On 2020-01-15 05:02, Kyotaro Horiguchi wrote:
FWIW, I restate this (perhaps) more clearly.
At Wed, 15 Jan 2020 11:02:24 +0900 (JST), Kyotaro Horiguchi
<horikyota....@gmail.com> wrote in
recvoery_target_* is not cleared after startup. If a server crashed
just after the last shutdown checkpoint, any recovery_target_* setting
prevents the server from starting regardless of its value.
recvoery_target_* is not automatically cleared after a successful
archive recovery. After that, if the server crashed just after the
last shutdown checkpoint, any recovery_target_* setting prevents the
server from starting regardless of its value.
Thank you for this clarification. Here is a new patch that addresses
that and also the other comments raised about my previous patch.
--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
From a0ef72e61a31d6519ead6f4d9fb9efe2a2c94990 Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <pe...@eisentraut.org>
Date: Mon, 27 Jan 2020 12:14:36 +0100
Subject: [PATCH v6] Fail if recovery target is not reached
Before, if a recovery target is configured, but the archive ended
before the target was reached, recovery would end and the server would
promote without further notice. That was deemed to be pretty wrong.
With this change, if the recovery target is not reached, it is a fatal
error.
Discussion:
https://www.postgresql.org/message-id/flat/993736dd3f1713ec1f63fc3b65383...@lako.no
---
doc/src/sgml/config.sgml | 5 ++++
src/backend/access/transam/xlog.c | 19 +++++++++---
src/test/perl/PostgresNode.pm | 33 +++++++++++++++++++--
src/test/recovery/t/003_recovery_targets.pl | 24 ++++++++++++++-
4 files changed, 73 insertions(+), 8 deletions(-)
diff --git a/doc/src/sgml/config.sgml b/doc/src/sgml/config.sgml
index e07dc01e80..c1128f89ec 100644
--- a/doc/src/sgml/config.sgml
+++ b/doc/src/sgml/config.sgml
@@ -3571,6 +3571,11 @@ <title>Recovery Target</title>
If <xref linkend="guc-hot-standby"/> is not enabled, a setting of
<literal>pause</literal> will act the same as
<literal>shutdown</literal>.
</para>
+ <para>
+ In any case, if a recovery target is configured but the archive
+ recovery ends before the target is reached, the server will shut down
+ with a fatal error.
+ </para>
</listitem>
</varlistentry>
diff --git a/src/backend/access/transam/xlog.c
b/src/backend/access/transam/xlog.c
index 882d5e8a73..be4c923ab1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -6200,7 +6200,7 @@ StartupXLOG(void)
XLogCtlInsert *Insert;
CheckPoint checkPoint;
bool wasShutdown;
- bool reachedStopPoint = false;
+ bool reachedRecoveryTarget = false;
bool haveBackupLabel = false;
bool haveTblspcMap = false;
XLogRecPtr RecPtr,
@@ -7103,7 +7103,7 @@ StartupXLOG(void)
*/
if (recoveryStopsBefore(xlogreader))
{
- reachedStopPoint = true; /* see
below */
+ reachedRecoveryTarget = true;
break;
}
@@ -7258,7 +7258,7 @@ StartupXLOG(void)
/* Exit loop if we reached inclusive recovery
target */
if (recoveryStopsAfter(xlogreader))
{
- reachedStopPoint = true;
+ reachedRecoveryTarget = true;
break;
}
@@ -7270,7 +7270,7 @@ StartupXLOG(void)
* end of main redo apply loop
*/
- if (reachedStopPoint)
+ if (reachedRecoveryTarget)
{
if (!reachedConsistency)
ereport(FATAL,
@@ -7327,7 +7327,18 @@ StartupXLOG(void)
/* there are no WAL records following the checkpoint */
ereport(LOG,
(errmsg("redo is not required")));
+
}
+
+ /*
+ * This check is intentionally after the above log messages that
+ * indicate how far recovery went.
