Compare commits

..

64 Commits

Author SHA1 Message Date
Ian Barwick
922b8de46b doc: link to main documentation section about RemoveIPC 2021-07-26 11:05:49 +09:00
Ian Barwick
92cd8c30b5 doc: link to service commands section from switchover docs 2021-07-26 09:48:02 +09:00
Ian Barwick
fa234919fb doc: document pg_bindir setting
Per suggestion in GitHub #705.
2021-07-20 13:41:00 +09:00
Ian Barwick
03c5220e87 doc: link to sample configuration file
Unfortunately it hasn't been possible yet to include all available
configuration items in the main documentation, but we should at least
make it easier to find the full list.
2021-07-20 13:40:57 +09:00
Ian Barwick
65bd0878fa doc: update release notes 2021-07-01 13:42:30 +09:00
Ian Barwick
f1592f009a repmgrd: ensure short option "-s" is accepted
The long option --show-pid-file was fine.
2021-07-01 13:33:31 +09:00
Ian Barwick
490e51fd7b Remove reference to PostgreSQL 9.3 in --help output
It is not supported by repmgr 5.2 and later.
2021-07-01 13:33:15 +09:00
Ian Barwick
f5c555a7f5 Fix incorrect comment 2021-07-01 13:33:03 +09:00
Ian Barwick
de51df9759 node rejoin: emit rejoin target note information as NOTICE
As it's possible to specify the connection information for any available
node, but currently not possible to rejoin to a node other than the
primary, explicitly mention what the rejoin target will be.
2021-07-01 13:32:13 +09:00
Ian Barwick
b094d7b407 standby clone: set "slot_name" in node record if required
If executing "repmgr standby clone --replication-conf-only" on a node
which was set up without replication slots, but the repmgr configuration
was since changed to "use_replication_slots=1", repmgr will attempt to
create the replication slot. This will however fail if "slot_name"
is not set in the node's record, so have repmgr set the slot_name in
this case.

It might be preferable to preemptively create the slot name for each
node when configuring the cluster, however this would be a behavioural
change which would be better off in a major release (for example, it's
conceivable a user runs sanity checks on the node records and expects
to find the slot names empty if replication slots are not in use).
2021-07-01 13:11:11 +09:00
Ian Barwick
44cdb7d001 doc: clarify "connection_check_type='query'" 2021-03-03 13:07:37 +09:00
Ian Barwick
f5f8db8ca1 docs: update repmgr.conf.sample
Fix description for connection_check_type='connection'.
2021-03-02 11:47:24 +09:00
Ian Barwick
0c568da254 docs: update README
Note latest version number.
2021-03-02 11:22:39 +09:00
Ian Barwick
46042298f7 doc: update README
Fix and update broken link.
2021-03-02 11:15:58 +09:00
Ian Barwick
01e89c3ca9 doc: update GitHub links to new location 2021-03-02 10:22:36 +09:00
Ian Barwick
7b97a8f04d doc: remove generated .fo files 2021-03-02 10:22:23 +09:00
Ian Barwick
c362d6d03f Change copyright information to "EnterpriseDB Corporation"
RM20485.
2021-03-02 10:22:17 +09:00
Ian Barwick
75a2e2a8b4 doc: fix XML markup
An incorrect column count was causing PDF builds to fail.
2021-02-19 20:39:25 +09:00
Ian Barwick
65d59db6d1 doc: update repmgr.conf.sample 2021-01-14 15:28:04 +09:00
Ian Barwick
10cd15bedc doc: "repmgr node rejoin" clarifications
- make it clearer a node can only be joined to the primary
- update patch status
2021-01-06 12:37:14 +09:00
Ian Barwick
cae2255d58 Update copyright notices to 2021 2021-01-04 12:57:26 +09:00
Ian Barwick
bc0bc2696e Add missing PQconninfoFree() call 2020-12-24 18:07:50 +09:00
Ian Barwick
7db9afdc10 repmgrd: edit code comment for clarity 2020-12-22 13:59:29 +09:00
Ian Barwick
5eb8bccd33 doc: minor grammar tweak 2020-12-22 13:59:16 +09:00
Josh Soref
4760c11937 doc: various spelling fixes
Via GitHub #687.
2020-12-22 13:48:29 +09:00
Josh Soref
cb6501aa48 Fix various typos in code comments.
Via GitHub #687.
2020-12-22 13:48:25 +09:00
Josh Soref
6100007a32 repmgr: various log ouput typo fixes
Via GitHub #687.
2020-12-22 13:21:20 +09:00
Josh Soref
a415d85530 repmgr: fix typo in "repmgr node --help" output
Via GitHub #687.
2020-12-22 13:21:15 +09:00
Ian Barwick
45a9a784e3 Improve HINT about upgrading the repmgr extension
Per feedback in GitHub #685.
2020-12-15 08:42:41 +09:00
Ian Barwick
8ce212327c standby switchover: remove extraneous space in log message 2020-12-15 08:42:36 +09:00
Ian Barwick
54851e8df1 Finalize release date 2020-12-07 17:54:14 +09:00
Ian Barwick
a2166d0024 doc: add 5.2.1 release date 2020-12-07 17:53:38 +09:00
Ian Barwick
abe55e60e6 doc: update 5.2.1 release notes 2020-12-04 14:50:04 +09:00
Ian Barwick
220bcbd620 Minor string formatting optimization 2020-12-04 10:18:05 +09:00
Ian Barwick
e824dfd499 standby clone: tweak error message
Probably a remnant from the 9.1 era, where it was not possible to
take a base backup from a standby.
2020-12-04 10:18:02 +09:00
Ian Barwick
9fb9decf13 Fix return value of pg_reload_conf() database utility function
Would always return "false", but as the value wasn't used anywhere,
the issue was inconsequential.

However while we're at it, actually check the return value in the
two places it's called, to help diagnose any issues in the unlikely
event they occur.

Per issue reported via GitHub PR #671 from user duzhgg.
2020-12-02 09:27:14 +09:00
Ian Barwick
ad6dde4218 doc: update README
- remove partial sentence
- remove links to very dated blog entries
2020-12-02 09:26:06 +09:00
Ian Barwick
7cbaec6469 doc: update README
Link to release notes as a simple way of providing the latest release
information.
2020-12-02 09:25:59 +09:00
Ian Barwick
9d2c5921ee Add missing connection close
In a corner-case situation where a standby is unable to attach to
the new primary due to a mismatch in the WAL stream, the connection
used to verify the recovery status of the new primary was not being
closed, leading to a risk of connection exhaustion on the new primary.

Addresses GitHub #682.
2020-12-02 09:23:00 +09:00
Ian Barwick
21f94e6de3 doc: tweak "repmgr standby clone" reference
As recovery.conf starts to fade away, mention that last.
2020-12-02 09:22:49 +09:00
Ian Barwick
7418c7b8f0 Fix typo 2020-12-02 09:22:38 +09:00
Ian Barwick
7cee09dd95 standby clone: fix data directory permissions handling for Pg11 and later
Previously, repmgr would forcibly change the permissions on a data
directory to 0700. However from PostgreSQL 11, 0750 is also valid,
so that value should not be changed.
2020-12-01 11:49:13 +09:00
Ian Barwick
53774d6998 Bump version to 5.2.1 2020-11-30 17:33:42 +09:00
Ian Barwick
5a251ef268 standby clone: add --recovery-min-apply-delay to help output 2020-11-30 16:43:17 +09:00
Ian Barwick
b83ce6b147 doc: update 5.2.1 release notes 2020-11-30 16:14:37 +09:00
Ian Barwick
5f9f1f65ae repmgrd: fix issue with incorrect reconnect_interval
Addresses GitHub #673.
2020-11-25 20:42:29 +09:00
Ian Barwick
9d7eebef1b Update Makefile
We don't actually need $(LIBS) in there; this was cargo-culted in
from somewhere.
2020-11-24 17:37:54 +09:00
Ian Barwick
5cff7fab64 Avoid compiler warnings for various strncpy() operations
Here the compiler may complain that the source length is being used,
though in all cases the source length was previously used to
define the length of the destination buffer, so it's not actually
a problem.
2020-11-24 15:42:54 +09:00
Ian Barwick
2a8ac36aec repmgr: prevent termination in corner-case situation
If neither the local node nor the upstream are available, and
"standby_disconnect_on_failover" is set, attempting to fetch
the walreceiver PID will result in repmgrd terminating.

Add a check that the connection is valid before attempting to
fetch the walreceiver PID.

Addresses GitHub #675.
2020-11-17 16:35:03 +09:00
Ian Barwick
a5a5b506f9 standby clone: various clarifications for --replication-conf-only option
In particular, the emitted HINT was not really appropriate for Pg13 and
later.
2020-11-17 10:01:30 +09:00
Ian Barwick
bd6871817d Update corner-case error message
Not possible to build repmgr compatible with Pg12+ against Pg11
and earlier due to the addition of FullTransactionId.
2020-11-17 10:01:26 +09:00
Ian Barwick
bd35e503ec standby clone: add option --recovery-min-apply-delay
This overrides the equivalent setting in repmgr.conf, if present.

Note this option was available in repmgr versions prior to 4.0, but
was assumed to be redundant. However recently a use-case was made
for its reintroduction.
2020-11-10 16:05:14 +09:00
Romain Jacquier
15801b8f9e Fix help witness
Fix the `repmgr witness --help` command where at the "Unregister" section the message shown was
```
"witness register" unregisters a witness node.
```
instead of
```
"witness unregister" unregisters a witness node.
```

GitHub #676.
2020-11-09 13:35:23 +09:00
Ian Barwick
4d3262d306 standby clone: handle missing "postgresql.auto.conf"
In PostgreSQL 12 and later we need to append replication configuration
to "postgresql.auto.conf" to guarantee it will be read last, and hence
override any preceding replication configuration which may be haunting
the configuration files.

We've been assuming that "postgresql.auto.conf" will always be present,
but at least one corner case has been observed where that was not the
case on the node being cloned from. Moreover it's perfectly acceptable
that this file does not exist (it will be recreated the next time
ALTER SYSTEM is executed), so we should be prepared to handle that case.

In passing, improve handling of more unlikely errors which might be
encountered when processing "postgresql.auto.conf".
2020-10-30 12:26:56 +09:00
Ian Barwick
36a8dfcf4f Standardize code style 2020-10-30 11:06:46 +09:00
Ian Barwick
94612a336a config: fix parsing of "replication_type"
This is a legacy parameter which can currently only contain one value,
"physical" (the default).

It can be safely omitted.

