On Thu, May 26, 2022 at 6:17 AM Bruce Momjian <bruce@momjian.us> wrote:
What do we currently know?
* Caused by CREATE/REINDEX CONCURRENTLY * No way to detect which indexes were created concurrently * amcheck will be required to find corrupt indexes (options?)
Actually in the initial bug report I posted an SQL that is testing sequential scan VS index scan to find the missing records in the index.
This is how we initially found the problem.
I am not sure how these missing records are related to the `amcheck` problems because the `ctid`s does not match exactly.