-
Notifications
You must be signed in to change notification settings - Fork 7
Description
Is your feature request related to a problem? Please describe.
People forget to teardown replication sometimes, since people normally cut over and only teardown forward replication, then move on with their lives.
When this happens, often people forget to teardown in pgbelt before deleting the source databases. This bites people later.
Example: Dropping a table a year after cutover - no pgbelt teardown was run so the table was still part of the replication set. Manual resolution was to run DROP on the table, verify that the only related object was the replication set, then run DROP with CASCADE.
Describe the solution you'd like
In this scenario, it would be nicer to teardown against the single databases as a previous target.
Describe alternatives you've considered
DROP with CASCADE is a bit of a blunt hammer and if misused can cause unintended DROPS of related objects, causing data problems.
Additional context
N/A