Addresses GitHub #672.
2020-10-30 10:17:03 +09:00
Ian Barwick
de567d584f doc: update README 2020-10-22 21:21:16 +09:00
Ian Barwick
0fb3432398 Finalize release date 2020-10-22 17:15:37 +09:00
Ian Barwick
5e59e543d6 Additional fix to upgrade script
Drop old "repl_events" table.
2020-10-22 12:36:31 +09:00
Ian Barwick
4f6b642320 Fix extension script for unpackaged upgrades to 5.2
Apparently "ALTER TABLE" (which we were using to convert the
"repl_events" table) does not mark the table as being part of the
extension. Instead, we need to create the new table and copy the
data, as is done with the other tables.
2020-10-22 11:18:03 +09:00
Ian Barwick
c1973438f7 doc: add missing "unpackaged" reference 2020-10-21 17:31:22 +09:00
Ian Barwick
fe5904e04e standby clone: improve Barman source server check
Use "remote_command()" to execute the remote psql command, and
provide the -X option to psql to ensure it doesn't read ~/.psqlrc.
2020-10-20 17:21:17 +09:00
Ian Barwick
2cb1f4f728 doc: update release notes 2020-10-20 14:05:25 +09:00
Ian Barwick
5b90842c55 Bump version to 5.2.0
Also update the minimum version check to PostgreSQL 9.4.
2020-10-20 13:34:16 +09:00
31 changed files with 249 additions and 1227 deletions

View File

@@ -1,9 +1,4 @@
5.3.0 2021-10-12
standby switchover: improve handling of node rejoin failure (Ian)
repmgrd: prefix all shared library functions with "repmgr_" to
minimize the risk of clashes with other shared libraries (Ian)
repmgrd: at startup, if node record is marked as "inactive", attempt
to set it to "active" (Ian)
5.2.2. 2021-??-??
standby clone: set "slot_name" in node record if required (Ian)
node rejoin: emit rejoin target note information as NOTICE (Ian)
repmgrd: ensure short option "-s" is accepted (Ian)

View File

@@ -13,7 +13,6 @@ DATA = \
repmgr--unpackaged--4.0.sql \
repmgr--unpackaged--5.1.sql \
repmgr--unpackaged--5.2.sql \
repmgr--unpackaged--5.3.sql \
repmgr--4.0.sql \
repmgr--4.0--4.1.sql \
repmgr--4.1.sql \
@@ -28,9 +27,7 @@ DATA = \
repmgr--5.0--5.1.sql \
repmgr--5.1.sql \
repmgr--5.1--5.2.sql \
repmgr--5.2.sql \
repmgr--5.2--5.3.sql \
repmgr--5.3.sql
repmgr--5.2.sql
REGRESS = repmgr_extension

View File

@@ -90,4 +90,3 @@ Further reading
* [repmgr documentation](https://repmgr.org/docs/current/index.html)
* [How to Automate PostgreSQL 12 Replication and Failover with repmgr - Part 1](https://www.2ndquadrant.com/en/blog/how-to-automate-postgresql-12-replication-and-failover-with-repmgr-part-1/)
* [How to Automate PostgreSQL 12 Replication and Failover with repmgr - Part 2](https://www.2ndquadrant.com/en/blog/how-to-automate-postgresql-12-replication-and-failover-with-repmgr-part-2/)
* [How to implement repmgr for PostgreSQL automatic failover](https://www.enterprisedb.com/postgres-tutorials/how-implement-repmgr-postgresql-automatic-failover)

View File

@@ -581,16 +581,6 @@ struct ConfigFileSetting config_file_settings[] =
{ .strmaxlen = sizeof(config_file_options.repmgrd_pid_file) },
{ .postprocess_func = &repmgr_canonicalize_path }
},
/* repmgrd_exit_on_inactive_node */
{
"repmgrd_exit_on_inactive_node",
CONFIG_BOOL,
{ .boolptr = &config_file_options.repmgrd_exit_on_inactive_node},
{ .booldefault = DEFAULT_REPMGRD_EXIT_ON_INACTIVE_NODE },
{},
{},
{}
},
/* standby_disconnect_on_failover */
{
"standby_disconnect_on_failover",

View File

@@ -206,7 +206,6 @@ typedef struct
int primary_notification_timeout;
int repmgrd_standby_startup_timeout;
char repmgrd_pid_file[MAXPGPATH];
bool repmgrd_exit_on_inactive_node;
bool standby_disconnect_on_failover;
int sibling_nodes_disconnect_timeout;
ConnectionCheckType connection_check_type;

24
configure vendored
View File

@@ -1,6 +1,6 @@
#! /bin/sh
# Guess values for system-dependent variables and create Makefiles.
# Generated by GNU Autoconf 2.69 for repmgr 5.3.0.
# Generated by GNU Autoconf 2.69 for repmgr 5.2.1.
#
# Report bugs to <repmgr@googlegroups.com>.
#
@@ -582,8 +582,8 @@ MAKEFLAGS=
# Identity of this package.
PACKAGE_NAME='repmgr'
PACKAGE_TARNAME='repmgr'
PACKAGE_VERSION='5.3.0'
PACKAGE_STRING='repmgr 5.3.0'
PACKAGE_VERSION='5.2.1'
PACKAGE_STRING='repmgr 5.2.1'
PACKAGE_BUGREPORT='repmgr@googlegroups.com'
PACKAGE_URL='https://repmgr.org/'
@@ -1181,7 +1181,7 @@ if test "$ac_init_help" = "long"; then
# Omit some internal or obsolete options to make the list less imposing.
# This message is too long to be a string in the A/UX 3.1 sh.
cat <<_ACEOF
\`configure' configures repmgr 5.3.0 to adapt to many kinds of systems.
\`configure' configures repmgr 5.2.1 to adapt to many kinds of systems.
Usage: $0 [OPTION]... [VAR=VALUE]...
@@ -1242,7 +1242,7 @@ fi
if test -n "$ac_init_help"; then
case $ac_init_help in
short | recursive ) echo "Configuration of repmgr 5.3.0:";;
short | recursive ) echo "Configuration of repmgr 5.2.1:";;
esac
cat <<\_ACEOF
@@ -1316,7 +1316,7 @@ fi
test -n "$ac_init_help" && exit $ac_status
if $ac_init_version; then
cat <<\_ACEOF
repmgr configure 5.3.0
repmgr configure 5.2.1
generated by GNU Autoconf 2.69
Copyright (C) 2012 Free Software Foundation, Inc.
@@ -1335,7 +1335,7 @@ cat >config.log <<_ACEOF
This file contains any messages produced by compilers while
running configure, to aid debugging if configure makes a mistake.
It was created by repmgr $as_me 5.3.0, which was
It was created by repmgr $as_me 5.2.1, which was
generated by GNU Autoconf 2.69. Invocation command line was
$ $0 $@
@@ -1811,11 +1811,11 @@ fi
pgac_pg_config_version=$($PG_CONFIG --version 2>/dev/null)
major_version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[^0-9]\+ \([0-9]\{1,2\}\).*$/\1/')
$SED -e 's/^PostgreSQL \([0-9]\{1,2\}\).*$/\1/')
if test "$major_version_num" -lt '10'; then
version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[^0-9]\+ \([0-9]*\)\.\([0-9]*\)\([a-zA-Z0-9.]*\)$/\1.\2/')
$SED -e 's/^PostgreSQL \([0-9]*\)\.\([0-9]*\)\([a-zA-Z0-9.]*\)$/\1.\2/')
if test -z "$version_num"; then
as_fn_error $? "could not detect the PostgreSQL version, wrong or broken pg_config?" "$LINENO" 5
@@ -1829,7 +1829,7 @@ if test "$major_version_num" -lt '10'; then
fi
else
version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[^0-9]\+ \(.\+\)$/\1/')
$SED -e 's/^PostgreSQL \(.\+\)$/\1/')
if test -z "$version_num"; then
as_fn_error $? "could not detect the PostgreSQL version, wrong or broken pg_config?" "$LINENO" 5
@@ -2487,7 +2487,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
# report actual input values of CONFIG_FILES etc. instead of their
# values after options handling.
ac_log="
This file was extended by repmgr $as_me 5.3.0, which was
This file was extended by repmgr $as_me 5.2.1, which was
generated by GNU Autoconf 2.69. Invocation command line was
CONFIG_FILES = $CONFIG_FILES
@@ -2550,7 +2550,7 @@ _ACEOF
cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
ac_cs_version="\\
repmgr config.status 5.3.0
repmgr config.status 5.2.1
configured by $0, generated by GNU Autoconf 2.69,
with options \\"\$ac_cs_config\\"

View File

@@ -1,4 +1,4 @@
AC_INIT([repmgr], [5.3.0], [repmgr@googlegroups.com], [repmgr], [https://repmgr.org/])
AC_INIT([repmgr], [5.2.1], [repmgr@googlegroups.com], [repmgr], [https://repmgr.org/])
AC_COPYRIGHT([Copyright (c) 2010-2021, EnterpriseDB Corporation])
@@ -19,11 +19,11 @@ fi
pgac_pg_config_version=$($PG_CONFIG --version 2>/dev/null)
major_version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[[^0-9]]\+ \([[0-9]]\{1,2\}\).*$/\1/')
$SED -e 's/^PostgreSQL \([[0-9]]\{1,2\}\).*$/\1/')
if test "$major_version_num" -lt '10'; then
version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[[^0-9]]\+ \([[0-9]]*\)\.\([[0-9]]*\)\([[a-zA-Z0-9.]]*\)$/\1.\2/')
$SED -e 's/^PostgreSQL \([[0-9]]*\)\.\([[0-9]]*\)\([[a-zA-Z0-9.]]*\)$/\1.\2/')
if test -z "$version_num"; then
AC_MSG_ERROR([could not detect the PostgreSQL version, wrong or broken pg_config?])
@@ -37,7 +37,7 @@ if test "$major_version_num" -lt '10'; then
fi
else
version_num=$(echo "$pgac_pg_config_version"|
$SED -e 's/^[[^0-9]]\+ \(.\+\)$/\1/')
$SED -e 's/^PostgreSQL \(.\+\)$/\1/')
if test -z "$version_num"; then
AC_MSG_ERROR([could not detect the PostgreSQL version, wrong or broken pg_config?])

