On Nov 23, 2025 at 11:46 PM Fujii Masao
<masao(dot)fujii(at)gmail(dot)com> wrote:
> The difference is that PGC_USERSET also allows per–replication-user
> overrides when needed, which gives users more flexibility without
> losing the ability to set a server-wide setting, I think.
> ...
> I think there are valid use cases for applying this setting to
> physical replication as well.
Thanks for the comments. I agree, this parameter also seems usable
for physical replication, if you use it with caution. In this case,
it really becomes useful to be able to configure a parameter for
each connection. I have added these changes to my patch.
Also, earlier I did not mention another difference between my patch
and those discussed earlier. Previously, even in immediate mode,
WalSndCaughtUp flag was checked before calling WalSndDone,
and this made it impossible to shut down even in immediate mode
with WalSndCaughtUp = false when the server has full output buffers.
This does not happen in the current patch implementation. I added
an additional test case for this situation.
Regards,
Andrey Silitskiy
From 6c06cc76cf32d7f47c93b6de9f92aa54eaeead3c Mon Sep 17 00:00:00 2001
From: "a.silitskiy" <[email protected]>
Date: Thu, 27 Nov 2025 15:52:34 +0700
Subject: [PATCH v2] Introduce a new GUC 'wal_sender_shutdown_mode'.
Previously, at shutdown, walsender processes were always waiting to send all
pending data and ensure that all data is flushed in remote node. But in some cases
an unexpected wait may be unacceptable. For example, in logical replication,
apply_workers may hang on locks for some time, excluding the possibility of
sender's shutdown.
New guc allows to change shutdown mode of walsenders without changing
default behavior.
The shutdown modes are:
1) 'wait_flush' (the default). In this mode, the walsender will wait for all
WALs to be flushed on the receiver side, before exiting the process.
2) 'immediate'. In this mode, the walsender will exit without confirming the
remote flush. This may break the consistency between sender and receiver.
This mode might be useful for a system that has a high-latency network (to
reduce the amount of time for shutdown), or to allow the shutdown of
publisher even when when the subscriber's apply_worker is waiting for any
locks to be released.
Author: Andrey Silitskiy
Co-authored by: Hayato Kuroda
Discussion: https://postgr.es/m/TYAPR01MB586668E50FC2447AD7F92491F5E89%40TYAPR01MB5866.jpnprd01.prod.outlook.com
---
doc/src/sgml/config.sgml | 33 +++++
src/backend/replication/walsender.c | 44 ++++++
src/backend/utils/misc/guc_parameters.dat | 7 +
src/backend/utils/misc/guc_tables.c | 6 +
src/backend/utils/misc/postgresql.conf.sample | 2 +
src/include/replication/walsender.h | 7 +
.../t/037_walsnd_immediate_shutdown.pl | 136 ++++++++++++++++++
7 files changed, 235 insertions(+)
create mode 100644 src/test/subscription/t/037_walsnd_immediate_shutdown.pl
diff --git a/doc/src/sgml/config.sgml b/doc/src/sgml/config.sgml
index 737b90736bf..3aa09f90a65 100644
--- a/doc/src/sgml/config.sgml
+++ b/doc/src/sgml/config.sgml
@@ -4708,6 +4708,39 @@ restore_command = 'copy "C:\\server\\archivedir\\%f" "%p"' # Windows
</listitem>
</varlistentry>
+ <varlistentry id="guc-wal_sender_shutdown_mode" xreflabel="wal_sender_shutdown_mode">
+ <term><varname>wal_sender_shutdown_mode</varname> (<type>enum</type>)
+ <indexterm>
+ <primary><varname>wal_sender_shutdown_mode</varname> configuration parameter</primary>
+ </indexterm>
+ </term>
+ <listitem>
+ <para>
+ Specifies the mode in which walsender process will terminate
+ after receival of shutdown request. Valid values are <literal>wait_flush</literal>
+ and <literal>immediate</literal>. Default value is <literal>wait_flush</literal>.
+ Can be set for each walsender.
+ </para>
+ <para>
+ In <literal>wait_flush</literal> mode, the walsender will wait for all
+ WALs to be flushed on the receiver side, before exiting the process. May
+ lead to unexpected lag of server shutdown.
