Improve logical replication connection-failure messages.

tglsfdc · tglsfdc · commit 0d8bd0a72ea2 · 2024-07-11T13:21:13.000-04:00
These messages mostly said "could not connect to the publisher: %s" which is lacking context. Add some verbiage to indicate which subscription or worker process is failing. Nisha Moond Discussion: https://postgr.es/m/CABdArM7q1=zqL++cYd0hVMg3u_tc0S=0Of=Um-KvDhLony0cSg@mail.gmail.com
diff --git a/src/backend/commands/subscriptioncmds.c b/src/backend/commands/subscriptioncmds.c
@@ -755,7 +755,8 @@ CreateSubscription(ParseState *pstate, CreateSubscriptionStmt *stmt,
 		if (!wrconn)
 			ereport(ERROR,
 					(errcode(ERRCODE_CONNECTION_FAILURE),
-					 errmsg("could not connect to the publisher: %s", err)));
+					 errmsg("subscription \"%s\" could not connect to the publisher: %s",
+							stmt->subname, err)));
 
 		PG_TRY();
 		{
@@ -888,7 +889,8 @@ AlterSubscription_refresh(Subscription *sub, bool copy_data,
 	if (!wrconn)
 		ereport(ERROR,
 				(errcode(ERRCODE_CONNECTION_FAILURE),
-				 errmsg("could not connect to the publisher: %s", err)));
+				 errmsg("subscription \"%s\" could not connect to the publisher: %s",
+						sub->name, err)));
 
 	PG_TRY();
 	{
@@ -1521,7 +1523,8 @@ AlterSubscription(ParseState *pstate, AlterSubscriptionStmt *stmt,
 		if (!wrconn)
 			ereport(ERROR,
 					(errcode(ERRCODE_CONNECTION_FAILURE),
-					 errmsg("could not connect to the publisher: %s", err)));
+					 errmsg("subscription \"%s\" could not connect to the publisher: %s",
+							sub->name, err)));
 
 		PG_TRY();
 		{
diff --git a/src/backend/replication/logical/slotsync.c b/src/backend/replication/logical/slotsync.c
@@ -1463,7 +1463,8 @@ ReplSlotSyncWorkerMain(char *startup_data, size_t startup_data_len)
 	if (!wrconn)
 		ereport(ERROR,
 				errcode(ERRCODE_CONNECTION_FAILURE),
-				errmsg("could not connect to the primary server: %s", err));
+				errmsg("synchronization worker \"%s\" could not connect to the primary server: %s",
+					   app_name.data, err));
 
 	/*
 	 * Register the disconnection callback.
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
@@ -1342,7 +1342,8 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos)
 	if (LogRepWorkerWalRcvConn == NULL)
 		ereport(ERROR,
 				(errcode(ERRCODE_CONNECTION_FAILURE),
-				 errmsg("could not connect to the publisher: %s", err)));
+				 errmsg("table synchronization worker for subscription \"%s\" could not connect to the publisher: %s",
+						MySubscription->name, err)));
 
 	Assert(MyLogicalRepWorker->relstate == SUBREL_STATE_INIT ||
 		   MyLogicalRepWorker->relstate == SUBREL_STATE_DATASYNC ||
diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c
@@ -4500,7 +4500,8 @@ run_apply_worker()
 	if (LogRepWorkerWalRcvConn == NULL)
 		ereport(ERROR,
 				(errcode(ERRCODE_CONNECTION_FAILURE),
-				 errmsg("could not connect to the publisher: %s", err)));
+				 errmsg("apply worker for subscription \"%s\" could not connect to the publisher: %s",
+						MySubscription->name, err)));
 
 	/*
 	 * We don't really use the output identify_system for anything but it does
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
@@ -898,7 +898,8 @@ pg_sync_replication_slots(PG_FUNCTION_ARGS)
 	if (!wrconn)
 		ereport(ERROR,
 				errcode(ERRCODE_CONNECTION_FAILURE),
-				errmsg("could not connect to the primary server: %s", err));
+				errmsg("synchronization worker \"%s\" could not connect to the primary server: %s",
+					   app_name.data, err));
 
 	SyncReplicationSlots(wrconn);
 
diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c
@@ -195,6 +195,7 @@ WalReceiverMain(char *startup_data, size_t startup_data_len)
 	char	   *err;
 	char	   *sender_host = NULL;
 	int			sender_port = 0;
+	char	   *appname;
 
 	Assert(startup_data_len == 0);
 
@@ -298,13 +299,13 @@ WalReceiverMain(char *startup_data, size_t startup_data_len)
 	sigprocmask(SIG_SETMASK, &UnBlockSig, NULL);
 
 	/* Establish the connection to the primary for XLOG streaming */
-	wrconn = walrcv_connect(conninfo, true, false, false,
-							cluster_name[0] ? cluster_name : "walreceiver",
-							&err);
+	appname = cluster_name[0] ? cluster_name : "walreceiver";
+	wrconn = walrcv_connect(conninfo, true, false, false, appname, &err);
 	if (!wrconn)
 		ereport(ERROR,
 				(errcode(ERRCODE_CONNECTION_FAILURE),
-				 errmsg("could not connect to the primary server: %s", err)));
+				 errmsg("streaming replication receiver \"%s\" could not connect to the primary server: %s",
+						appname, err)));
 
 	/*
 	 * Save user-visible connection string.  This clobbers the original
diff --git a/src/test/regress/expected/subscription.out b/src/test/regress/expected/subscription.out
@@ -139,7 +139,7 @@ ERROR:  invalid connection string syntax: invalid connection option "i_dont_exis
 -- fail, connection string parses, but doesn't work (and does so without
 -- connecting, so this is reliable and safe)
 CREATE SUBSCRIPTION regress_testsub5 CONNECTION 'port=-1' PUBLICATION testpub;
-ERROR:  could not connect to the publisher: invalid port number: "-1"
+ERROR:  subscription "regress_testsub5" could not connect to the publisher: invalid port number: "-1"
 -- fail - invalid connection string during ALTER
 ALTER SUBSCRIPTION regress_testsub CONNECTION 'foobar';
 ERROR:  invalid connection string syntax: missing "=" after "foobar" in connection info string