View File

@@ -4242,7 +4242,7 @@ _create_event(PGconn *conn, t_configuration_options *options, int node_id, char
}
break;
case 'p':
/* %p: primary id ("standby_switchover"/"repmgrd_failover_promote": former primary id) */
/* %p: primary id ("standby_switchover": former primary id) */
src_ptr++;
if (event_info->node_id != UNKNOWN_NODE_ID)
{
@@ -6008,43 +6008,6 @@ is_wal_replay_paused(PGconn *conn, bool check_pending_wal)
return is_paused;
}
/* repmgrd status functions */
CheckStatus
get_repmgrd_status(PGconn *conn)
{
PQExpBufferData query;
PGresult *res = NULL;
CheckStatus repmgrd_status = CHECK_STATUS_CRITICAL;
initPQExpBuffer(&query);
appendPQExpBufferStr(&query,
" SELECT "
" CASE "
" WHEN repmgr.repmgrd_is_running() "
" THEN "
" CASE "
" WHEN repmgr.repmgrd_is_paused() THEN 1 ELSE 0 "
" END "
" ELSE 2 "
" END AS repmgrd_status");
res = PQexec(conn, query.data);
if (PQresultStatus(res) != PGRES_TUPLES_OK)
{
log_db_error(conn, query.data, _("unable to execute repmgrd status query"));
}
else
{
repmgrd_status = atoi(PQgetvalue(res, 0, 0));
}
termPQExpBuffer(&query);
PQclear(res);
return repmgrd_status;
}
/* miscellaneous debugging functions */

View File

@@ -602,9 +602,6 @@ int get_upstream_last_seen(PGconn *conn, t_server_type node_type);
bool is_wal_replay_paused(PGconn *conn, bool check_pending_wal);
/* repmgrd status functions */
CheckStatus get_repmgrd_status(PGconn *conn);
/* miscellaneous debugging functions */
const char *print_node_status(NodeStatus node_status);
const char *print_pqping_status(PGPing ping_status);

View File

@@ -16,84 +16,17 @@
</para>
<!-- remember to update the release date in ../repmgr_version.h.in -->
<sect1 id="release-5.3.0">
<title id="release-current">Release 5.3.0</title>
<para><emphasis>Tue 12 October, 2021</emphasis></para>
<sect1 id="release-5.2.2">
<title id="release-current">Release 5.2.2</title>
<para><emphasis>??? ? ???, 2021</emphasis></para>
<para>
&repmgr; 5.3.0 is a major release.
&repmgr; 5.2.2 is a minor release.
</para>
<para>
This release provides support for <ulink url="https://www.postgresql.org/docs/14/release-14.html">PostgreSQL 14</ulink>,
released in September 2021.
</para>
<sect2>
<title>Improvements</title>
<para>
<itemizedlist>
<listitem>
<para>
<link linkend="repmgr-standby-switchover"><command>repmgr standby switchover</command></link>:
Improve handling of node rejoin failure on the demotion candidate.
</para>
<para>
Previously &repmgr; did not check whether <command>repmgr node rejoin</command> actually
succeeded on the demotion candidate, and would always wait up to <varname>node_rejoin_timeout</varname>
seconds for it to attach to the promotion candidate, even if this would never happen.
</para>
<para>
This makes it easier to identify unexpected events during a switchover operation, such as
the demotion candidate being unexpectedly restarted by an external process.
</para>
<para>
Note that the output of the <link linkend="repmgr-node-rejoin"><command>repmgr node rejoin</command></link>
operation on the demotion candidate will now be logged to a temporary file on that node;
the location of the file will be reported in the error message, if one is emitted.
</para>
</listitem>
<listitem>
<para>
&repmgrd;: at startup, if node record is marked as "inactive", attempt
to set it to "active".
</para>
<para>
This behaviour can be overridden by setting the configuration parameter
<varname>repmgrd_exit_on_inactive_node</varname> to <literal>true</literal>.
</para>
</listitem>
<listitem>
<para>
<command><link linkend="repmgr-node-rejoin">repmgr node rejoin</link></command>:
emit rejoin target note information as <literal>NOTICE</literal>.
</para>
<para>
This makes it clearer what &repmgr; is trying to do.
</para>
</listitem>
<listitem>
<para>
<link linkend="repmgr-node-check">repmgr node check</link>:
option <option>--repmgrd</option> added to check &repmgrd;
status.
</para>
</listitem>
<listitem>
<para>
Add <literal>%p</literal> <link linkend="event-notifications">event notification parameter</link>
providing the node ID of the former primary for the <literal>repmgrd_failover_promote</literal> event.
</para>
</listitem>
</itemizedlist>
</para>
</sect2>
<sect2>
<title>Bug fixes</title>
<para>
<itemizedlist>
<itemizedlist>
<listitem>
<para>
<command><link linkend="repmgr-standby-clone">repmgr standby clone</link></command>:
@@ -105,24 +38,22 @@
</para>
</listitem>
<listitem>
<para>
&repmgrd;: rename internal shared library functions to minimize the
risk of clashes with other shared libraries.
</para>
<para>
This does not affect user-facing SQL functions. However an upgrade
of the installed extension version is required.
</para>
</listitem>
<para>
<command><link linkend="repmgr-node-rejoin">repmgr node rejoin</link></command>:
emit rejoin target note information as <literal>NOTICE</literal>.
</para>
<para>
This makes it clearer what &repmgr; is trying to do.
</para>
</listitem>
<listitem>
<para>
&repmgrd;: ensure short option <option>-s</option> is accepted.
</para>
</listitem>
</itemizedlist>
</itemizedlist>
</para>
</sect2>
</sect1>

View File

@@ -95,8 +95,7 @@
</para>
<para>
The following parameters are provided for a subset of event notifications; their meaning may
change according to context:
The following parameters are provided for a subset of event notifications:
</para>
<variablelist>
@@ -109,9 +108,6 @@
<para>
node ID of the demoted primary (<xref linkend="repmgr-standby-switchover"/> only)
</para>
<para>
node ID of the former primary (<literal>repmgrd_failover_promote</literal> only)
</para>
</listitem>
</varlistentry>
<varlistentry>
@@ -137,7 +133,7 @@
<para>
The values provided for <literal>%c</literal> and <literal>%a</literal>
may contain spaces, so should always be quoted.
will probably contain spaces, so should always be quoted.
</para>
<para>

View File

@@ -125,29 +125,12 @@
is correctly configured.
</simpara>
</listitem>
</itemizedlist>
</para>
</refsect1>
<refsect1>
<title>repmgrd</title>
<para>
A separate check is available to verify whether &repmgrd; is running,
This is not included in the general output, as this does not
per-se constitute a check of the node's replication status.
</para>
<itemizedlist spacing="compact" mark="bullet">
<listitem>
<simpara>
<option>--repmgrd</option>: checks whether &repmgrd; is running.
If &repmgrd; is running but paused, status <literal>1</literal>
(<literal>WARNING</literal>) is returned.
</simpara>
</listitem>
</itemizedlist>
</refsect1>
<refsect1>
<title>Additional checks</title>
<para>

View File

@@ -26,7 +26,7 @@
<abstract>
<para>
This is the official documentation of &repmgr; &repmgrversion; for
use with PostgreSQL 9.4 - PostgreSQL 14.
use with PostgreSQL 9.4 - PostgreSQL 13.
</para>
<para>
&repmgr; is being continually developed and we strongly recommend using the

View File

@@ -485,32 +485,6 @@
</listitem>
</varlistentry>
<varlistentry>
<term><option>repmgrd_exit_on_inactive_node</option></term>
<listitem>
<indexterm>
<primary>repmgrd_exit_on_inactive_node</primary>
</indexterm>
<para>
This parameter is available in &repmgr; 5.3 and later.
</para>
<para>
If a node was marked as inactive but is running, and this option is set to
<literal>true</literal>, &repmgrd; will abort on startup.
</para>
<para>
By default, <option>repmgrd_exit_on_inactive_node</option> is set
to <literal>false</literal>, in which case &repmgrd; will set the
node record to active on startup.
</para>
<para>
Setting this parameter to <literal>true</literal> causes &repmgrd;
to behave in the same way it did in &repmgr; 5.2 and earlier.
</para>
</listitem>
</varlistentry>
</variablelist>
<para>
@@ -1079,29 +1053,6 @@ REPMGRD_OPTS="--daemonize=false"
</para>
</sect2>
<sect2 id="repmgrd-daemon-monitoring">
<title>repmgrd daemon monitoring</title>
<indexterm>
<primary>repmgrd</primary>
<secondary>monitoring</secondary>
</indexterm>
<indexterm>
<primary>monitoring</primary>
<secondary>repmgrd</secondary>
</indexterm>
<para>
The command <command><link linkend="repmgr-service-status">repmgr service status</link></command>
provides an overview of the &repmgrd; daemon status (including pause status)
on all nodes in the cluster.
</para>
<para>
From &repmgr; 5.3, <command><link linkend="repmgr-node-check">repmgr node check --repmgrd</link></command>
can be used to check the status of &repmgrd; (including pause status)
on the local node.
</para>
</sect2>
</sect1>
<sect1 id="repmgrd-connection-settings">

View File

@@ -1,64 +0,0 @@
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "CREATE EXTENSION repmgr" to load this file. \quit
CREATE OR REPLACE FUNCTION set_local_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_local_node_id'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION repmgr.get_local_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_local_node_id'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION standby_set_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_set_last_updated'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION standby_get_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_get_last_updated'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION set_upstream_last_seen(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_last_seen'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION get_upstream_last_seen()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_last_seen'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION get_upstream_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_node_id'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION set_upstream_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_node_id'
LANGUAGE C STRICT;
/* failover functions */
CREATE OR REPLACE FUNCTION notify_follow_primary(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_notify_follow_primary'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION get_new_primary()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_new_primary'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION reset_voting_status()
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_reset_voting_status'
LANGUAGE C STRICT;
CREATE OR REPLACE FUNCTION get_wal_receiver_pid()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_wal_receiver_pid'
LANGUAGE C STRICT;

