I just export/import the database via sql admin. Much simpler to manage.
I use mongrel_rails cluster and apache, and subversion and capistrano. One day soon I will look into doing all this with scripts and / or subversion / capistrano, but for now, this is not too bad:
On my production box:
remove the production database (I keep backups on the development side, which is where all my changes take place):
mysql > drop database projectname_production
create a new empty production database, so I can import new sql into it in a moment:
mysql > create database projectname_production
Then, on development my development box:
dump out the entire development database into an sql file:
mysqldump -u mysql -p projectname_development > projectname_production.sql
copy entire database to production machine
on production machine: mongrel_rails cluster::stop apachectl stop
mysql -u root -p projectname_production < projectname_production.sql
mongrel_rails cluster::start apachectl start
That's what I do --
this may be a silly answer, but just putting it across,
change your database.yml file entries like:
exchange your username, password, database, host etc, in production to test then run the corresponding rake task. (i don't remember the task, may be "rake clone:db_struture" or "rake test:db:prepare"
Note: doing this will copy all your schema but it may not copy the constraints you defined on those database objects.