Fix alignment and syntax

This commit introduces three new options:
  - start_command
  - stop_command
  - restart_command

If these are set, repmgr will issue the specified command instead
of the default pg_ctl commands

HISTORY

@@ -1,4 +1,18 @@
-3.1.4   2016-07-
+3.1.5   2016-08-15
+        repmgrd: in a failover situation, prevent endless looping when
+          attempting to establish the status of a node with
+          `failover=manual` (Ian)
+        repmgrd: improve handling of failover events on standbys with
+          `failover=manual`, and create a new event notification
+          for this, `standby_disconnect_manual` (Ian)
+        repmgr: add further event notifications (Gianni)
+        repmgr: when executing `standby switchover`, don't collect remote
+          command output unless required (Gianni, Ian)
+        repmgrd: improve standby monitoring query (Ian, based on suggestion
+          from  Álvaro)
+        repmgr: various command line handling improvements (Ian)
+
+3.1.4   2016-07-12
         repmgr: new configuration option for setting "restore_command"
           in the recovery.conf file generated by repmgr (Martín)
         repmgr: add --csv option to "repmgr cluster show" (Gianni)

README.md

@@ -155,9 +155,15 @@ system.
 
 - RedHat/CentOS: RPM packages for `repmgr` are available via Yum through
   the PostgreSQL Global Development Group RPM repository ( http://yum.postgresql.org/ ).
-  You need to follow the instructions for your distribution (RedHat, CentOS,
+  Follow the instructions for your distribution (RedHat, CentOS,
   Fedora, etc.) and architecture as detailed at yum.postgresql.org.
 
+  2ndQuadrant also provides its own RPM packages which are made available
+  at the same time as each `repmgr` release, as it can take some days for
+  them to become available via the main PGDG repository. See here for details:
+
+     http://repmgr.org/yum-repository.html
+
 - Debian/Ubuntu: the most recent `repmgr` packages are available from the
   PostgreSQL Community APT repository ( http://apt.postgresql.org/ ).
   Instructions can be found in the APT section of the PostgreSQL Wiki
@@ -215,6 +221,34 @@ command line options:
 - `-b/--pg_bindir`
 
 
+### Command line options and environment variables
+
+For some commands, e.g. `repmgr standby clone`, database connection parameters
+need to be provided. Like other PostgreSQL utilities, following standard
+parameters can be used:
+
+- `-d/--dbname=DBNAME`
+- `-h/--host=HOSTNAME`
+- `-p/--port=PORT`
+- `-U/--username=USERNAME`
+
+If `-d/--dbname` contains an `=` sign or starts with a valid URI prefix (`postgresql://`
+or `postgres://`), it is treated as a conninfo string. See the PostgreSQL
+documentation for further details:
+
+  https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-CONNSTRING
+
+Note that if a `conninfo` string is provided, values set in this will override any
+provided as individual parameters. For example, with `-d 'host=foo' --host bar`, `foo`
+will be chosen over `bar`.
+
+Like other PostgreSQL utilities, `repmgr` will default to any values set in environment
+variables if explicit command line parameters are not provided. See the PostgreSQL
+documentation for further details:
+
+  https://www.postgresql.org/docs/current/static/libpq-envars.html
+
+
 Setting up a simple replication cluster with repmgr
 ---------------------------------------------------
 
@@ -383,14 +417,32 @@ Clone the standby with:
     [2016-01-07 17:21:28] [NOTICE] you can now start your PostgreSQL server
     [2016-01-07 17:21:28] [HINT] for example : pg_ctl -D /path/to/node2/data/ start
 
-This will clone the PostgreSQL data directory files from the master at repmgr_node1
+This will clone the PostgreSQL data directory files from the master at `repmgr_node1`
-using PostgreSQL's pg_basebackup utility. A `recovery.conf` file containing the
+using PostgreSQL's `pg_basebackup` utility. A `recovery.conf` file containing the
 correct parameters to start streaming from this master server will be created
-automatically, and unless otherwise the `postgresql.conf` and `pg_hba.conf`
+automatically, and unless otherwise specified, the `postgresql.conf` and `pg_hba.conf`
 files will be copied from the master.
 
