I'm trying to figure out how Rails handles database connections. The
business problem is:
* My users have access to multiple, structurally identical schemas
* So, for each request, Rails needs to be able to talk to a different
* It's essential that users not pull data from the wrong schema
A couple more rules:
* During the request, the schema used will not change
* All the schemas can be accessed from one database server using one
set of login credentials
The solution I'm contemplating is to execute a "use" statement for
each request, so that all subsequent queries use the correct table.
My questions are:
* How many database connections does each Rails process create?
* Does this vary by adapter? (I'm using MS SQL Server)
* Does apache pool database connections in any way?
* Are database connections independent of Rails processes on apache?