+ </para>
+ <para>
+ In <literal>immediate</literal> mode, the walsender will exit without waiting
+ for data replication to the receiver. This may break data consistency between
+ sender and receiver after shutdown, which can be especially important in
+ case of physical replication and switch-over.
+ </para>
+ <para>
+ This mode might be useful for a system that has a high-latency network (to
+ reduce the amount of time for shutdown), or to allow the shutdown of
+ logical replication walsender even when the subscriber's apply_worker
+ is waiting for any locks to be released.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="guc-synchronized-standby-slots" xreflabel="synchronized_standby_slots">
<term><varname>synchronized_standby_slots</varname> (<type>string</type>)
<indexterm>
diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c
index fc8f8559073..298b167d1e2 100644
--- a/src/backend/replication/walsender.c
+++ b/src/backend/replication/walsender.c
@@ -130,6 +130,9 @@ int max_wal_senders = 10; /* the maximum number of concurrent
* walsenders */
int wal_sender_timeout = 60 * 1000; /* maximum time to send one WAL
* data message */
+
+int wal_sender_shutdown_mode = WALSND_SHUTDOWN_MODE_WAIT_FLUSH;
+
bool log_replication_commands = false;
/*
@@ -262,6 +265,7 @@ static void WalSndKill(int code, Datum arg);
pg_noreturn static void WalSndShutdown(void);
static void XLogSendPhysical(void);
static void XLogSendLogical(void);
+pg_noreturn static void WalSndDoneImmediate(void);
static void WalSndDone(WalSndSendDataCallback send_data);
static void IdentifySystem(void);
static void UploadManifest(void);
@@ -1650,6 +1654,11 @@ ProcessPendingWrites(void)
/* Try to flush pending output to the client */
if (pq_flush_if_writable() != 0)
WalSndShutdown();
+
+ /* If we got shut down request in immediate shutdown mode, exit the process */
+ if ((got_STOPPING || got_SIGUSR2) &&
+ wal_sender_shutdown_mode == WALSND_SHUTDOWN_MODE_IMMEDIATE)
+ WalSndDoneImmediate();
}
/* reactivate latch so WalSndLoop knows to continue */
@@ -2927,6 +2936,14 @@ WalSndLoop(WalSndSendDataCallback send_data)
if (pq_flush_if_writable() != 0)
WalSndShutdown();
+ /*
+ * When immediate shutdown of walsender is requested, we do not
+ * wait for successfull sending of all data.
+ */
+ if ((got_STOPPING || got_SIGUSR2) &&
+ wal_sender_shutdown_mode == WALSND_SHUTDOWN_MODE_IMMEDIATE)
+ WalSndDoneImmediate();
+
/* If nothing remains to be sent right now ... */
if (WalSndCaughtUp && !pq_is_send_pending())
{
@@ -3575,6 +3592,33 @@ XLogSendLogical(void)
}
}
+/*
+ * Shutdown walsender in immediate mode.
+ *
+ * NB: This should only be called when immediate shutdown of walsender
+ * was requested and shutdown signal has been received from postmaster.
+ */
+static void
+WalSndDoneImmediate()
+{
+ QueryCompletion qc;
+
+ /* Try to inform receiver that XLOG streaming is done */
+ SetQueryCompletion(&qc, CMDTAG_COPY, 0);
+ EndCommand(&qc, DestRemote, false);
+
+ /*
+ * Note that the output buffer may be full during immediate shutdown of
+ * walsender. If pq_flush() is called at that time, the walsender process
+ * will be stuck. Therefore, call pq_flush_if_writable() instead. Successfull
+ * receival of done message in immediate shutdown mode is not guaranteed.
+ */
+ pq_flush_if_writable();
+
+ proc_exit(0);
+ abort();
+}
+
/*
* Shutdown if the sender is caught up.