+ */
+ if (ArchiveRecoveryRequested &&
+ recoveryTarget != RECOVERY_TARGET_UNSET &&
+ !reachedRecoveryTarget)
+ ereport(FATAL,
+ (errmsg("recovery ended before
configured recovery target was reached")));
}
/*
diff --git a/src/test/perl/PostgresNode.pm b/src/test/perl/PostgresNode.pm
index 2e0cf4a2f3..be44e8784f 100644
--- a/src/test/perl/PostgresNode.pm
+++ b/src/test/perl/PostgresNode.pm
@@ -653,6 +653,9 @@ Restoring WAL segments from archives using restore_command
can be enabled
by passing the keyword parameter has_restoring => 1. This is disabled by
default.
+Is has_restoring is used, standby mode is used by default. To use
+recovery mode instead, pass the keyword parameter standby => 0.
+
The backup is copied, leaving the original unmodified. pg_hba.conf is
unconditionally set to enable replication connections.
@@ -669,6 +672,7 @@ sub init_from_backup
$params{has_streaming} = 0 unless defined $params{has_streaming};
$params{has_restoring} = 0 unless defined $params{has_restoring};
+ $params{standby} = 1 unless defined $params{standby};
print
"# Initializing node \"$node_name\" from backup \"$backup_name\" of
node \"$root_name\"\n";
@@ -699,7 +703,7 @@ port = $port
"unix_socket_directories = '$host'");
}
$self->enable_streaming($root_node) if $params{has_streaming};
- $self->enable_restoring($root_node) if $params{has_restoring};
+ $self->enable_restoring($root_node, $params{standby}) if
$params{has_restoring};
return;
}
@@ -939,7 +943,7 @@ primary_conninfo='$root_connstr'
# Internal routine to enable archive recovery command on a standby node
sub enable_restoring
{
- my ($self, $root_node) = @_;
+ my ($self, $root_node, $standby) = @_;
my $path = TestLib::perl2host($root_node->archive_dir);
my $name = $self->name;
@@ -961,7 +965,30 @@ sub enable_restoring
'postgresql.conf', qq(
restore_command = '$copy_command'
));
- $self->set_standby_mode();
+ if ($standby)
+ {
+ $self->set_standby_mode();
+ }
+ else
+ {
+ $self->set_recovery_mode();
+ }
+ return;
+}
+
+=pod
+
+=item $node->set_recovery_mode()
+
+Place recovery.signal file.
+
+=cut
+
+sub set_recovery_mode
+{
+ my ($self) = @_;
+
+ $self->append_conf('recovery.signal', '');
return;
}
diff --git a/src/test/recovery/t/003_recovery_targets.pl
b/src/test/recovery/t/003_recovery_targets.pl
index d8fbd50011..fd14bab208 100644
--- a/src/test/recovery/t/003_recovery_targets.pl
+++ b/src/test/recovery/t/003_recovery_targets.pl
@@ -3,7 +3,8 @@
use warnings;
use PostgresNode;
use TestLib;
-use Test::More tests => 8;
+use Test::More tests => 9;
+use Time::HiRes qw(usleep);
# Create and test a standby from given backup, with a certain recovery target.
# Choose $until_lsn later than the transaction commit that causes the row
@@ -145,3 +146,24 @@ sub test_recovery_standby
my $logfile = slurp_file($node_standby->logfile());
ok($logfile =~ qr/multiple recovery targets specified/,
'multiple conflicting settings');
+
+# Check behavior when recovery ends before target is reached
+
+$node_standby = get_new_node('standby_8');
+$node_standby->init_from_backup($node_master, 'my_backup',
+ has_restoring
=> 1, standby => 0);
+$node_standby->append_conf('postgresql.conf',
+ "recovery_target_name =
'does_not_exist'");
+
+run_log(['pg_ctl', '-D', $node_standby->data_dir,
+ '-l', $node_standby->logfile, 'start']);
+
+# wait up to 10 seconds for postgres to terminate
+foreach my $i (0..100)
+{
+ last if ! -f $node_standby->data_dir . '/postmaster.pid';
+ usleep(100_000);
+}
+$logfile = slurp_file($node_standby->logfile());
+ok($logfile =~ qr/FATAL: recovery ended before configured recovery target was
reached/,
+ 'recovery end before target reached is a fatal error');
base-commit: 3e4818e9dd5be294d97ca67012528cb1c0b0ccaa
--
2.25.0