-Make any adjustments to the PostgreSQL configuration files now, then start the
+Be aware that when initially cloning a standby, you will need to ensure
-standby server.
+that all required WAL files remain available while the cloning is taking
+place. To ensure this happens when using the default `pg_basebackup` method,
+`repmgr` will set `pg_basebackup`'s `--xlog-method` parameter to `stream`,
+which will ensure all WAL files generated during the cloning process are
+streamed in parallel with the main backup. Note that this requires two
+replication connections to be available.
+
+To override this behaviour, in `repmgr.conf` set `pg_basebackup`'s
+`--xlog-method` parameter to `fetch`:
+
+    pg_basebackup_options='--xlog-method=fetch'
+
+and ensure that `wal_keep_segments` is set to an appropriately high value.
+See the `pg_basebackup` documentation for details:
+
+    https://www.postgresql.org/docs/current/static/app-pgbasebackup.html
+
+Make any adjustments to the standby's PostgreSQL configuration files now,
+then start the server.
 
 * * *
 
@@ -470,7 +522,11 @@ so should be used with care.
 Further options can be passed to the `pg_basebackup` utility via
 the setting `pg_basebackup_options` in `repmgr.conf`. See the PostgreSQL
 documentation for more details of available options:
+<<<<<<< HEAD
+  http://www.postgresql.org/docs/current/static/app-pgbasebackup.html
+=======
   https://www.postgresql.org/docs/current/static/app-pgbasebackup.html
+>>>>>>> 72f9b0145afab1060dd1202c8f8937653c8b2e39
 
 ### Using rsync to clone a standby
 
@@ -488,7 +544,6 @@ and destination server as the contents of files existing on both servers need
 to be compared, meaning this method is not necessarily faster than making a
 fresh clone with `pg_basebackup`.
 
-
 ### Dealing with PostgreSQL configuration files
 
 By default, `repmgr` will attempt to copy the standard configuration files
@@ -503,6 +558,21 @@ which enables any valid `rsync` options to be passed to that command, e.g.:
 
     rsync_options='--exclude=postgresql.local.conf'
 
+### Controlling `primary_conninfo` in `recovery.conf`
+
+`repmgr` will create the `primary_conninfo` setting in `recovery.conf` based
+on the connection parameters provided to `repmgr standby clone` and PostgreSQL's
+standard connection defaults, including any environment variables set on the
+local node.
+
+To include specific connection parameters other than the standard host, port,
+username and database values (e.g. `sslmode`), include these in a `conninfo`-style
+tring passed to `repmgr` with `-d/--dbname` (see above for details), and/or set
+appropriate environment variables.
+
+Note that PostgreSQL will always set explicit defaults for `sslmode` and
+`sslcompression`.
+
 
 Setting up cascading replication with repmgr
 --------------------------------------------
@@ -576,6 +646,10 @@ To enable `repmgr` to use replication slots, set the boolean parameter
 Note that `repmgr` will fail with an error if this option is specified when
 working with PostgreSQL 9.3.
 
+Replication slots must be enabled in `postgresql.conf` by setting the parameter
+`max_replication_slots` to at least the number of expected standbys (changes
+to this parameter require a server restart).
+
 When cloning a standby, `repmgr` will automatically generate an appropriate
 slot name, which is stored in the `repl_nodes` table, and create the slot
 on the master:
@@ -598,18 +672,6 @@ Note that a slot name will be created by default for the master but not
 actually used unless the master is converted to a standby using e.g.
 `repmgr standby switchover`.
 
-Be aware that when initially cloning a standby, you will need to ensure
-that all required WAL files remain available while the cloning is taking
-place. If using the default `pg_basebackup` method, we recommend setting
-`pg_basebackup`'s `--xlog-method` parameter to `stream` like this:
-
-    pg_basebackup_options='--xlog-method=stream'
-
-See the `pg_basebackup` documentation for details:
-    https://www.postgresql.org/docs/current/static/app-pgbasebackup.html
-
-Otherwise it's necessary to set `wal_keep_segments` to an appropriately high
-value.
 
