Hi Michael, Kyotaro, Thanks for the clarifications and feedback.
I spent some time looking at how core tools like pg_rewind use this value. After going through the code paths and the existing pg_rewind tests, I understand the concern that minRecoveryPoint should primarily be treated as a consistency-floor value. I also see that pg_rewind currently depends on minRecoveryPoint in some standby-related decisions, but I agree that this does not necessarily mean it should be treated as a general indicator of how far recovery has progressed. Given that, I’m okay to treat the current behavior as expected for now and avoid pushing further on this until we have Adam’s view on whether he still wants to pursue the change with stronger justification and tests, or whether he prefers to drop or rework it based on this semantic direction. Best Regards, Nitin Jadhav Azure Database for PostgreSQL Microsoft
