Hello
This might be not the right way, but I can't think of a better way to not
switch to a different method than split of lastSourceFailed processing and
starting new source. Something like refactoring in first attached patch. I move
RequestXLogStreaming logic from lastSourceFailed processing and add new
pendingRestartSource indicator.
Second patch is mostly the same as previous version but uses new
pendingRestartSource mechanism instead of lastSourceFailed.
Thanks in advance!
regards, Sergei
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index ecd12fc53a..daddaeb236 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -798,10 +798,11 @@ static XLogSource readSource = 0; /* XLOG_FROM_* code */
* different from readSource in that this is always set, even when we don't
* currently have a WAL file open. If lastSourceFailed is set, our last
* attempt to read from currentSource failed, and we should try another source
- * next.
+ * next. If pendingRestartSource is set we want restart current source
*/
static XLogSource currentSource = 0; /* XLOG_FROM_* code */
static bool lastSourceFailed = false;
+static bool pendingRestartSource = false;
typedef struct XLogPageReadPrivate
{
@@ -11828,48 +11829,6 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
if (!StandbyMode)
return false;
- /*
- * If primary_conninfo is set, launch walreceiver to try
- * to stream the missing WAL.
- *
- * If fetching_ckpt is true, RecPtr points to the initial
- * checkpoint location. In that case, we use RedoStartLSN
- * as the streaming start position instead of RecPtr, so
- * that when we later jump backwards to start redo at
- * RedoStartLSN, we will have the logs streamed already.
- */
- if (PrimaryConnInfo && strcmp(PrimaryConnInfo, "") != 0)
- {
- XLogRecPtr ptr;
- TimeLineID tli;
-
- if (fetching_ckpt)
- {
- ptr = RedoStartLSN;
- tli = ControlFile->checkPointCopy.ThisTimeLineID;
- }
- else
- {
- ptr = RecPtr;
-
- /*
- * Use the record begin position to determine the
- * TLI, rather than the position we're reading.
- */
- tli = tliOfPointInHistory(tliRecPtr, expectedTLEs);
-
- if (curFileTLI > 0 && tli < curFileTLI)
- elog(ERROR, "according to history file, WAL location %X/%X belongs to timeline %u, but previous recovered WAL file came from timeline %u",
- (uint32) (tliRecPtr >> 32),
- (uint32) tliRecPtr,
- tli, curFileTLI);
- }
- curFileTLI = tli;
- RequestXLogStreaming(tli, ptr, PrimaryConnInfo,
- PrimarySlotName);
- receivedUpto = 0;
- }
-
/*
* Move to XLOG_FROM_STREAM state in either case. We'll
* get immediate failure if we didn't launch walreceiver,
@@ -11969,10 +11928,83 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
lastSourceFailed ? "failure" : "success");
/*
- * We've now handled possible failure. Try to read from the chosen
- * source.
+ * Prepare to read from the chosen source if we asked restart source
+ * or last source was failed
+ */
+ if (pendingRestartSource || lastSourceFailed)
+ {
+ /*
+ * make sure that walreceiver is not active. this is needed for
+ * all supported sources
+ */
+ if (WalRcvRunning())
+ ShutdownWalRcv();
+
+ switch (currentSource)
+ {
+ case XLOG_FROM_ARCHIVE:
+ case XLOG_FROM_PG_WAL:
+
+ /*
+ * We do not need additional actions here
+ */
+ break;
+
+ case XLOG_FROM_STREAM:
+
+ /*
+ * If primary_conninfo is set, launch walreceiver to try
+ * to stream the missing WAL.
+ *
+ * If fetching_ckpt is true, RecPtr points to the initial
+ * checkpoint location. In that case, we use RedoStartLSN
+ * as the streaming start position instead of RecPtr, so
+ * that when we later jump backwards to start redo at
+ * RedoStartLSN, we will have the logs streamed already.
+ */
+ if (PrimaryConnInfo && strcmp(PrimaryConnInfo, "") != 0)
+ {
+ XLogRecPtr ptr;
+ TimeLineID tli;
+
+ if (fetching_ckpt)
+ {
+ ptr = RedoStartLSN;
+ tli = ControlFile->checkPointCopy.ThisTimeLineID;
+ }
+ else
+ {
+ ptr = RecPtr;
+
+ /*
+ * Use the record begin position to determine the
+ * TLI, rather than the position we're reading.