 Further information on replication slots in the PostgreSQL documentation:
     https://www.postgresql.org/docs/current/interactive/warm-standby.html#STREAMING-REPLICATION-SLOTS
@@ -907,7 +969,7 @@ actions happening, but we strongly recommend executing `repmgr` directly.
 
 `repmgrd` can be started simply with e.g.:
 
-    repmgrd -f /etc/repmgr.conf --verbose > $HOME/repmgr/repmgr.log 2>&1
+    repmgrd -f /etc/repmgr.conf --verbose >> $HOME/repmgr/repmgr.log 2>&1
 
 For permanent operation, we recommend using the options `-d/--daemonize` to
 detach the `repmgrd` process, and `-p/--pid-file` to write the process PID
@@ -929,7 +991,7 @@ table looks like this:
 
 
 Start `repmgrd` on each standby and verify that it's running by examining
-the log output, which at default log level will look like this:
+the log output, which at log level INFO will look like this:
 
     [2016-01-05 13:15:40] [INFO] checking cluster configuration with schema 'repmgr_test'
     [2016-01-05 13:15:40] [INFO] checking node 2 in cluster 'test'
@@ -1029,7 +1091,7 @@ the length of time it takes to determine that the connection is not possible.
 In particular explicitly setting a parameter for `connect_timeout` should
 be considered; the effective minimum value of `2` (seconds) will ensure
 that a connection failure at network level is reported as soon as possible,
-otherwise dependeing on the system settings (e.g. `tcp_syn_retries` in Linux)
+otherwise depending on the system settings (e.g. `tcp_syn_retries` in Linux)
 a delay of a minute or more is possible.
 
 For further details on `conninfo` network connection parameters, see:
@@ -1070,9 +1132,16 @@ table , it's advisable to regularly purge historical data with
 `repmgr cluster cleanup`; use the `-k/--keep-history` to specify how
 many day's worth of data should be retained.
 
+It's possible to use `repmgrd` to provide monitoring only for some or all
+nodes by setting `failover = manual` in the node's `repmgr.conf`. In the
+event of the node's upstream failing, no failover action will be taken
+and the node will require manual intervention to be reattached to replication.
+If this occurs, event notification `standby_disconnect_manual` will be
+created.
+
 Note that when a standby node is not streaming directly from its upstream
-node, i.e. recovering WAL from an archive, `apply_lag` will always
+node, e.g. recovering WAL from an archive, `apply_lag` will always appear as
-appear as `0 bytes`.
+`0 bytes`.
 
 
 Using a witness server with repmgrd
@@ -1169,6 +1238,7 @@ The following event types are available:
   * `standby_promote`
   * `standby_follow`
   * `standby_switchover`
+  * `standby_disconnect_manual`
   * `witness_create`
   * `witness_create`
   * `repmgrd_start`
@@ -1330,17 +1400,32 @@ which contains connection details for the local database.
     when analyzing connectivity from a particular node.
 
     This command requires a valid `repmgr.conf` file to be provided; no
-    additional arguments are required.
+    additional arguments are needed.
 
     Example:
 
         $ repmgr -f /etc/repmgr.conf cluster show
 
         Role      | Name  | Upstream | Connection String
-        ----------+-------|----------|--------------------------------------------
+        ----------+-------|----------|----------------------------------------
-        * master  | node1 |          | host=repmgr_node1 dbname=repmgr user=repmgr
+        * master  | node1 |          | host=db_node1 dbname=repmgr user=repmgr
-          standby | node2 | node1    | host=repmgr_node1 dbname=repmgr user=repmgr
+          standby | node2 | node1    | host=db_node2 dbname=repmgr user=repmgr
-          standby | node3 | node2    | host=repmgr_node1 dbname=repmgr user=repmgr
+          standby | node3 | node2    | host=db_node3 dbname=repmgr user=repmgr
+
+    To show database connection errors when polling nodes, run the command in
+    `--verbose` mode.
+
+    The `cluster show` command now accepts the optional parameter `--csv`, which
+    outputs the replication cluster's status in a simple CSV format, suitable for
+    parsing by scripts:
+
+        $ repmgr -f /etc/repmgr.conf cluster show --csv
+        1,-1
+        2,0
+        3,1
+
+    The first column is the node's ID, and the second column represents the
+    node's status (0 = master, 1 = standby, -1 = failed).
 