View File

@@ -1,192 +0,0 @@
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "CREATE EXTENSION repmgr" to load this file. \quit
CREATE TABLE repmgr.nodes (
node_id INTEGER PRIMARY KEY,
upstream_node_id INTEGER NULL REFERENCES nodes (node_id) DEFERRABLE,
active BOOLEAN NOT NULL DEFAULT TRUE,
node_name TEXT NOT NULL,
type TEXT NOT NULL CHECK (type IN('primary','standby','witness','bdr')),
location TEXT NOT NULL DEFAULT 'default',
priority INT NOT NULL DEFAULT 100,
conninfo TEXT NOT NULL,
repluser VARCHAR(63) NOT NULL,
slot_name TEXT NULL,
config_file TEXT NOT NULL
);
SELECT pg_catalog.pg_extension_config_dump('repmgr.nodes', '');
CREATE TABLE repmgr.events (
node_id INTEGER NOT NULL,
event TEXT NOT NULL,
successful BOOLEAN NOT NULL DEFAULT TRUE,
event_timestamp TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT CURRENT_TIMESTAMP,
details TEXT NULL
);
SELECT pg_catalog.pg_extension_config_dump('repmgr.events', '');
CREATE TABLE repmgr.monitoring_history (
primary_node_id INTEGER NOT NULL,
standby_node_id INTEGER NOT NULL,
last_monitor_time TIMESTAMP WITH TIME ZONE NOT NULL,
last_apply_time TIMESTAMP WITH TIME ZONE,
last_wal_primary_location PG_LSN NOT NULL,
last_wal_standby_location PG_LSN,
replication_lag BIGINT NOT NULL,
apply_lag BIGINT NOT NULL
);
CREATE INDEX idx_monitoring_history_time
ON repmgr.monitoring_history (last_monitor_time, standby_node_id);
SELECT pg_catalog.pg_extension_config_dump('repmgr.monitoring_history', '');
CREATE VIEW repmgr.show_nodes AS
SELECT n.node_id,
n.node_name,
n.active,
n.upstream_node_id,
un.node_name AS upstream_node_name,
n.type,
n.priority,
n.conninfo
FROM repmgr.nodes n
LEFT JOIN repmgr.nodes un
ON un.node_id = n.upstream_node_id;
CREATE TABLE repmgr.voting_term (
term INT NOT NULL
);
CREATE UNIQUE INDEX voting_term_restrict
ON repmgr.voting_term ((TRUE));
CREATE RULE voting_term_delete AS
ON DELETE TO repmgr.voting_term
DO INSTEAD NOTHING;
/* ================= */
/* repmgrd functions */
/* ================= */
/* monitoring functions */
CREATE FUNCTION set_local_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_local_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION get_local_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_local_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION standby_set_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_set_last_updated'
LANGUAGE C STRICT;
CREATE FUNCTION standby_get_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_get_last_updated'
LANGUAGE C STRICT;
CREATE FUNCTION set_upstream_last_seen(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_last_seen'
LANGUAGE C STRICT;
CREATE FUNCTION get_upstream_last_seen()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_last_seen'
LANGUAGE C STRICT;
CREATE FUNCTION get_upstream_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION set_upstream_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_node_id'
LANGUAGE C STRICT;
/* failover functions */
CREATE FUNCTION notify_follow_primary(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_notify_follow_primary'
LANGUAGE C STRICT;
CREATE FUNCTION get_new_primary()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_new_primary'
LANGUAGE C STRICT;
CREATE FUNCTION reset_voting_status()
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_reset_voting_status'
LANGUAGE C STRICT;
CREATE FUNCTION get_repmgrd_pid()
RETURNS INT
AS 'MODULE_PATHNAME', 'get_repmgrd_pid'
LANGUAGE C STRICT;
CREATE FUNCTION get_repmgrd_pidfile()
RETURNS TEXT
AS 'MODULE_PATHNAME', 'get_repmgrd_pidfile'
LANGUAGE C STRICT;
CREATE FUNCTION set_repmgrd_pid(INT, TEXT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'set_repmgrd_pid'
LANGUAGE C CALLED ON NULL INPUT;
CREATE FUNCTION repmgrd_is_running()
RETURNS BOOL
AS 'MODULE_PATHNAME', 'repmgrd_is_running'
LANGUAGE C STRICT;
CREATE FUNCTION repmgrd_pause(BOOL)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgrd_pause'
LANGUAGE C STRICT;
CREATE FUNCTION repmgrd_is_paused()
RETURNS BOOL
AS 'MODULE_PATHNAME', 'repmgrd_is_paused'
LANGUAGE C STRICT;
CREATE FUNCTION get_wal_receiver_pid()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_wal_receiver_pid'
LANGUAGE C STRICT;
/* views */
CREATE VIEW repmgr.replication_status AS
SELECT m.primary_node_id, m.standby_node_id, n.node_name AS standby_name,
n.type AS node_type, n.active, last_monitor_time,
CASE WHEN n.type='standby' THEN m.last_wal_primary_location ELSE NULL END AS last_wal_primary_location,
m.last_wal_standby_location,
CASE WHEN n.type='standby' THEN pg_catalog.pg_size_pretty(m.replication_lag) ELSE NULL END AS replication_lag,
CASE WHEN n.type='standby' THEN
CASE WHEN replication_lag > 0 THEN age(now(), m.last_apply_time) ELSE '0'::INTERVAL END
ELSE NULL
END AS replication_time_lag,
CASE WHEN n.type='standby' THEN pg_catalog.pg_size_pretty(m.apply_lag) ELSE NULL END AS apply_lag,
AGE(NOW(), CASE WHEN pg_catalog.pg_is_in_recovery() THEN repmgr.standby_get_last_updated() ELSE m.last_monitor_time END) AS communication_time_lag
FROM repmgr.monitoring_history m
JOIN repmgr.nodes n ON m.standby_node_id = n.node_id
WHERE (m.standby_node_id, m.last_monitor_time) IN (
SELECT m1.standby_node_id, MAX(m1.last_monitor_time)
FROM repmgr.monitoring_history m1 GROUP BY 1
);

View File

@@ -1,245 +0,0 @@
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "CREATE EXTENSION repmgr" to load this file. \quit
-- extract the current schema name
-- NOTE: this assumes there will be only one schema matching 'repmgr_%';
-- user is responsible for ensuring this is the case
CREATE TEMPORARY TABLE repmgr_old_schema (schema_name TEXT);
INSERT INTO repmgr_old_schema (schema_name)
SELECT nspname AS schema_name
FROM pg_catalog.pg_namespace
WHERE nspname LIKE 'repmgr_%'
LIMIT 1;
-- move old objects into new schema
DO $repmgr$
DECLARE
old_schema TEXT;
BEGIN
SELECT schema_name FROM repmgr_old_schema
INTO old_schema;
EXECUTE format('ALTER TABLE %I.repl_nodes SET SCHEMA repmgr', old_schema);
EXECUTE format('ALTER TABLE %I.repl_events SET SCHEMA repmgr', old_schema);
EXECUTE format('ALTER TABLE %I.repl_monitor SET SCHEMA repmgr', old_schema);
EXECUTE format('DROP VIEW IF EXISTS %I.repl_show_nodes', old_schema);
EXECUTE format('DROP VIEW IF EXISTS %I.repl_status', old_schema);
END$repmgr$;
-- convert "repmgr_$cluster.repl_nodes" to "repmgr.nodes"
CREATE TABLE repmgr.nodes (
node_id INTEGER PRIMARY KEY,
upstream_node_id INTEGER NULL REFERENCES repmgr.nodes (node_id) DEFERRABLE,
active BOOLEAN NOT NULL DEFAULT TRUE,
node_name TEXT NOT NULL,
type TEXT NOT NULL CHECK (type IN('primary','standby','witness','bdr')),
location TEXT NOT NULL DEFAULT 'default',
priority INT NOT NULL DEFAULT 100,
conninfo TEXT NOT NULL,
repluser VARCHAR(63) NOT NULL,
slot_name TEXT NULL,
config_file TEXT NOT NULL
);
INSERT INTO repmgr.nodes
(node_id, upstream_node_id, active, node_name, type, location, priority, conninfo, repluser, slot_name, config_file)
SELECT id, upstream_node_id, active, name,
CASE WHEN type = 'master' THEN 'primary' ELSE type END,
'default', priority, conninfo, 'unknown', slot_name, 'unknown'
FROM repmgr.repl_nodes
ORDER BY id;
-- convert "repmgr_$cluster.repl_event" to "event"
CREATE TABLE repmgr.events (
node_id INTEGER NOT NULL,
event TEXT NOT NULL,
successful BOOLEAN NOT NULL DEFAULT TRUE,
event_timestamp TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT CURRENT_TIMESTAMP,
details TEXT NULL
);
INSERT INTO repmgr.events
(node_id, event, successful, event_timestamp, details)
SELECT node_id, event, successful, event_timestamp, details
FROM repmgr.repl_events;
-- create new table "repmgr.voting_term"
CREATE TABLE repmgr.voting_term (
term INT NOT NULL
);
CREATE UNIQUE INDEX voting_term_restrict
ON repmgr.voting_term ((TRUE));
CREATE RULE voting_term_delete AS
ON DELETE TO repmgr.voting_term
DO INSTEAD NOTHING;
INSERT INTO repmgr.voting_term (term) VALUES (1);
-- convert "repmgr_$cluster.repl_monitor" to "monitoring_history"
CREATE TABLE repmgr.monitoring_history (
primary_node_id INTEGER NOT NULL,
standby_node_id INTEGER NOT NULL,
last_monitor_time TIMESTAMP WITH TIME ZONE NOT NULL,
last_apply_time TIMESTAMP WITH TIME ZONE,
last_wal_primary_location PG_LSN NOT NULL,
last_wal_standby_location PG_LSN,
replication_lag BIGINT NOT NULL,
apply_lag BIGINT NOT NULL
);
INSERT INTO repmgr.monitoring_history
(primary_node_id, standby_node_id, last_monitor_time, last_apply_time, last_wal_primary_location, last_wal_standby_location, replication_lag, apply_lag)
SELECT primary_node, standby_node, last_monitor_time, last_apply_time, last_wal_primary_location::pg_lsn, last_wal_standby_location::pg_lsn, replication_lag, apply_lag
FROM repmgr.repl_monitor;
CREATE INDEX idx_monitoring_history_time
ON repmgr.monitoring_history (last_monitor_time, standby_node_id);
CREATE VIEW repmgr.show_nodes AS
SELECT n.node_id,
n.node_name,
n.active,
n.upstream_node_id,
un.node_name AS upstream_node_name,
n.type,
n.priority,
n.conninfo
FROM repmgr.nodes n
LEFT JOIN repmgr.nodes un
ON un.node_id = n.upstream_node_id;
/* ================= */
/* repmgrd functions */
/* ================= */
/* monitoring functions */
CREATE FUNCTION set_local_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_local_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION get_local_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_local_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION standby_set_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_set_last_updated'
LANGUAGE C STRICT;
CREATE FUNCTION standby_get_last_updated()
RETURNS TIMESTAMP WITH TIME ZONE
AS 'MODULE_PATHNAME', 'repmgr_standby_get_last_updated'
LANGUAGE C STRICT;
CREATE FUNCTION set_upstream_last_seen(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_last_seen'
LANGUAGE C STRICT;
CREATE FUNCTION get_upstream_last_seen()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_last_seen'
LANGUAGE C STRICT;
CREATE FUNCTION get_upstream_node_id()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_upstream_node_id'
LANGUAGE C STRICT;
CREATE FUNCTION set_upstream_node_id(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_set_upstream_node_id'
LANGUAGE C STRICT;
/* failover functions */
CREATE FUNCTION notify_follow_primary(INT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_notify_follow_primary'
LANGUAGE C STRICT;
CREATE FUNCTION get_new_primary()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_new_primary'
LANGUAGE C STRICT;
CREATE FUNCTION reset_voting_status()
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgr_reset_voting_status'
LANGUAGE C STRICT;
CREATE FUNCTION get_repmgrd_pid()
RETURNS INT
AS 'MODULE_PATHNAME', 'get_repmgrd_pid'
LANGUAGE C STRICT;
CREATE FUNCTION get_repmgrd_pidfile()
RETURNS TEXT
AS 'MODULE_PATHNAME', 'get_repmgrd_pidfile'
LANGUAGE C STRICT;
CREATE FUNCTION set_repmgrd_pid(INT, TEXT)
RETURNS VOID
AS 'MODULE_PATHNAME', 'set_repmgrd_pid'
LANGUAGE C CALLED ON NULL INPUT;
CREATE FUNCTION repmgrd_is_running()
RETURNS BOOL
AS 'MODULE_PATHNAME', 'repmgrd_is_running'
LANGUAGE C STRICT;
CREATE FUNCTION repmgrd_pause(BOOL)
RETURNS VOID
AS 'MODULE_PATHNAME', 'repmgrd_pause'
LANGUAGE C STRICT;
CREATE FUNCTION repmgrd_is_paused()
RETURNS BOOL
AS 'MODULE_PATHNAME', 'repmgrd_is_paused'
LANGUAGE C STRICT;
CREATE FUNCTION get_wal_receiver_pid()
RETURNS INT
AS 'MODULE_PATHNAME', 'repmgr_get_wal_receiver_pid'
LANGUAGE C STRICT;
/* views */
CREATE VIEW repmgr.replication_status AS
SELECT m.primary_node_id, m.standby_node_id, n.node_name AS standby_name,
n.type AS node_type, n.active, last_monitor_time,
CASE WHEN n.type='standby' THEN m.last_wal_primary_location ELSE NULL END AS last_wal_primary_location,
m.last_wal_standby_location,
CASE WHEN n.type='standby' THEN pg_catalog.pg_size_pretty(m.replication_lag) ELSE NULL END AS replication_lag,
CASE WHEN n.type='standby' THEN
CASE WHEN replication_lag > 0 THEN age(now(), m.last_apply_time) ELSE '0'::INTERVAL END
ELSE NULL
END AS replication_time_lag,
CASE WHEN n.type='standby' THEN pg_catalog.pg_size_pretty(m.apply_lag) ELSE NULL END AS apply_lag,
AGE(NOW(), CASE WHEN pg_catalog.pg_is_in_recovery() THEN repmgr.standby_get_last_updated() ELSE m.last_monitor_time END) AS communication_time_lag
FROM repmgr.monitoring_history m
JOIN repmgr.nodes n ON m.standby_node_id = n.node_id
WHERE (m.standby_node_id, m.last_monitor_time) IN (
SELECT m1.standby_node_id, MAX(m1.last_monitor_time)
FROM repmgr.monitoring_history m1 GROUP BY 1
);
/* drop old tables */
DROP TABLE repmgr.repl_nodes;
DROP TABLE repmgr.repl_monitor;
DROP TABLE repmgr.repl_events;
-- remove temporary table
DROP TABLE repmgr_old_schema;

