Add 'no_error' argument to pg_wal_replay_wait()

This argument allow skipping throwing an error.  Instead, the result status
can be obtained using pg_wal_replay_wait_status() function.

Catversion is bumped.

Reported-by: Michael Paquier
Discussion: https://postgr.es/m/ZtUF17gF0pNpwZDI%40paquier.xyz
Reviewed-by: Pavel Borisov

6 files changed, 68 insertions, 8 deletions

diff --git a/src/backend/access/transam/xlogfuncs.c b/src/backend/access/transam/xlogfuncs.c
index ddca78d3717..bca1d395683 100644
--- a/src/backend/access/transam/xlogfuncs.c
+++ b/src/backend/access/transam/xlogfuncs.c
@@ -751,15 +751,18 @@ pg_promote(PG_FUNCTION_ARGS)
 	PG_RETURN_BOOL(false);
 }
 
+static WaitLSNResult lastWaitLSNResult = WAIT_LSN_RESULT_SUCCESS;
+
 /*
- * Waits until recovery replays the target LSN with optional timeout.
+ * Waits until recovery replays the target LSN with optional timeout.  Unless
+ * 'no_error' provided throws an error on failure
  */
 Datum
 pg_wal_replay_wait(PG_FUNCTION_ARGS)
 {
 	XLogRecPtr	target_lsn = PG_GETARG_LSN(0);
 	int64		timeout = PG_GETARG_INT64(1);
-	WaitLSNResult result;
+	bool		no_error = PG_GETARG_BOOL(2);
 
 	if (timeout < 0)
 		ereport(ERROR,
@@ -800,13 +803,16 @@ pg_wal_replay_wait(PG_FUNCTION_ARGS)
 	 */
 	Assert(MyProc->xmin == InvalidTransactionId);
 
-	result = WaitForLSNReplay(target_lsn, timeout);
+	lastWaitLSNResult = WaitForLSNReplay(target_lsn, timeout);
+
+	if (no_error)
+		PG_RETURN_VOID();
 
 	/*
 	 * Process the result of WaitForLSNReplay().  Throw appropriate error if
 	 * needed.
 	 */
-	switch (result)
+	switch (lastWaitLSNResult)
 	{
 		case WAIT_LSN_RESULT_SUCCESS:
 			/* Nothing to do on success */
@@ -832,3 +838,27 @@ pg_wal_replay_wait(PG_FUNCTION_ARGS)
 
 	PG_RETURN_VOID();
 }
+
+Datum
+pg_wal_replay_wait_status(PG_FUNCTION_ARGS)
+{
+	const char *result_string = "";
+
+	/* Process the result of WaitForLSNReplay(). */
+	switch (lastWaitLSNResult)
+	{
+		case WAIT_LSN_RESULT_SUCCESS:
+			result_string = "success";
+			break;
+
+		case WAIT_LSN_RESULT_TIMEOUT:
+			result_string = "timeout";
+			break;
+
+		case WAIT_LSN_RESULT_NOT_IN_RECOVERY:
+			result_string = "not in recovery";
+			break;
+	}
+
+	PG_RETURN_TEXT_P(cstring_to_text(result_string));
+}
diff --git a/src/backend/access/transam/xlogwait.c b/src/backend/access/transam/xlogwait.c
index 58fb10aa5a8..8860a9c73da 100644
--- a/src/backend/access/transam/xlogwait.c
+++ b/src/backend/access/transam/xlogwait.c
@@ -2,7 +2,8 @@
  *
  * xlogwait.c
  *	  Implements waiting for the given replay LSN, which is used in
- *	  CALL pg_wal_replay_wait(target_lsn pg_lsn, timeout float8).
+ *	  CALL pg_wal_replay_wait(target_lsn pg_lsn,
+ *							  timeout float8, no_error bool).
  *
  * Copyright (c) 2024, PostgreSQL Global Development Group
  *
diff --git a/src/backend/catalog/system_functions.sql b/src/backend/catalog/system_functions.sql
index 9c223edfac9..20d3b9b73fd 100644
--- a/src/backend/catalog/system_functions.sql
+++ b/src/backend/catalog/system_functions.sql
@@ -414,7 +414,9 @@ CREATE OR REPLACE FUNCTION
   json_populate_recordset(base anyelement, from_json json, use_json_as_text boolean DEFAULT false)
   RETURNS SETOF anyelement LANGUAGE internal STABLE ROWS 100  AS 'json_populate_recordset' PARALLEL SAFE;
 