 * `cluster cleanup`
 
@@ -1359,20 +1444,22 @@ which contains connection details for the local database.
 `repmgr` or `repmgrd` will return one of the following error codes on program
 exit:
 
-* SUCCESS (0)              Program ran successfully.
+* SUCCESS (0)               Program ran successfully.
-* ERR_BAD_CONFIG (1)       Configuration file could not be parsed or was invalid
+* ERR_BAD_CONFIG (1)        Configuration file could not be parsed or was invalid
-* ERR_BAD_RSYNC (2)        An rsync call made by the program returned an error
+* ERR_BAD_RSYNC (2)         An rsync call made by the program returned an error (repmgr only)
-* ERR_NO_RESTART (4)       An attempt to restart a PostgreSQL instance failed
+* ERR_NO_RESTART (4)        An attempt to restart a PostgreSQL instance failed
-* ERR_DB_CON (6)           Error when trying to connect to a database
+* ERR_DB_CON (6)            Error when trying to connect to a database
-* ERR_DB_QUERY (7)         Error while executing a database query
+* ERR_DB_QUERY (7)          Error while executing a database query
-* ERR_PROMOTED (8)         Exiting program because the node has been promoted to master
+* ERR_PROMOTED (8)          Exiting program because the node has been promoted to master
-* ERR_BAD_PASSWORD (9)     Password used to connect to a database was rejected
+* ERR_STR_OVERFLOW (10)     String overflow error
-* ERR_STR_OVERFLOW (10)    String overflow error
+* ERR_FAILOVER_FAIL (11)    Error encountered during failover (repmgrd only)
-* ERR_FAILOVER_FAIL (11)   Error encountered during failover (repmgrd only)
+* ERR_BAD_SSH (12)          Error when connecting to remote host via SSH (repmgr only)
-* ERR_BAD_SSH (12)         Error when connecting to remote host via SSH
+* ERR_SYS_FAILURE (13)      Error when forking (repmgrd only)
-* ERR_SYS_FAILURE (13)     Error when forking (repmgrd only)
+* ERR_BAD_BASEBACKUP (14)   Error when executing pg_basebackup (repmgr only)
-* ERR_BAD_BASEBACKUP (14)  Error when executing pg_basebackup
+* ERR_MONITORING_FAIL (16)  Unrecoverable error encountered during monitoring (repmgrd only)
-* ERR_MONITORING_FAIL (16) Unrecoverable error encountered during monitoring (repmgrd only)
+* ERR_BAD_BACKUP_LABEL (17) Corrupt or unreadable backup label encountered (repmgr only)
+* ERR_SWITCHOVER_FAIL (18)  Error encountered during switchover (repmgr only)
+
 
 Support and Assistance
 ----------------------
@@ -1418,5 +1505,6 @@ Thanks from the repmgr core team.
 Further reading
 ---------------
 
+* http://blog.2ndquadrant.com/improvements-in-repmgr-3-1-4/
 * http://blog.2ndquadrant.com/managing-useful-clusters-repmgr/
 * http://blog.2ndquadrant.com/easier_postgresql_90_clusters/

