Point in time recovery for CloudSQL migration database (PostgreSQL)?
Slava Fomin II
Hello!
I have a PostgreSQL replica database created in Cloud SQL using the Database Migration tool. Effectively, it's a streaming logical replication of the existing production PostgreSQL database implemented using pglogical extension.
I'm using this replica as a means for backup, archiving and analytics. In order to use it effectively, I need a way to create a separate copy of the replica, probably using point in time recovery to be able to time-travel to the specific point in the past in order to analyze the state of the data or to restore the database from the backup.
However, I can't find a way to actually do this in GCP. There is no option to clone the existing replica database, I can only promote the replica to the stand-alone instance, but this will stop the replication process and I don't want that.
Why does such limitations exist for migration replicas and is there an option to somehow overcome this? Right now, the solution provided by Cloud SQL Migrations are just not suitable for my use case it seams.
Thanks!