Add further debug info to help debug 019_replslot_limit.pl failures.

See also afdeff1052. Failures after that commit provided a few more hints,
but not yet enough to understand what's going on.

In 019_replslot_limit.pl shut down nodes with fast instead of immediate mode
if we observe the failure mode. That should tell us whether the failures we're
observing are just a timing issue under high load. PGCTLTIMEOUT should prevent
buildfarm animals from hanging endlessly.

Also adds a bit more logging to replication slot drop and ShutdownPostgres().

Discussion: https://postgr.es/m/20220225192941.hqnvefgdzaro6gzg@alap3.anarazel.de

src/backend/utils/init/postinit.c

@@ -1262,6 +1262,23 @@ ShutdownPostgres(int code, Datum arg)
 	 * them explicitly.
 	 */
 	LockReleaseAll(USER_LOCKMETHOD, true);
+
+	/*
+	 * temp debugging aid to analyze 019_replslot_limit failures
+	 *
+	 * If an error were thrown outside of a transaction nothing up to now
+	 * would have released lwlocks. We probably will add an
+	 * LWLockReleaseAll(). But for now make it easier to understand such cases
+	 * by warning if any lwlocks are held.
+	 */
+#ifdef USE_ASSERT_CHECKING
+	{
+		int held_lwlocks = LWLockHeldCount();
+		if (held_lwlocks)
+			elog(WARNING, "holding %d lwlocks at the end of ShutdownPostgres()",
+				 held_lwlocks);
+	}
+#endif
 }