config.c

@@ -219,6 +219,9 @@ parse_config(t_configuration_options *options)
 	memset(options->node_name, 0, sizeof(options->node_name));
 	memset(options->promote_command, 0, sizeof(options->promote_command));
 	memset(options->follow_command, 0, sizeof(options->follow_command));
+	memset(options->stop_command, 0, sizeof(options->stop_command));
+	memset(options->start_command, 0, sizeof(options->start_command));
+	memset(options->restart_command, 0, sizeof(options->restart_command));
 	memset(options->rsync_options, 0, sizeof(options->rsync_options));
 	memset(options->ssh_options, 0, sizeof(options->ssh_options));
 	memset(options->pg_bindir, 0, sizeof(options->pg_bindir));
@@ -341,6 +344,12 @@ parse_config(t_configuration_options *options)
 			strncpy(options->promote_command, value, MAXLEN);
 		else if (strcmp(name, "follow_command") == 0)
 			strncpy(options->follow_command, value, MAXLEN);
+		else if (strcmp(name, "stop_command") == 0)
+			strncpy(options->stop_command, value, MAXLEN);
+		else if (strcmp(name, "start_command") == 0)
+			strncpy(options->start_command, value, MAXLEN);
+		else if (strcmp(name, "restart_command") == 0)
+			strncpy(options->restart_command, value, MAXLEN);
 		else if (strcmp(name, "master_response_timeout") == 0)
 			options->master_response_timeout = repmgr_atoi(value, "master_response_timeout", &config_errors, false);
 		/*

config.h

@@ -62,6 +62,9 @@ typedef struct
 	char		node_name[MAXLEN];
 	char		promote_command[MAXLEN];
 	char		follow_command[MAXLEN];
+	char		stop_command[MAXLEN];
+	char		start_command[MAXLEN];
+	char		restart_command[MAXLEN];
 	char		loglevel[MAXLEN];
 	char		logfacility[MAXLEN];
 	char		rsync_options[QUERY_STR_LEN];
@@ -87,7 +90,7 @@ typedef struct
  * The following will initialize the structure with a minimal set of options;
  * actual defaults are set in parse_config() before parsing the configuration file
  */
-#define T_CONFIGURATION_OPTIONS_INITIALIZER { "", -1, NO_UPSTREAM_NODE, "", MANUAL_FAILOVER, -1, "", "", "", "", "", "", "", -1, -1, -1, "", "", "", "", "", 0, 0, 0, 0, "", { NULL, NULL }, { NULL, NULL } }
+#define T_CONFIGURATION_OPTIONS_INITIALIZER { "", -1, NO_UPSTREAM_NODE, "", MANUAL_FAILOVER, -1, "", "", "", "", "", "", "", "", "", "", -1, -1, -1, "", "", "", "", "", 0, 0, 0, 0, "", { NULL, NULL }, {NULL, NULL} }
 
 typedef struct ErrorListCell
 {

dbutils.c

@@ -34,7 +34,7 @@ char repmgr_schema_quoted[MAXLEN] = "";
 static int _get_node_record(PGconn *conn, char *cluster, char *sqlquery, t_node_info *node_info);
 
 PGconn *
-_establish_db_connection(const char *conninfo, const bool exit_on_error, const bool log_notice)
+_establish_db_connection(const char *conninfo, const bool exit_on_error, const bool log_notice, const bool verbose_only)
 {
 	/* Make a connection to the database */
 	PGconn	   *conn = NULL;
@@ -50,15 +50,23 @@ _establish_db_connection(const char *conninfo, const bool exit_on_error, const b
 	/* Check to see that the backend connection was successfully made */
 	if ((PQstatus(conn) != CONNECTION_OK))
 	{
-		if (log_notice)
+		bool emit_log = true;
+
+		if (verbose_only == true && verbose_logging == false)
+			emit_log = false;
+
+		if (emit_log)
 		{
-			log_notice(_("connection to database failed: %s\n"),
+			if (log_notice)
-					PQerrorMessage(conn));
+			{
-		}
+				log_notice(_("connection to database failed: %s\n"),
-		else
+						   PQerrorMessage(conn));
-		{
+			}
-			log_err(_("connection to database failed: %s\n"),
+			else
-					PQerrorMessage(conn));
+			{
+				log_err(_("connection to database failed: %s\n"),
+						PQerrorMessage(conn));
+			}
 		}
 
 		if (exit_on_error)
@@ -71,16 +79,35 @@ _establish_db_connection(const char *conninfo, const bool exit_on_error, const b
 	return conn;
 }
 
+
+/*
+ * Establish a database connection, optionally exit on error
+ */
 PGconn *
 establish_db_connection(const char *conninfo, const bool exit_on_error)
 {
-	return _establish_db_connection(conninfo, exit_on_error, false);
+	return _establish_db_connection(conninfo, exit_on_error, false, false);
 }
 