+ */
+ tli = tliOfPointInHistory(tliRecPtr, expectedTLEs);
+
+ if (curFileTLI > 0 && tli < curFileTLI)
+ elog(ERROR, "according to history file, WAL location %X/%X belongs to timeline %u, but previous recovered WAL file came from timeline %u",
+ (uint32) (tliRecPtr >> 32),
+ (uint32) tliRecPtr,
+ tli, curFileTLI);
+ }
+ curFileTLI = tli;
+ RequestXLogStreaming(tli, ptr, PrimaryConnInfo,
+ PrimarySlotName);
+ receivedUpto = 0;
+ }
+ break;
+ default:
+ elog(ERROR, "unexpected WAL source %d", currentSource);
+ }
+ }
+
+ /*
+ * We've now handled possible failure and pending source restart. Try
+ * to read from the chosen source.
*/
lastSourceFailed = false;
+ pendingRestartSource = false;
switch (currentSource)
{
diff --git a/doc/src/sgml/config.sgml b/doc/src/sgml/config.sgml
index 6d42b7afe7..bf37cc4637 100644
--- a/doc/src/sgml/config.sgml
+++ b/doc/src/sgml/config.sgml
@@ -3839,9 +3839,14 @@ ANY <replaceable class="parameter">num_sync</replaceable> ( <replaceable class="
<varname>primary_conninfo</varname> string.
</para>
<para>
- This parameter can only be set at server start.
+ This parameter can only be set in the <filename>postgresql.conf</filename>
+ file or on the server command line.
This setting has no effect if the server is not in standby mode.
</para>
+ <para>
+ WAL receiver will be restarted after <varname>primary_conninfo</varname>
+ was changed.
+ </para>
</listitem>
</varlistentry>
<varlistentry id="guc-primary-slot-name" xreflabel="primary_slot_name">
@@ -3856,9 +3861,14 @@ ANY <replaceable class="parameter">num_sync</replaceable> ( <replaceable class="
connecting to the sending server via streaming replication to control
resource removal on the upstream node
(see <xref linkend="streaming-replication-slots"/>).
- This parameter can only be set at server start.
+ This parameter can only be set in the <filename>postgresql.conf</filename>
+ file or on the server command line.
This setting has no effect if <varname>primary_conninfo</varname> is not
- set.
+ set or the server is not in standby mode.
+ </para>
+ <para>
+ WAL receiver will be restarted after <varname>primary_slot_name</varname>
+ was changed.
</para>
</listitem>
</varlistentry>
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index daddaeb236..156954b7d9 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -12169,6 +12169,42 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
return false; /* not reached */
}
+void
+ProcessStartupSigHup(void)
+{
+ char *conninfo = pstrdup(PrimaryConnInfo);
+ char *slotname = pstrdup(PrimarySlotName);
+ bool conninfoChanged;
+ bool slotnameChanged;
+
+ ProcessConfigFile(PGC_SIGHUP);
+
+ /*
+ * We need restart XLOG_FROM_STREAM source if replication settings was
+ * changed
+ */
+ conninfoChanged = (strcmp(conninfo, PrimaryConnInfo) != 0);
+ slotnameChanged = (strcmp(slotname, PrimarySlotName) != 0);
+
+ if ((conninfoChanged || slotnameChanged) &&
+ currentSource == XLOG_FROM_STREAM
+ && WalRcvRunning())
+ {
+ if (conninfoChanged && slotnameChanged)
+ ereport(LOG,
+ (errmsg("The WAL receiver is going to be restarted due to change of primary_conninfo and primary_slot_name")));
+ else
+ ereport(LOG,
+ (errmsg("The WAL receiver is going to be restarted due to change of %s",
+ conninfoChanged ? "primary_conninfo" : "primary_slot_name")));
+
+ pendingRestartSource = true;
+ }
+
+ pfree(conninfo);
+ pfree(slotname);
+}
+
/*
* Determine what log level should be used to report a corrupt WAL record
* in the current WAL page, previously read by XLogPageRead().