View File

@@ -35,7 +35,6 @@
static bool copy_file(const char *src_file, const char *dest_file);
static void format_archive_dir(PQExpBufferData *archive_dir);
static t_server_action parse_server_action(const char *action);
static const char *output_repmgrd_status(CheckStatus status);
static void exit_optformat_error(const char *error, int errcode);
@@ -53,11 +52,9 @@ static CheckStatus do_node_check_role(PGconn *conn, OutputMode mode, t_node_info
static CheckStatus do_node_check_slots(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output);
static CheckStatus do_node_check_missing_slots(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output);
static CheckStatus do_node_check_data_directory(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output);
static CheckStatus do_node_check_repmgrd(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output);
static CheckStatus do_node_check_replication_config_owner(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output);
static CheckStatus do_node_check_db_connection(PGconn *conn, OutputMode mode);
/*
* NODE STATUS
*
@@ -944,16 +941,6 @@ do_node_check(void)
exit(return_code);
}
if (runtime_options.repmgrd == true)
{
return_code = do_node_check_repmgrd(conn,
runtime_options.output_mode,
&node_info,
NULL);
PQfinish(conn);
exit(return_code);
}
if (runtime_options.replication_config_owner == true)
{
return_code = do_node_check_replication_config_owner(conn,
@@ -2037,6 +2024,7 @@ do_node_check_missing_slots(PGconn *conn, OutputMode mode, t_node_info *node_inf
return status;
}
CheckStatus
do_node_check_data_directory(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output)
{
@@ -2171,53 +2159,6 @@ do_node_check_data_directory(PGconn *conn, OutputMode mode, t_node_info *node_in
return status;
}
CheckStatus
do_node_check_repmgrd(PGconn *conn, OutputMode mode, t_node_info *node_info, CheckStatusList *list_output)
{
CheckStatus status = CHECK_STATUS_OK;
if (mode == OM_CSV && list_output == NULL)
{
log_error(_("--csv output not provided with --repmgrd option"));
PQfinish(conn);
exit(ERR_BAD_CONFIG);
}
status = get_repmgrd_status(conn);
switch (mode)
{
case OM_OPTFORMAT:
printf("--repmgrd=%s\n",
output_check_status(status));
break;
case OM_NAGIOS:
printf("REPMGRD %s: %s\n",
output_check_status(status),
output_repmgrd_status(status));
break;
case OM_CSV:
case OM_TEXT:
if (list_output != NULL)
{
check_status_list_set(list_output,
"repmgrd",
status,
output_repmgrd_status(status));
}
else
{
printf("%s (%s)\n",
output_check_status(status),
output_repmgrd_status(status));
}
default:
break;
}
return status;
}
/*
* This is not included in the general list output
*/
@@ -2879,8 +2820,7 @@ do_node_rejoin(void)
log_notice(_("temporarily removing \"standby.signal\""));
log_detail(_("this is required so pg_rewind can fix the unclean shutdown"));
make_standby_signal_path(config_file_options.data_directory,
standby_signal_file_path);
make_standby_signal_path(standby_signal_file_path);
if (unlink(standby_signal_file_path) < 0 && errno != ENOENT)
{
@@ -2905,7 +2845,7 @@ do_node_rejoin(void)
* of whether the pg_rewind operation failed.
*/
log_notice(_("recreating \"standby.signal\""));
write_standby_signal(config_file_options.data_directory);
write_standby_signal();
}
if (ret == false)
@@ -3629,25 +3569,6 @@ copy_file(const char *src_file, const char *dest_file)
}
static const char *
output_repmgrd_status(CheckStatus status)
{
switch (status)
{
case CHECK_STATUS_OK:
return "repmgrd running";
case CHECK_STATUS_WARNING:
return "repmgrd running but paused";
case CHECK_STATUS_CRITICAL:
return "repmgrd not running";
case CHECK_STATUS_UNKNOWN:
return "repmgrd status unknown";
}
return "UNKNOWN";
}
void
do_node_help(void)
{
@@ -3690,7 +3611,6 @@ do_node_help(void)
printf(_(" --role check node has expected role\n"));
printf(_(" --slots check for inactive replication slots\n"));
printf(_(" --missing-slots check for missing replication slots\n"));
printf(_(" --repmgrd check if repmgrd is running\n"));
printf(_(" --data-directory-config check repmgr's data directory configuration\n"));
puts("");

