Hi Sawada-San.
Some review comments for v24-0001.
======
src/backend/replication/slot.c
ReplicationSlotsDropDBSlots:
1.
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (i = 0; i < max_replication_slots; i++)
{
ReplicationSlot *s;
+ bool invalidated;
char *slotname;
int active_pid;
@@ -1462,11 +1494,22 @@ restart:
if (!SlotIsLogical(s))
continue;
+ SpinLockAcquire(&s->mutex);
+ invalidated = s->data.invalidated == RS_INVAL_NONE;
+ SpinLockRelease(&s->mutex);
+
+ /*
+ * Count slots on other databases too so we can disable logical
+ * decoding only if no slots in the cluster.
+ */
+ if (invalidated)
+ n_valid_logicalslots++;
+
IMO it's misleading/backwards to have bool invalidated == true when
the invalidated member is RS_INVAL_NONE.
It's fine to rename the bool var as 'invalidated', but don't you also
need to reverse the assignment to (s->data.invalidated !=
RS_INVAL_NONE) instead of (s->data.invalidated == RS_INVAL_NONE)?
======
Kind Regards,
Peter Smith.
Fujitsu Australia