On Tue, Jun 02, 2020 at 04:46:55PM +0900, Masahiko Sawada wrote:
> How about avoiding such an inconsistent situation? In that case,
> replica identity works as NOTHING, but pg_class.relreplident is still
> ‘i’, confusing users. It seems to me that dropping an index specified
> by REPLICA IDENTITY USING INDEX is not a valid operation.
This looks first like complicating RemoveRelations() or the internal
object removal APIs with a dedicated lookup at this index's pg_index
tuple, but you could just put that in index_drop when REINDEX
CONCURRENTLY is not used. Still, I am not sure if it is worth
complicating those code paths. It would be better to get more
opinions about that first.
Consistency is a good goal. Why don't we clear the relreplident from the
relation while dropping the index? relation_mark_replica_identity() already
does that but do other things too. Let's move the first code block from
relation_mark_replica_identity to another function and call this new function
while dropping the index.
--