*
diff --git a/src/backend/utils/misc/guc_parameters.dat b/src/backend/utils/misc/guc_parameters.dat
index 3b9d8349078..6911635a85a 100644
--- a/src/backend/utils/misc/guc_parameters.dat
+++ b/src/backend/utils/misc/guc_parameters.dat
@@ -3422,6 +3422,13 @@
check_hook => 'check_wal_segment_size',
},
+{ name => 'wal_sender_shutdown_mode', type => 'enum', context => 'PGC_USERSET', group => 'REPLICATION_SENDING',
+ short_desc => 'Sets the mode in which walsender will be terminated after shutdown request.',
+ variable => 'wal_sender_shutdown_mode',
+ boot_val => 'WALSND_SHUTDOWN_MODE_WAIT_FLUSH',
+ options => 'wal_sender_shutdown_mode_options',
+},
+
{ name => 'wal_sender_timeout', type => 'int', context => 'PGC_USERSET', group => 'REPLICATION_SENDING',
short_desc => 'Sets the maximum time to wait for WAL replication.',
flags => 'GUC_UNIT_MS',
diff --git a/src/backend/utils/misc/guc_tables.c b/src/backend/utils/misc/guc_tables.c
index f87b558c2c6..3306ebb39f9 100644
--- a/src/backend/utils/misc/guc_tables.c
+++ b/src/backend/utils/misc/guc_tables.c
@@ -335,6 +335,12 @@ static const struct config_enum_entry constraint_exclusion_options[] = {
{NULL, 0, false}
};
+static const struct config_enum_entry wal_sender_shutdown_mode_options[] = {
+ {"wait_flush", WALSND_SHUTDOWN_MODE_WAIT_FLUSH, false},
+ {"immediate", WALSND_SHUTDOWN_MODE_IMMEDIATE, false},
+ {NULL, 0, false}
+};
+
/*
* Although only "on", "off", "remote_apply", "remote_write", and "local" are
* documented, we accept all the likely variants of "on" and "off".
diff --git a/src/backend/utils/misc/postgresql.conf.sample b/src/backend/utils/misc/postgresql.conf.sample
index dc9e2255f8a..111ae7dea48 100644
--- a/src/backend/utils/misc/postgresql.conf.sample
+++ b/src/backend/utils/misc/postgresql.conf.sample
@@ -346,6 +346,8 @@
#wal_sender_timeout = 60s # in milliseconds; 0 disables
#track_commit_timestamp = off # collect timestamp of transaction commit
# (change requires restart)
+#wal_sender_shutdown_mode = wait_flush # walsender termination mode after
+ # receival of shutdown request
# - Primary Server -
diff --git a/src/include/replication/walsender.h b/src/include/replication/walsender.h
index c3e8e191339..228f3f0aa2c 100644
--- a/src/include/replication/walsender.h
+++ b/src/include/replication/walsender.h
@@ -24,6 +24,12 @@ typedef enum
CRS_USE_SNAPSHOT,
} CRSSnapshotAction;
+typedef enum
+{
+ WALSND_SHUTDOWN_MODE_WAIT_FLUSH = 0,
+ WALSND_SHUTDOWN_MODE_IMMEDIATE
+} WalSndShutdownMode;
+
/* global state */
extern PGDLLIMPORT bool am_walsender;
extern PGDLLIMPORT bool am_cascading_walsender;
@@ -33,6 +39,7 @@ extern PGDLLIMPORT bool wake_wal_senders;
/* user-settable parameters */
extern PGDLLIMPORT int max_wal_senders;
extern PGDLLIMPORT int wal_sender_timeout;
+extern PGDLLIMPORT int wal_sender_shutdown_mode;
extern PGDLLIMPORT bool log_replication_commands;
extern void InitWalSender(void);
diff --git a/src/test/subscription/t/037_walsnd_immediate_shutdown.pl b/src/test/subscription/t/037_walsnd_immediate_shutdown.pl
new file mode 100644
index 00000000000..8bbf64df8e0
--- /dev/null
+++ b/src/test/subscription/t/037_walsnd_immediate_shutdown.pl
@@ -0,0 +1,136 @@
+# Checks that publisher is able to shut down without
+# waiting for sending of all pending data to subscriber
+# with wal_sender_shutdown_mode = immediate
+
+use strict;
+use warnings FATAL => 'all';
+
+use PostgreSQL::Test::Cluster;
+use Test::More;
+
+sub run_with_timeout
+{
+ my ($seconds, $code_to_run) = @_;
+
+ eval
+ {
+ local $SIG{ALRM} = sub { die "timeout\n" };
+ alarm $seconds;
+ $code_to_run->();
+ alarm 0;
+ };
+
+ return $@; # Empty string means success or contains error text.