diff --git a/src/backend/postmaster/startup.c b/src/backend/postmaster/startup.c
index 5048a2c2aa..9bf5c792fe 100644
--- a/src/backend/postmaster/startup.c
+++ b/src/backend/postmaster/startup.c
@@ -147,7 +147,7 @@ HandleStartupProcInterrupts(void)
if (got_SIGHUP)
{
got_SIGHUP = false;
- ProcessConfigFile(PGC_SIGHUP);
+ ProcessStartupSigHup();
}
/*
diff --git a/src/backend/utils/misc/guc.c b/src/backend/utils/misc/guc.c
index 156d147c85..87bd7443ef 100644
--- a/src/backend/utils/misc/guc.c
+++ b/src/backend/utils/misc/guc.c
@@ -3485,7 +3485,7 @@ static struct config_string ConfigureNamesString[] =
},
{
- {"primary_conninfo", PGC_POSTMASTER, REPLICATION_STANDBY,
+ {"primary_conninfo", PGC_SIGHUP, REPLICATION_STANDBY,
gettext_noop("Sets the connection string to be used to connect to the sending server."),
NULL,
GUC_SUPERUSER_ONLY
@@ -3496,7 +3496,7 @@ static struct config_string ConfigureNamesString[] =
},
{
- {"primary_slot_name", PGC_POSTMASTER, REPLICATION_STANDBY,
+ {"primary_slot_name", PGC_SIGHUP, REPLICATION_STANDBY,
gettext_noop("Sets the name of the replication slot to use on the sending server."),
NULL
},
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index f90a6a9139..75ec605a3f 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -318,6 +318,7 @@ extern void SetWalWriterSleeping(bool sleeping);
extern void XLogRequestWalReceiverReply(void);
+extern void ProcessStartupSigHup(void);
extern void assign_max_wal_size(int newval, void *extra);
extern void assign_checkpoint_completion_target(double newval, void *extra);
diff --git a/src/test/recovery/t/001_stream_rep.pl b/src/test/recovery/t/001_stream_rep.pl
index beb45551a2..07ac9642ba 100644
--- a/src/test/recovery/t/001_stream_rep.pl
+++ b/src/test/recovery/t/001_stream_rep.pl
@@ -3,7 +3,7 @@ use strict;
use warnings;
use PostgresNode;
use TestLib;
-use Test::More tests => 26;
+use Test::More tests => 27;
# Initialize master node
my $node_master = get_new_node('master');
@@ -146,7 +146,9 @@ $node_standby_2->append_conf('postgresql.conf',
"primary_slot_name = $slotname_2");
$node_standby_2->append_conf('postgresql.conf',
"wal_receiver_status_interval = 1");
-$node_standby_2->restart;
+# should be able change primary_slot_name without restart
+# will wait effect in get_slot_xmins above
+$node_standby_2->reload;
# Fetch xmin columns from slot's pg_replication_slots row, after waiting for
# given boolean condition to be true to ensure we've reached a quiescent state
@@ -282,3 +284,21 @@ is($catalog_xmin, '',
is($xmin, '', 'xmin of cascaded slot null with hs feedback reset');
is($catalog_xmin, '',
'catalog xmin of cascaded slot still null with hs_feedback reset');
+
+note "check change primary_conninfo without restart";
+$node_standby_2->append_conf('postgresql.conf',
+ "primary_slot_name = ''");
+$node_standby_2->enable_streaming($node_master);
+$node_standby_2->reload;
+
+# be sure do not streaming from cascade
+$node_standby_1->stop;
+
+my $newval = $node_master->safe_psql('postgres',
+'INSERT INTO replayed(val) SELECT coalesce(max(val),0) + 1 AS newval FROM replayed RETURNING val'
+);
+$node_master->wait_for_catchup($node_standby_2, 'replay',
+ $node_master->lsn('insert'));
+my $is_replayed = $node_standby_2->safe_psql('postgres',
+ qq[SELECT 1 FROM replayed WHERE val = $newval]);
+is($is_replayed, qq(1), "standby_2 didn't replay master value $newval");