View File

@@ -20,7 +20,6 @@
*/
#include <sys/stat.h>
#include <time.h>
#include "repmgr.h"
#include "dirutil.h"
@@ -174,6 +173,21 @@ do_standby_clone(void)
initialize_conninfo_params(&recovery_conninfo, false);
/*
* --replication-conf-only provided - we'll handle that separately
*/
if (runtime_options.replication_conf_only == true)
{
return _do_create_replication_conf();
}
/*
* conninfo params for the actual upstream node (which might be different
* to the node we're cloning from) to write to recovery.conf
*/
mode = get_standby_clone_mode();
/*
* Copy the provided data directory; if a configuration file was provided,
* use the (mandatory) value from that; if -D/--pgdata was provided, use
@@ -201,19 +215,6 @@ do_standby_clone(void)
exit(ERR_BAD_CONFIG);
}
/*
* --replication-conf-only provided - we'll handle that separately
*/
if (runtime_options.replication_conf_only == true)
{
return _do_create_replication_conf();
}
/*
* conninfo params for the actual upstream node (which might be different
* to the node we're cloning from) to write to recovery.conf
*/
mode = get_standby_clone_mode();
if (mode == barman)
{
@@ -669,15 +670,6 @@ do_standby_clone(void)
log_hint(_("consider using the -c/--fast-checkpoint option"));
}
if (mode == pg_basebackup)
{
/*
* In --dry-run mode, this will just output the pg_basebackup command which
* would be executed.
*/
run_basebackup(&local_node_record);
}
PQfinish(source_conn);
log_info(_("all prerequisites for \"standby clone\" are met"));
@@ -1546,7 +1538,7 @@ _do_create_replication_conf(void)
}
else
{
if (write_standby_signal(local_data_directory) == false)
if (write_standby_signal() == false)
{
log_error(_("unable to write \"standby.signal\" file"));
}
@@ -2007,7 +1999,7 @@ do_standby_register(void)
/* only do this if record does not exist */
if (record_status != RECORD_FOUND)
{
log_warning(_("--upstream-node-id not supplied, assuming upstream node is primary (node ID: %i)"),
log_warning(_("--upstream-node-id not supplied, assuming upstream node is primary (node ID %i)"),
primary_node_id);
/* check our standby is connected */
@@ -3652,9 +3644,8 @@ do_standby_switchover(void)
PQExpBufferData remote_command_str;
PQExpBufferData command_output;
PQExpBufferData node_rejoin_options;
PQExpBufferData logmsg;
PQExpBufferData errmsg;
PQExpBufferData detailmsg;
PQExpBufferData event_details;
int r,
i;
@@ -3671,9 +3662,6 @@ do_standby_switchover(void)
/* store list of configuration files on the demotion candidate */
KeyValueList remote_config_files = {NULL, NULL};
/* temporary log file for "repmgr node rejoin" on the demotion candidate */
char node_rejoin_log[MAXPGPATH] = "";
NodeInfoList sibling_nodes = T_NODE_INFO_LIST_INITIALIZER;
SiblingNodeStats sibling_nodes_stats = T_SIBLING_NODES_STATS_INITIALIZER;
@@ -3816,24 +3804,24 @@ do_standby_switchover(void)
* the demotion candidate as the rejoin will fail if we are unable to to write to that.
*/
initPQExpBuffer(&logmsg);
initPQExpBuffer(&errmsg);
initPQExpBuffer(&detailmsg);
if (check_replication_config_owner(PQserverVersion(local_conn),
config_file_options.data_directory,
&logmsg, &detailmsg) == false)
&errmsg, &detailmsg) == false)
{
log_error("%s", logmsg.data);
log_error("%s", errmsg.data);
log_detail("%s", detailmsg.data);
termPQExpBuffer(&logmsg);
termPQExpBuffer(&errmsg);
termPQExpBuffer(&detailmsg);
PQfinish(local_conn);
exit(ERR_BAD_CONFIG);
}
termPQExpBuffer(&logmsg);
termPQExpBuffer(&errmsg);
termPQExpBuffer(&detailmsg);
/* check remote server connection and retrieve its record */
@@ -4781,7 +4769,6 @@ do_standby_switchover(void)
repmgrd_info = (RepmgrdInfo **) pg_malloc0(sizeof(RepmgrdInfo *) * all_nodes.node_count);
log_notice(_("attempting to pause repmgrd on %i nodes"), all_nodes.node_count);
for (cell = all_nodes.head; cell; cell = cell->next)
{
repmgrd_info[i] = pg_malloc0(sizeof(RepmgrdInfo));
@@ -4810,7 +4797,7 @@ do_standby_switchover(void)
unreachable_node_count++;
item_list_append_format(&repmgrd_connection_errors,
_("unable to connect to node \"%s\" (ID: %i):\n%s"),
_("unable to connect to node \"%s\" (ID %i):\n%s"),
cell->node_info->node_name,
cell->node_info->node_id,
PQerrorMessage(cell->node_info->conn));
@@ -4841,9 +4828,8 @@ do_standby_switchover(void)
initPQExpBuffer(&msg);
appendPQExpBuffer(&msg,
_("unable to connect to %i of %i node(s), unable to pause all repmgrd instances"),
unreachable_node_count,
all_nodes.node_count);
_("unable to connect to %i node(s), unable to pause all repmgrd instances"),
unreachable_node_count);
initPQExpBuffer(&detail);
@@ -4894,7 +4880,7 @@ do_standby_switchover(void)
*/
if (repmgrd_info[i]->pg_running == false)
{
log_warning(_("node \"%s\" (ID: %i) unreachable, unable to pause repmgrd"),
log_warning(_("node \"%s\" (ID %i) unreachable, unable to pause repmgrd"),
cell->node_info->node_name,
cell->node_info->node_id);
i++;
@@ -4907,7 +4893,7 @@ do_standby_switchover(void)
*/
if (repmgrd_info[i]->running == false)
{
log_notice(_("repmgrd not running on node \"%s\" (ID: %i), not pausing"),
log_warning(_("repmgrd not running on node \"%s\" (ID %i)"),
cell->node_info->node_name,
cell->node_info->node_id);
i++;
@@ -4928,14 +4914,14 @@ do_standby_switchover(void)
if (runtime_options.dry_run == true)
{
log_info(_("would pause repmgrd on node \"%s\" (ID: %i)"),
log_info(_("would pause repmgrd on node \"%s\" (ID %i)"),
cell->node_info->node_name,
cell->node_info->node_id);
}
else
{
/* XXX check result */
log_debug("pausing repmgrd on node \"%s\" (ID: %i)",
log_debug("pausing repmgrd on node \"%s\" (ID %i)",
cell->node_info->node_name,
cell->node_info->node_id);
@@ -5235,18 +5221,6 @@ do_standby_switchover(void)
format_lsn(replication_info.last_wal_receive_lsn),
format_lsn(remote_last_checkpoint_lsn));
/*
* optionally add a delay before promoting the standby; this is mainly
* useful for testing (e.g. for reappearance of the original primary) and
* is not documented.
*/
if (config_file_options.promote_delay > 0)
{
log_debug("sleeping %i seconds before promoting standby",
config_file_options.promote_delay);
sleep(config_file_options.promote_delay);
}
/*
* Promote standby (local node).
*
@@ -5372,21 +5346,6 @@ do_standby_switchover(void)
pfree(conninfo_normalized);
}
/* */
snprintf(node_rejoin_log, MAXPGPATH,
#if defined(__i386__) || defined(__i386)
"/tmp/node-rejoin.%u.log",
(unsigned)time(NULL)
#else
"/tmp/node-rejoin.%lu.log",
(unsigned long)time(NULL)
#endif
);
appendPQExpBuffer(&remote_command_str,
" > %s 2>&1 && echo \"1\" || echo \"0\"",
node_rejoin_log);
termPQExpBuffer(&node_rejoin_options);
log_debug("executing:\n %s", remote_command_str.data);
@@ -5400,161 +5359,78 @@ do_standby_switchover(void)
termPQExpBuffer(&remote_command_str);
initPQExpBuffer(&logmsg);
initPQExpBuffer(&detailmsg);
/* TODO: verify this node's record was updated correctly */
/* This is failure to execute the ssh command */
if (command_success == false)
{
log_error(_("rejoin failed with error code %i"), r);
switchover_success = false;
appendPQExpBuffer(&logmsg,
_("unable to execute \"repmgr node rejoin\" on demotion candidate \"%s\" (ID: %i)"),
remote_node_record.node_name,
remote_node_record.node_id);
appendPQExpBufferStr(&detailmsg,
command_output.data);
create_event_notification_extended(local_conn,
&config_file_options,
config_file_options.node_id,
"standby_switchover",
false,
command_output.data,
&event_info);
}
else
{
standy_join_status join_success = JOIN_UNKNOWN;
PQExpBufferData event_details;
standy_join_status join_success = check_standby_join(local_conn,
&local_node_record,
&remote_node_record);
/* "rempgr node rejoin" failed on the demotion candidate */
if (command_output.data[0] == '0')
initPQExpBuffer(&event_details);
switch (join_success) {
case JOIN_FAIL_NO_PING:
appendPQExpBuffer(&event_details,
_("node \"%s\" (ID: %i) promoted to primary, but demote node \"%s\" (ID: %i) did not beome available"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
switchover_success = false;
break;
case JOIN_FAIL_NO_REPLICATION:
appendPQExpBuffer(&event_details,
_("node \"%s\" (ID: %i) promoted to primary, but demote node \"%s\" (ID: %i) did not connect to the new primary"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
switchover_success = false;
break;
case JOIN_SUCCESS:
appendPQExpBuffer(&event_details,
_("node \"%s\" (ID: %i) promoted to primary, node \"%s\" (ID: %i) demoted to standby"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
}
create_event_notification_extended(local_conn,
&config_file_options,
config_file_options.node_id,
"standby_switchover",
switchover_success,
event_details.data,
&event_info);
if (switchover_success == true)
{
appendPQExpBuffer(&logmsg,
_("execution of \"repmgr node rejoin\" on demotion candidate \"%s\" (ID: %i) failed"),
remote_node_record.node_name,
remote_node_record.node_id);
/*
* Speculatively check if the demotion candidate has been restarted, e.g. by
* an external watchdog process which isn't aware a switchover is happening.
* This falls into the category "thing outside of our control which shouldn't
* happen, but if it does, make it easier to find out what happened".
*/
remote_conn = establish_db_connection(remote_node_record.conninfo, false);
if (PQstatus(remote_conn) == CONNECTION_OK)
{
if (get_recovery_type(remote_conn) == RECTYPE_PRIMARY)
{
appendPQExpBuffer(&detailmsg,
_("PostgreSQL instance on demotion candidate \"%s\" (ID: %i) is running as a primary\n"),
remote_node_record.node_name,
remote_node_record.node_id);
log_warning("%s", detailmsg.data);
}
}
PQfinish(remote_conn);
appendPQExpBuffer(&detailmsg,
"check log file \"%s\" on \"%s\" for details",
node_rejoin_log,
remote_node_record.node_name);
switchover_success = false;
join_success = JOIN_COMMAND_FAIL;
log_notice("%s", event_details.data);
}
else
{
join_success = check_standby_join(local_conn,
&local_node_record,
&remote_node_record);
switch (join_success) {
case JOIN_FAIL_NO_PING:
appendPQExpBuffer(&logmsg,
_("node \"%s\" (ID: %i) promoted to primary, but demotion candidate \"%s\" (ID: %i) did not become available"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
switchover_success = false;
break;
case JOIN_FAIL_NO_REPLICATION:
appendPQExpBuffer(&logmsg,
_("node \"%s\" (ID: %i) promoted to primary, but demotion candidate \"%s\" (ID: %i) did not connect to the new primary"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
switchover_success = false;
break;
case JOIN_SUCCESS:
appendPQExpBuffer(&logmsg,
_("node \"%s\" (ID: %i) promoted to primary, node \"%s\" (ID: %i) demoted to standby"),
config_file_options.node_name,
config_file_options.node_id,
remote_node_record.node_name,
remote_node_record.node_id);
break;
/* check_standby_join() does not return this */
case JOIN_COMMAND_FAIL:
break;
/* should never happen*/
case JOIN_UNKNOWN:
appendPQExpBuffer(&logmsg,
"unable to determine success of node rejoin action for demotion candidate \"%s\" (ID: %i)",
remote_node_record.node_name,
remote_node_record.node_id);
switchover_success = false;
break;
}
if (switchover_success == false)
{
appendPQExpBuffer(&detailmsg,
"check the PostgreSQL log file on demotion candidate \"%s\" (ID: %i)",
remote_node_record.node_name,
remote_node_record.node_id);
}
log_error("%s", event_details.data);
}
termPQExpBuffer(&event_details);
}
if (switchover_success == true)
{
/* TODO: verify demotion candidates's node record was updated correctly */
log_notice("%s", logmsg.data);
}
else
{
log_error("%s", logmsg.data);
}
initPQExpBuffer(&event_details);
appendPQExpBufferStr(&event_details, logmsg.data);
if (detailmsg.data[0] != '\0')
{
log_detail("%s", detailmsg.data);
appendPQExpBuffer(&event_details, "\n%s",
detailmsg.data);
}
create_event_notification_extended(local_conn,
&config_file_options,
config_file_options.node_id,
"standby_switchover",
switchover_success,
event_details.data,
&event_info);
termPQExpBuffer(&event_details);
termPQExpBuffer(&logmsg);
termPQExpBuffer(&detailmsg);
termPQExpBuffer(&command_output);
/*
* If --siblings-follow specified, attempt to make them follow the new
* primary
@@ -5667,7 +5543,7 @@ do_standby_switchover(void)
if (repmgrd_info[i]->paused == true && runtime_options.repmgrd_force_unpause == false)
{
log_debug("repmgrd on node \"%s\" (ID: %i) paused before switchover, --repmgrd-force-unpause not provided, not unpausing",
log_debug("repmgrd on node \"%s\" (ID %i) paused before switchover, --repmgrd-force-unpause not provided, not unpausing",
cell->node_info->node_name,
cell->node_info->node_id);
@@ -5675,7 +5551,7 @@ do_standby_switchover(void)
continue;
}
log_debug("unpausing repmgrd on node \"%s\" (ID: %i)",
log_debug("unpausing repmgrd on node \"%s\" (ID %i)",
cell->node_info->node_name,
cell->node_info->node_id);
@@ -5686,7 +5562,7 @@ do_standby_switchover(void)
if (repmgrd_pause(cell->node_info->conn, false) == false)
{
item_list_append_format(&repmgrd_unpause_errors,
_("unable to unpause node \"%s\" (ID: %i)"),
_("unable to unpause node \"%s\" (ID %i)"),
cell->node_info->node_name,
cell->node_info->node_id);
error_node_count++;
@@ -5695,7 +5571,7 @@ do_standby_switchover(void)
else
{
item_list_append_format(&repmgrd_unpause_errors,
_("unable to connect to node \"%s\" (ID: %i):\n%s"),
_("unable to connect to node \"%s\" (ID %i):\n%s"),
cell->node_info->node_name,
cell->node_info->node_id,
PQerrorMessage(cell->node_info->conn));
@@ -6979,13 +6855,6 @@ run_basebackup(t_node_info *node_record)
termPQExpBuffer(&params);
if (runtime_options.dry_run == true)
{
log_info(_("would execute:\n %s"), script.data);
termPQExpBuffer(&script);
return SUCCESS;
}
log_info(_("executing:\n %s"), script.data);
/*
@@ -8149,9 +8018,9 @@ create_recovery_file(t_node_info *node_record, t_conninfo_param_list *primary_co
free(escaped);
}
/*
* Caller requests the generated file to be written into a buffer
*/
if (as_file == false)
{
/* create file in buffer */
@@ -8171,17 +8040,20 @@ create_recovery_file(t_node_info *node_record, t_conninfo_param_list *primary_co
return true;
}
/*
* PostgreSQL 12 and later: modify postgresql.auto.conf
*
*/
if (server_version_num >= 120000)
{
if (modify_auto_conf(dest, &recovery_config) == false)
{
return false;
}
if (write_standby_signal(dest) == false)
if (write_standby_signal() == false)
{
return false;
}