+/*
+ * Attempt to establish a database connection, never exit on error, only
+ * output error messages if --verbose option used
+ */
 PGconn *
-test_db_connection(const char *conninfo, const bool exit_on_error)
+establish_db_connection_quiet(const char *conninfo)
 {
-	return _establish_db_connection(conninfo, exit_on_error, true);
+	return _establish_db_connection(conninfo, false, false, true);
+}
+
+/*
+ * Attempt to establish a database connection, never exit on error,
+ * output connection error messages as NOTICE (useful when connection
+ * failure is expected)
+ */
+PGconn *
+test_db_connection(const char *conninfo)
+{
+	return _establish_db_connection(conninfo, false, true, false);
 }
 
 

dbutils.h

@@ -81,11 +81,12 @@ typedef struct s_replication_slot
 
 PGconn *_establish_db_connection(const char *conninfo,
 								 const bool exit_on_error,
-								 const bool log_notice);
+								 const bool log_notice,
+								 const bool verbose_only);
 PGconn *establish_db_connection(const char *conninfo,
 								const bool exit_on_error);
-PGconn *test_db_connection(const char *conninfo,
+PGconn *establish_db_connection_quiet(const char *conninfo);
-						   const bool exit_on_error);
+PGconn *test_db_connection(const char *conninfo);
 PGconn *establish_db_connection_by_params(const char *keywords[],
 								  const char *values[],
 								  const bool exit_on_error);

errcode.h

@@ -29,7 +29,6 @@
 #define ERR_DB_CON 6
 #define ERR_DB_QUERY 7
 #define ERR_PROMOTED 8
-#define ERR_BAD_PASSWORD 9
 #define ERR_STR_OVERFLOW 10
 #define ERR_FAILOVER_FAIL 11
 #define ERR_BAD_SSH 12

log.c

@@ -142,7 +142,7 @@ log_verbose(int level, const char *fmt, ...)
 
 
 bool
-logger_init(t_configuration_options * opts, const char *ident)
+logger_init(t_configuration_options *opts, const char *ident)
 {
 	char	   *level = opts->loglevel;
 	char	   *facility = opts->logfacility;

log.h

@@ -130,5 +130,7 @@ __attribute__((format(PG_PRINTF_ATTRIBUTE, 2, 3)));
 
 extern int	log_type;
 extern int	log_level;
+extern int	verbose_logging;
+extern int	terse_logging;
 
-#endif
+#endif /* _REPMGR_LOG_H_ */

repmgr.conf.sample

@@ -101,6 +101,29 @@
 # (if not provided, defaults to system $PATH)
 #pg_bindir=/usr/bin/
 
+# service control commands
+#
+# repmgr provides options to to override the default pg_ctl commands
+# used to stop, start  and restart the PostgreSQL cluster
+#
+# NOTE: These commands must be runnable on remote nodes as well for switchover
+# to function correctly.
+#
+# If you use sudo, the user repmgr runs as (usually 'postgres')  must have
+# passwordless sudo access to execute the command
+#
+# For example, to use systemd, you may use the following configuration:
+#
+#    # this is required when running sudo over ssh without -t:
+#    Defaults:postgres !requiretty
+#    postgres ALL = NOPASSWD: /usr/bin/systemctl stop postgresql-9.5, \
+#       /usr/bin/systemctl start postgresql-9.5, \
+#       /usr/bin/systemctl restart postgresql-9.5
+#
+# start_command = systemctl start postgresql-9.5
+# stop_command = systemctl stop postgresql-9.5
+# restart_command = systemctl restart postgresql-9.5
+
 # external command options
 
 #rsync_options=--archive --checksum --compress --progress --rsh="ssh -o \"StrictHostKeyChecking no\""
@@ -144,10 +167,18 @@
 #reconnect_interval=10
 