+}
+
+sub ok_with_timeout
+{
+ my ($seconds, $code, $msg) = @_;
+ my $err = run_with_timeout($seconds, $code);
+
+ if (!$err)
+ {
+ pass($msg);
+ }
+ elsif ($err =~ /timeout/)
+ {
+ fail("$msg (timed out after $seconds seconds)");
+ }
+ else
+ {
+ fail("$msg (unknown error: $err)");
+ }
+}
+
+my $out;
+
+# create publisher
+my $publisher = PostgreSQL::Test::Cluster->new('publisher');
+$publisher->init(allows_streaming => 'logical');
+# set wal_sender_shutdown_mode GUC parameter to immediate
+$publisher->append_conf('postgresql.conf',
+ 'wal_sender_timeout = 0
+ wal_sender_shutdown_mode = immediate');
+$publisher->start();
+
+# create subscriber
+my $subscriber = PostgreSQL::Test::Cluster->new('subscriber');
+$subscriber->init();
+$subscriber->start();
+
+# create publication for test table
+$publisher->safe_psql('postgres', q{
+ CREATE TABLE pub_test (id int PRIMARY KEY);
+ CREATE PUBLICATION pub_all FOR TABLE pub_test;
+});
+
+# create matching table on subscriber
+$subscriber->safe_psql('postgres', q{
+ CREATE TABLE pub_test (id int PRIMARY KEY);
+});
+
+# form connection string to publisher
+my $pub_connstr = $publisher->connstr;
+
+# create the subscription on subscriber
+$subscriber->safe_psql('postgres', qq{
+ CREATE SUBSCRIPTION sub_all
+ CONNECTION '$pub_connstr'
+ PUBLICATION pub_all;
+});
+
+# wait for initial sync to finish
+$subscriber->wait_for_subscription_sync($publisher, 'sub_all');
+
+# create background psql session
+my $bpgsql = $subscriber->background_psql('postgres', on_error_stop => 0);
+
+
+# =============================================================================
+# Testcase start: Shutdown of publisher with empty output buffers
+
+# start transaction on subscriber to hold locks
+$bpgsql->query_safe(
+ "BEGIN; INSERT INTO pub_test VALUES (0);"
+);
+
+# run concurrent transaction on publisher and commit
+$out = $publisher->safe_psql('postgres', 'BEGIN; INSERT INTO pub_test VALUES (0); COMMIT;');
+ok($out eq "", "Concurrent transaction was committed on publisher");
+
+# test publisher shutdown
+ok_with_timeout(5, sub { $publisher->stop('fast') },
+ "Successfull fast shutdown of server with empty output buffers");
+
+# Testcase end: Shutdown of publisher with empty output buffers
+# =============================================================================
+
+$bpgsql->query_safe(
+ "ABORT;"
+);
+
+# restart publisher for the next testcase
+$publisher->start();
+
+$subscriber->wait_for_subscription_sync($publisher, 'sub_all');
+
+# =============================================================================
+# Testcase start: Shutdown of publisher with full output buffers
+
+# lock table to make apply_worker hang
+$bpgsql->query_safe(
+ "BEGIN; LOCK TABLE pub_test IN EXCLUSIVE MODE;"
+);
+
+# generate big amount of wal records for locked table
+$out = $publisher->safe_psql('postgres', 'BEGIN; INSERT INTO pub_test SELECT i from generate_series(1, 20000) s(i); COMMIT;');
+ok($out eq "", "Inserts into locked table successfully generated");
+
+# test publisher shutdown
+ok_with_timeout(5, sub { $publisher->stop('fast') },
+ "Successfull fast shutdown of server with full output buffers");
+
+# Testcase end: Shutdown of publisher with full output buffers
+# =============================================================================
+
+done_testing();
--
2.34.1