View File

@@ -120,7 +120,6 @@ typedef struct
bool missing_slots;
bool has_passfile;
bool replication_connection;
bool repmgrd;
bool data_directory_config;
bool replication_config_owner;
bool db_connection;
@@ -176,7 +175,7 @@ typedef struct
/* "node status" options */ \
false, \
/* "node check" options */ \
false, false, false, false, false, false, false, false, false, false, false, false, false, \
false, false, false, false, false, false, false, false, false, false, false, false, \
/* "node rejoin" options */ \
"", \
/* "node service" options */ \
@@ -220,9 +219,7 @@ typedef enum
typedef enum
{
JOIN_UNKNOWN = -1,
JOIN_SUCCESS,
JOIN_COMMAND_FAIL,
JOIN_FAIL_NO_PING,
JOIN_FAIL_NO_REPLICATION
} standy_join_status;
@@ -285,8 +282,8 @@ extern void get_node_config_directory(char *config_dir_buf);
extern void get_node_data_directory(char *data_dir_buf);
extern void init_node_record(t_node_info *node_record);
extern bool can_use_pg_rewind(PGconn *conn, const char *data_directory, PQExpBufferData *reason);
extern void make_standby_signal_path(const char *data_dir, char *buf);
extern bool write_standby_signal(const char *data_dir);
extern void make_standby_signal_path(char *buf);
extern bool write_standby_signal(void);
extern bool create_replication_slot(PGconn *conn, char *slot_name, t_node_info *upstream_node_record, PQExpBufferData *error_msg);
extern bool drop_replication_slot_if_exists(PGconn *conn, int node_id, char *slot_name);

View File

@@ -549,10 +549,6 @@ main(int argc, char **argv)
runtime_options.data_directory_config = true;
break;
case OPT_REPMGRD:
runtime_options.repmgrd = true;
break;
case OPT_REPLICATION_CONFIG_OWNER:
runtime_options.replication_config_owner = true;
break;
@@ -3665,11 +3661,11 @@ can_use_pg_rewind(PGconn *conn, const char *data_directory, PQExpBufferData *rea
void
make_standby_signal_path(const char *data_dir, char *buf)
make_standby_signal_path(char *buf)
{
snprintf(buf, MAXPGPATH,
"%s/%s",
data_dir,
config_file_options.data_directory,
STANDBY_SIGNAL_FILE);
}
@@ -3677,15 +3673,13 @@ make_standby_signal_path(const char *data_dir, char *buf)
* create standby.signal (PostgreSQL 12 and later)
*/
bool
write_standby_signal(const char *data_dir)
write_standby_signal(void)
{
char standby_signal_file_path[MAXPGPATH] = "";
FILE *file;
mode_t um;
Assert(data_dir != NULL);
make_standby_signal_path(data_dir, standby_signal_file_path);
make_standby_signal_path(standby_signal_file_path);
/* Set umask to 0600 */
um = umask((~(S_IRUSR | S_IWUSR)) & (S_IRWXG | S_IRWXO));

View File

@@ -100,7 +100,6 @@
#define OPT_DB_CONNECTION 1047
#define OPT_VERIFY_BACKUP 1048
#define OPT_RECOVERY_MIN_APPLY_DELAY 1049
#define OPT_REPMGRD 1050
/* These options are for internal use only */
#define OPT_CONFIG_ARCHIVE_DIR 2001
@@ -194,7 +193,6 @@ static struct option long_options[] =
{"role", no_argument, NULL, OPT_ROLE},
{"slots", no_argument, NULL, OPT_SLOTS},
{"missing-slots", no_argument, NULL, OPT_MISSING_SLOTS},
{"repmgrd", no_argument, NULL, OPT_REPMGRD},
{"has-passfile", no_argument, NULL, OPT_HAS_PASSFILE},
{"replication-connection", no_argument, NULL, OPT_REPL_CONN},
{"data-directory-config", no_argument, NULL, OPT_DATA_DIRECTORY_CONFIG},

View File

@@ -88,24 +88,59 @@ void _PG_fini(void);
static void repmgr_shmem_startup(void);
PG_FUNCTION_INFO_V1(repmgr_set_local_node_id);
PG_FUNCTION_INFO_V1(repmgr_get_local_node_id);
PG_FUNCTION_INFO_V1(repmgr_standby_set_last_updated);
PG_FUNCTION_INFO_V1(repmgr_standby_get_last_updated);
PG_FUNCTION_INFO_V1(repmgr_set_upstream_last_seen);
PG_FUNCTION_INFO_V1(repmgr_get_upstream_last_seen);
PG_FUNCTION_INFO_V1(repmgr_get_upstream_node_id);
PG_FUNCTION_INFO_V1(repmgr_set_upstream_node_id);
PG_FUNCTION_INFO_V1(repmgr_notify_follow_primary);
PG_FUNCTION_INFO_V1(repmgr_get_new_primary);
PG_FUNCTION_INFO_V1(repmgr_reset_voting_status);
Datum set_local_node_id(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(set_local_node_id);
Datum get_local_node_id(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_local_node_id);
Datum standby_set_last_updated(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(standby_set_last_updated);
Datum standby_get_last_updated(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(standby_get_last_updated);
Datum set_upstream_last_seen(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(set_upstream_last_seen);
Datum get_upstream_last_seen(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_upstream_last_seen);
Datum get_upstream_node_id(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_upstream_node_id);
Datum set_upstream_node_id(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(set_upstream_node_id);
Datum notify_follow_primary(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(notify_follow_primary);
Datum get_new_primary(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_new_primary);
Datum reset_voting_status(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(reset_voting_status);
Datum set_repmgrd_pid(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(set_repmgrd_pid);
Datum get_repmgrd_pid(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_repmgrd_pid);
Datum get_repmgrd_pidfile(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_repmgrd_pidfile);
Datum repmgrd_is_running(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(repmgrd_is_running);
Datum repmgrd_pause(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(repmgrd_pause);
Datum repmgrd_is_paused(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(repmgrd_is_paused);
PG_FUNCTION_INFO_V1(repmgr_get_wal_receiver_pid);
Datum get_wal_receiver_pid(PG_FUNCTION_ARGS);
PG_FUNCTION_INFO_V1(get_wal_receiver_pid);
/*
@@ -198,7 +233,7 @@ repmgr_shmem_startup(void)
/* ==================== */
Datum
repmgr_set_local_node_id(PG_FUNCTION_ARGS)
set_local_node_id(PG_FUNCTION_ARGS)
{
int local_node_id = UNKNOWN_NODE_ID;
int stored_node_id = UNKNOWN_NODE_ID;
@@ -268,7 +303,7 @@ repmgr_set_local_node_id(PG_FUNCTION_ARGS)
Datum
repmgr_get_local_node_id(PG_FUNCTION_ARGS)
get_local_node_id(PG_FUNCTION_ARGS)
{
int local_node_id = UNKNOWN_NODE_ID;
@@ -285,7 +320,7 @@ repmgr_get_local_node_id(PG_FUNCTION_ARGS)
/* update and return last updated with current timestamp */
Datum
repmgr_standby_set_last_updated(PG_FUNCTION_ARGS)
standby_set_last_updated(PG_FUNCTION_ARGS)
{
TimestampTz last_updated = GetCurrentTimestamp();
@@ -302,7 +337,7 @@ repmgr_standby_set_last_updated(PG_FUNCTION_ARGS)
/* get last updated timestamp */
Datum
repmgr_standby_get_last_updated(PG_FUNCTION_ARGS)
standby_get_last_updated(PG_FUNCTION_ARGS)
{
TimestampTz last_updated;
@@ -319,7 +354,7 @@ repmgr_standby_get_last_updated(PG_FUNCTION_ARGS)
Datum
repmgr_set_upstream_last_seen(PG_FUNCTION_ARGS)
set_upstream_last_seen(PG_FUNCTION_ARGS)
{
int upstream_node_id = UNKNOWN_NODE_ID;
@@ -342,7 +377,7 @@ repmgr_set_upstream_last_seen(PG_FUNCTION_ARGS)
Datum
repmgr_get_upstream_last_seen(PG_FUNCTION_ARGS)
get_upstream_last_seen(PG_FUNCTION_ARGS)
{
long secs;
int microsecs;
@@ -376,7 +411,7 @@ repmgr_get_upstream_last_seen(PG_FUNCTION_ARGS)
Datum
repmgr_get_upstream_node_id(PG_FUNCTION_ARGS)
get_upstream_node_id(PG_FUNCTION_ARGS)
{
int upstream_node_id = UNKNOWN_NODE_ID;
@@ -391,7 +426,7 @@ repmgr_get_upstream_node_id(PG_FUNCTION_ARGS)
}
Datum
repmgr_set_upstream_node_id(PG_FUNCTION_ARGS)
set_upstream_node_id(PG_FUNCTION_ARGS)
{
int upstream_node_id = UNKNOWN_NODE_ID;
int local_node_id = UNKNOWN_NODE_ID;
@@ -427,7 +462,7 @@ repmgr_set_upstream_node_id(PG_FUNCTION_ARGS)
Datum
repmgr_notify_follow_primary(PG_FUNCTION_ARGS)
notify_follow_primary(PG_FUNCTION_ARGS)
{
int primary_node_id = UNKNOWN_NODE_ID;
@@ -470,7 +505,7 @@ repmgr_notify_follow_primary(PG_FUNCTION_ARGS)
Datum
repmgr_get_new_primary(PG_FUNCTION_ARGS)
get_new_primary(PG_FUNCTION_ARGS)
{
int new_primary_node_id = UNKNOWN_NODE_ID;
@@ -492,7 +527,7 @@ repmgr_get_new_primary(PG_FUNCTION_ARGS)
Datum
repmgr_reset_voting_status(PG_FUNCTION_ARGS)
reset_voting_status(PG_FUNCTION_ARGS)
{
if (!shared_state)
PG_RETURN_NULL();
@@ -700,7 +735,7 @@ repmgrd_is_paused(PG_FUNCTION_ARGS)
Datum
repmgr_get_wal_receiver_pid(PG_FUNCTION_ARGS)
get_wal_receiver_pid(PG_FUNCTION_ARGS)
{
int wal_receiver_pid;

