Confluence - Migrate MySQL to PostgreSQL DB

[collene]

I am new here and doing what I am about to ask for the first time so bare with me and my questions :)

We are currently running Confluence v6.13 with a local MySQL database.  We are trying to migrate our MySQL database to PostgreSQL database.

So far we have done a migration of the MySQL data to a separate/external PostgreSQL database using pgloader.  We are separating the app and db to be on their own separate servers.

I have read the Atlassian documentation for database migration: https://confluence.atlassian.com/conf613/migrating-to-another-database-964961318.html 

I have some questions which I am unsure about:

1. Since I migrated the MySQL database's data to a PostgreSQL database, can I just point to the new PostgreSQL database via the confluence.cfg.xml file that is on the current installation of Confluence?
2. If the answer is No to Question 1, then is modifying the confluence.cfg.xml file only meant for changing to the same type of DB? As in local MySQL DB to a external/separate MySQL database or vice versa.
3. If the fresh install (as pointed in the documentation) is required for the migration then during the setup wizard would just pointing to the PostgreSQL database I created using pgloader work?
   1. Has anyone done something like this?
   2. What is considered a large data set?
   3. Would it be better, upon setup of the new installation, to point to a empty PostgreSQL database before modifying the confluence.cfg.xml file to point to the migrated one that we created from pgloader?

[Monique Khairuliana[ServiceRocket]]

Hey @collene 

For question #1, I don't think it will be that straightforward. You'd need to change the database's driver_class, hibernate.dialect, among others that will be different from each database type. That answers question #2, yes, it's meant for same database type migration.

For Q#3, If you point to the db that already have the migrated data during the setup wizard, it'll complain that the database is not empty, and prompt you to overwrite the data anyway. You're right, you'd have to point it to an empty database in postgreSQL first, as mentioned in your point #3. This will prompt confluence to generate the correct confluence.cfg.xml parameters needed for pgsql.