Ensure we have a snapshot when updating pg_index in index_drop().

I assumed that all index_drop() callers set an active snapshot
beforehand, but that is evidently not true.  One counterexample is
autovacuum, which doesn't set an active snapshot when cleaning up
orphan temp indexes.  To fix, unconditionally push an active
snapshot before updating pg_index in index_drop().

Oversight in commit b52adbad46.

Reported-by: Masahiko Sawada
Reviewed-by: Stepan Neretin, Masahiko Sawada
Discussion: https://postgr.es/m/CAD21AoBgF9etQrXbN9or_YHsmBRJHHNUEkhHp9rGK9CyQv5aTQ%40mail.gmail.com
This commit is contained in:
Nathan Bossart 2024-10-28 16:44:31 -05:00
parent 11b7de4a78
commit 70b9adb98e

View File

@ -2336,13 +2336,9 @@ index_drop(Oid indexId, bool concurrent, bool concurrent_lock_mode)
/*
* Updating pg_index might involve TOAST table access, so ensure we have a
* valid snapshot. We only expect to get here without a snapshot in the
* concurrent path.
* valid snapshot.
*/
if (concurrent)
PushActiveSnapshot(GetTransactionSnapshot());
else
Assert(HaveRegisteredOrActiveSnapshot());
PushActiveSnapshot(GetTransactionSnapshot());
/*
* fix INDEX relation, and check for expressional index
@ -2361,8 +2357,7 @@ index_drop(Oid indexId, bool concurrent, bool concurrent_lock_mode)
ReleaseSysCache(tuple);
table_close(indexRelation, RowExclusiveLock);
if (concurrent)
PopActiveSnapshot();
PopActiveSnapshot();
/*
* if it has any expression columns, we might have stored statistics about