View File

@@ -337,7 +337,6 @@ ssh_options='-q -o ConnectTimeout=10' # Options to append to "ssh"
# "--no-pid-file" will force PID file creation to be skipped.
# Note: there is normally no need to set this, particularly if
# repmgr was installed from packages.
#repmgrd_exit_on_inactive_node=false # If "true", and the node record is marked as "inactive", abort repmgrd startup
#standby_disconnect_on_failover=false # If "true", in a failover situation wait for all standbys to
# disconnect their WAL receivers before electing a new primary
# (PostgreSQL 9.5 and later only; repmgr user must be a superuser for this)

View File

@@ -1,7 +1,8 @@
# repmgr extension
comment = 'Replication manager for PostgreSQL'
default_version = '5.3'
default_version = '5.2'
module_pathname = '$libdir/repmgr'
relocatable = false
schema = repmgr

View File

@@ -135,7 +135,6 @@
#define DEFAULT_ASYNC_QUERY_TIMEOUT 60 /* seconds */
#define DEFAULT_PRIMARY_NOTIFICATION_TIMEOUT 60 /* seconds */
#define DEFAULT_REPMGRD_STANDBY_STARTUP_TIMEOUT -1 /*seconds */
#define DEFAULT_REPMGRD_EXIT_ON_INACTIVE_NODE false,
#define DEFAULT_STANDBY_DISCONNECT_ON_FAILOVER false
#define DEFAULT_SIBLING_NODES_DISCONNECT_TIMEOUT 30 /* seconds */
#define DEFAULT_CONNECTION_CHECK_TYPE CHECK_PING

View File

@@ -1,7 +1,5 @@
#define REPMGR_VERSION_DATE ""
#define REPMGR_VERSION "5.3"
#define REPMGR_VERSION_NUM 50300
#define REPMGR_EXTENSION_VERSION "5.3"
#define REPMGR_EXTENSION_NUM 50300
#define REPMGR_RELEASE_DATE "2021-10-12"
#define REPMGR_VERSION "5.2.1"
#define REPMGR_VERSION_NUM 50201
#define REPMGR_RELEASE_DATE "2020-12-07"
#define PG_ACTUAL_VERSION_NUM

View File

@@ -169,126 +169,45 @@ handle_sigint_physical(SIGNAL_ARGS)
/* perform some sanity checks on the node's configuration */
void
do_physical_node_check(PGconn *conn)
do_physical_node_check(void)
{
/*
* If node record is "inactive"; if not, attempt to set it to "active".
* Check if node record is active - if not, and `failover=automatic`, the
* node won't be considered as a promotion candidate; this often happens
* when a failed primary is recloned and the node was not re-registered,
* giving the impression failover capability is there when it's not. In
* this case abort with an error and a hint about registering.
*
* Usually it will have become inactive due to e.g. a standby being shut down
* while repmgrd was running in an unpaused state. In this case it's
* perfectly reasonable to automatically mark the node as "active".
* If `failover=manual`, repmgrd can continue to passively monitor the
* node, but we should nevertheless issue a warning and the same hint.
*/
if (local_node_info.active == false)
{
char *hint = "Check that \"repmgr (primary|standby) register\" was executed for this node";
RecoveryType recovery_type = get_recovery_type(conn);
/*
* If the local node's recovery status is incompatible with its registered
* status, e.g. registered as primary but running as a standby, refuse to start.
*
* This typically happens when a failed primary is recloned but the node was not
* re-registered, leaving the cluster in a potentially ambiguous state. In
* this case it would not be possible or desirable to attempt to set the
* node to active; the user should ensure the cluster is in the correct state.
*/
if (recovery_type != RECTYPE_UNKNOWN && local_node_info.type != UNKNOWN)
switch (config_file_options.failover)
{
bool require_reregister = false;
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
/* "failover" is an enum, all values should be covered here */
if (recovery_type == RECTYPE_STANDBY && local_node_info.type != STANDBY)
{
appendPQExpBuffer(&event_details,
_("node is registered as a %s but running as a standby"),
get_node_type_string(local_node_info.type));
require_reregister = true;
}
else if (recovery_type == RECTYPE_PRIMARY && local_node_info.type == STANDBY)
{
log_error(_("node is registered as a standby but running as a %s"), get_node_type_string(local_node_info.type));
require_reregister = true;
}
if (require_reregister == true)
{
log_error("%s", event_details.data);
case FAILOVER_AUTOMATIC:
log_error(_("this node is marked as inactive and cannot be used as a failover target"));
log_hint(_("%s"), hint);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_start",
"repmgrd_shutdown",
false,
event_details.data);
"node is inactive and cannot be used as a failover target");
termPQExpBuffer(&event_details);
terminate(ERR_BAD_CONFIG);
}
break;
termPQExpBuffer(&event_details);
}
/*
* Attempt to set node record active (unless explicitly configured not to)
*/
log_notice(_("setting node record for node \"%s\" (ID: %i) to \"active\""),
local_node_info.node_name,
local_node_info.node_id);
if (config_file_options.repmgrd_exit_on_inactive_node == false)
{
PGconn *primary_conn = get_primary_connection(conn, NULL, NULL);
bool success = true;
if (PQstatus(primary_conn) != CONNECTION_OK)
{
log_error(_("unable to connect to the primary node to activate the node record"));
success = false;
}
else
{
success = update_node_record_set_active(primary_conn, local_node_info.node_id, true);
PQfinish(primary_conn);
}
if (success == true)
{
local_node_info.active = true;
}
}
/*
* Corner-case where it was not possible to set the node to "active"
*/
if (local_node_info.active == false)
{
switch (config_file_options.failover)
{
/* "failover" is an enum, all values should be covered here */
case FAILOVER_AUTOMATIC:
log_error(_("this node is marked as inactive and cannot be used as a failover target"));
log_hint(_("%s"), hint);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_start",
false,
"node is inactive and cannot be used as a failover target");
terminate(ERR_BAD_CONFIG);
break;
case FAILOVER_MANUAL:
log_warning(_("this node is marked as inactive and will be passively monitored only"));
log_hint(_("%s"), hint);
break;
}
case FAILOVER_MANUAL:
log_warning(_("this node is marked as inactive and will be passively monitored only"));
log_hint(_("%s"), hint);
break;
}
}
@@ -1894,7 +1813,6 @@ monitor_streaming_standby(void)
int former_upstream_node_id = local_node_info.upstream_node_id;
NodeInfoList sibling_nodes = T_NODE_INFO_LIST_INITIALIZER;
PQExpBufferData event_details;
t_event_info event_info = T_EVENT_INFO_INITIALIZER;
update_node_record_set_primary(local_conn, local_node_info.node_id);
record_status = get_node_record(local_conn, local_node_info.node_id, &local_node_info);
@@ -1907,16 +1825,12 @@ monitor_streaming_standby(void)
initPQExpBuffer(&event_details);
appendPQExpBufferStr(&event_details,
_("promotion command failed but promotion completed successfully"));
event_info.node_id = former_upstream_node_id;
create_event_notification_extended(local_conn,
&config_file_options,
local_node_info.node_id,
"repmgrd_failover_promote",
true,
event_details.data,
&event_info);
create_event_notification(local_conn,
&config_file_options,
local_node_info.node_id,
"repmgrd_failover_promote",
true,
event_details.data);
termPQExpBuffer(&event_details);
@@ -3760,7 +3674,6 @@ promote_self(void)
{
PQExpBufferData event_details;
t_event_info event_info = T_EVENT_INFO_INITIALIZER;
/* update own internal node record */
record_status = get_node_record(local_conn, local_node_info.node_id, &local_node_info);
@@ -3777,16 +3690,13 @@ promote_self(void)
failed_primary.node_name,
failed_primary.node_id);
event_info.node_id = failed_primary.node_id;
/* local_conn is now the primary connection */
create_event_notification_extended(local_conn,
&config_file_options,
local_node_info.node_id,
"repmgrd_failover_promote",
true,
event_details.data,
&event_info);
create_event_notification(local_conn,
&config_file_options,
local_node_info.node_id,
"repmgrd_failover_promote",
true,
event_details.data);
termPQExpBuffer(&event_details);
}

View File

@@ -19,7 +19,7 @@
#ifndef _REPMGRD_PHYSICAL_H_
#define _REPMGRD_PHYSICAL_H_
void do_physical_node_check(PGconn *conn);
void do_physical_node_check(void);
void monitor_streaming_primary(void);
void monitor_streaming_standby(void);

View File

@@ -396,14 +396,13 @@ main(int argc, char **argv)
* extension is the latest available according to "pg_available_extensions" -
* - does our (major) version match that?
*/
log_verbose(LOG_DEBUG, "expected extension version: %i; extension version: %i",
REPMGR_EXTENSION_NUM, extversions.installed_version_num);
if ((REPMGR_EXTENSION_NUM/100) < (extversions.installed_version_num / 100))
log_verbose(LOG_DEBUG, "binary version: %i; extension version: %i",
REPMGR_VERSION_NUM, extversions.installed_version_num);
if ((REPMGR_VERSION_NUM/100) < (extversions.installed_version_num / 100))
{
log_error(_("this \"repmgr\" version is older than the installed \"repmgr\" extension version"));
log_detail(_("\"repmgr\" version %s providing extension version %s is installed but extension is version %s"),
log_detail(_("\"repmgr\" version %s is installed but extension is version %s"),
REPMGR_VERSION,
REPMGR_EXTENSION_VERSION,
extversions.installed_version);
log_hint(_("update the repmgr binaries to match the installed extension version"));
@@ -411,12 +410,11 @@ main(int argc, char **argv)
exit(ERR_BAD_CONFIG);
}
if ((REPMGR_EXTENSION_NUM/100) > (extversions.installed_version_num / 100))
if ((REPMGR_VERSION_NUM/100) > (extversions.installed_version_num / 100))
{
log_error(_("this \"repmgr\" version is newer than the installed \"repmgr\" extension version"));
log_detail(_("\"repmgr\" version %s providing extension version %s is installed but extension is version %s"),
log_detail(_("\"repmgr\" version %s is installed but extension is version %s"),
REPMGR_VERSION,
REPMGR_EXTENSION_VERSION,
extversions.installed_version);
log_hint(_("update the installed extension version by executing \"ALTER EXTENSION repmgr UPDATE\" in the repmgr database"));
@@ -512,7 +510,7 @@ main(int argc, char **argv)
log_debug("node id is %i, upstream node id is %i",
local_node_info.node_id,
local_node_info.upstream_node_id);
do_physical_node_check(local_conn);
do_physical_node_check();
}
if (daemonize == true)

View File

@@ -369,6 +369,7 @@ check_status_list_free(CheckStatusList *list)
}
const char *
output_check_status(CheckStatus status)
{