Incremental restore and restore database specific

[todeb]

Overview

PGO allow do incremental restore of restored or cloned database
PGO allow only clone and restore specified database, if no database specified, all databases are cloned/restored

Use Case

• Migrate huge databases to new cluster, perform full clone, then apply incremental restore to reduce downtime.
• Split database from existing cluster to new or combine existing databases in two clusters into one. Specify the database or databases that you want to move. So only the specified database will be cloned/restored and other databases that exists on destination and source clusters will be untouched.

Desired Behavior

Additional parameters in datasource for specifing database to restore and incremental. Or deploying new yaml file with restore specification.

Environment

Tell us about your environment:

• Platform: Kubernetes

[jmckulk]

Hi @todeb, happy to get these feature requests in, but I'm not fully understanding what you are looking for with an incremental restore. I understand that an incremental backup would be only changes since the previous full backup. When you are restoring with pgBackRest I'm not aware of any incremental restore option. Also, if you have a full clone, what else is left to restore?

Can you go into any more detail about what you are looking for here? Do you have any examples of how you would approach this using pgBackRest or any other backup/restore tool?

[todeb]

yes, you right there is no such thing in pgbackrest like incremenat restore.
Originally I meant to restore full database and if there are changes in source database, just restore these changes. maybe that can be achieved by replying WAL, although might be a bit tricky.

Then maybe we can just focus to be able to clone specific database from existing cluster or backup?

[benjaminjb]

Hi @todeb, I just wanted to update you on this before the weekend to say that we have captured "clone specific database from cluster or backup" in our backlog.