-CREATE OR REPLACE PROCEDURE pg_wal_replay_wait(target_lsn pg_lsn, timeout int8 DEFAULT 0)
+CREATE OR REPLACE PROCEDURE pg_wal_replay_wait(target_lsn pg_lsn,
+											   timeout int8 DEFAULT 0,
+											   no_error bool DEFAULT false)
   LANGUAGE internal AS 'pg_wal_replay_wait';
 
 CREATE OR REPLACE FUNCTION pg_logical_slot_get_changes(
diff --git a/src/include/catalog/catversion.h b/src/include/catalog/catversion.h
index 391bf04bf5d..8e0d93239fe 100644
--- a/src/include/catalog/catversion.h
+++ b/src/include/catalog/catversion.h
@@ -57,6 +57,6 @@
  */
 
 /*							yyyymmddN */
-#define CATALOG_VERSION_NO	202410222
+#define CATALOG_VERSION_NO	202410241
 
 #endif
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 6297b7c679d..90010d32a85 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6665,8 +6665,13 @@
 { oid => '8593',
   descr => 'wait for the target LSN to be replayed on standby with an optional timeout',
   proname => 'pg_wal_replay_wait', prokind => 'p', prorettype => 'void',
-  proargtypes => 'pg_lsn int8', proargnames => '{target_lsn,timeout}',
+  proargtypes => 'pg_lsn int8 bool', proargnames => '{target_lsn,timeout,no_error}',
   prosrc => 'pg_wal_replay_wait' },
+{ oid => '8594',
+  descr => 'the last result for pg_wal_replay_wait()',
+  proname => 'pg_wal_replay_wait_status', prorettype => 'text',
+  proargtypes => '',
+  prosrc => 'pg_wal_replay_wait_status' },
 
 { oid => '6224', descr => 'get resource managers loaded in system',
   proname => 'pg_get_wal_resource_managers', prorows => '50', proretset => 't',
diff --git a/src/test/recovery/t/043_wal_replay_wait.pl b/src/test/recovery/t/043_wal_replay_wait.pl
index cf77a9eec70..5857b943711 100644
--- a/src/test/recovery/t/043_wal_replay_wait.pl
+++ b/src/test/recovery/t/043_wal_replay_wait.pl
@@ -77,6 +77,20 @@ $node_standby->psql(
 ok( $stderr =~ /timed out while waiting for target LSN/,
 	"get timeout on waiting for unreachable LSN");
 
+$output = $node_standby->safe_psql(
+	'postgres', qq[
+	CALL pg_wal_replay_wait('${lsn2}', 10, true);
+	SELECT pg_wal_replay_wait_status();]);
+ok( $output eq "success",
+	"pg_wal_replay_wait_status() returns correct status after successful waiting"
+);
+$output = $node_standby->safe_psql(
+	'postgres', qq[
+	CALL pg_wal_replay_wait('${lsn3}', 10, true);
+	SELECT pg_wal_replay_wait_status();]);
+ok($output eq "timeout",
+	"pg_wal_replay_wait_status() returns correct status after timeout");
+
 # 4. Check that pg_wal_replay_wait() triggers an error if called on primary,
 # within another function, or inside a transaction with an isolation level
 # higher than READ COMMITTED.
@@ -193,6 +207,14 @@ $node_standby->safe_psql('postgres', "CALL pg_wal_replay_wait('${lsn5}');");
 
 ok(1, 'wait for already replayed LSN exits immediately even after promotion');
 
+$output = $node_standby->safe_psql(
+	'postgres', qq[
+	CALL pg_wal_replay_wait('${lsn4}', 10, true);
+	SELECT pg_wal_replay_wait_status();]);
+ok( $output eq "not in recovery",
+	"pg_wal_replay_wait_status() returns correct status after standby promotion"
+);
+
 $node_standby->stop;
 $node_primary->stop;