 # Autofailover options
-#failover=manual    # one of 'automatic', 'manual'
+#failover=manual    # one of 'automatic', 'manual' (default: manual)
-                    # (default: manual)
+                    # defines the action to take in the event of upstream failure
-#priority=100       # a value of zero or less prevents the node being promoted to primary
+                    #
+                    # 'automatic': repmgrd will automatically attempt to promote the
+                    #    node or follow the new upstream node
+                    # 'manual': repmgrd will take no action and the mode will require
+                    #    manual attention to reattach it to replication
+
+#priority=100       # indicate a preferred priorty for promoting nodes
+                    # a value of zero or less prevents the node being promoted to primary
                     # (default: 100)
+
 #promote_command='repmgr standby promote -f /path/to/repmgr.conf'
 #follow_command='repmgr standby follow -f /path/to/repmgr.conf -W'
 

repmgr.h

@@ -47,6 +47,15 @@
 #define NO_UPSTREAM_NODE	-1
 #define UNKNOWN_NODE_ID     -1
 
+#define OPT_HELP                         1
+#define OPT_CHECK_UPSTREAM_CONFIG        2
+#define OPT_RECOVERY_MIN_APPLY_DELAY     3
+#define OPT_IGNORE_EXTERNAL_CONFIG_FILES 4
+#define OPT_CONFIG_ARCHIVE_DIR           5
+#define OPT_PG_REWIND                    6
+#define OPT_PWPROMPT                     7
+#define OPT_CSV                          8
+#define OPT_INITDB_NO_PWPROMPT           9
 
 
 /* Run time options type */
@@ -92,11 +101,10 @@ typedef struct
 	char		recovery_min_apply_delay[MAXLEN];
 
 	/* deprecated command line options */
-	char		localport[MAXLEN];
+	char            localport[MAXLEN];
-	bool		initdb_no_pwprompt;
 }	t_runtime_options;
 
-#define T_RUNTIME_OPTIONS_INITIALIZER { "", "", "", "", "", "", "", DEFAULT_WAL_KEEP_SEGMENTS, false, false, false, false, false, false, false, false, false, false, "", "", "", "", "fast", "", 0, "", "", "", false }
+#define T_RUNTIME_OPTIONS_INITIALIZER { "", "", "", "", "", "", "", DEFAULT_WAL_KEEP_SEGMENTS, false, false, false, false, false, false, false, false, false, false, "", "", "", "", "fast", "", 0, "", "", ""}
 
 struct BackupLabel
 {

repmgrd.c

@@ -41,7 +41,10 @@
 #include "access/xlogdefs.h"
 #include "pqexpbuffer.h"
 
+/* Message strings passed in repmgrSharedState->location */
 
+#define PASSIVE_NODE "PASSIVE_NODE"
+#define LSN_QUERY_ERROR "LSN_QUERY_ERROR"
 
 /* Local info */
 t_configuration_options local_options = T_CONFIGURATION_OPTIONS_INITIALIZER;
@@ -59,6 +62,13 @@ t_node_info node_info;
 
 bool		failover_done = false;
 
+/*
+ * when `failover=manual`, and the upstream server has gone away,
+ * this flag is set to indicate we should connect to whatever the
+ * current master is to update monitoring information
+ */
+bool		manual_mode_upstream_disconnected = false;
+
 char	   *pid_file = NULL;
 
 static void help(void);
@@ -124,7 +134,7 @@ main(int argc, char **argv)
 		{"monitoring-history", no_argument, NULL, 'm'},
 		{"daemonize", no_argument, NULL, 'd'},
 		{"pid-file", required_argument, NULL, 'p'},
-		{"help", no_argument, NULL, '?'},
+		{"help", no_argument, NULL, OPT_HELP},
 		{"version", no_argument, NULL, 'V'},
 		{NULL, 0, NULL, 0}
 	};
@@ -158,6 +168,23 @@ main(int argc, char **argv)
 	{
 		switch (c)
 		{
+			case '?':
+				/* Actual help option given */
+				if (strcmp(argv[optind - 1], "-?") == 0)
+				{
+					help();
+					exit(SUCCESS);
+				}
+				/* unknown option reported by getopt */
+				else
+					goto unknown_option;
+				break;
+			case OPT_HELP:
+				help();
+				exit(SUCCESS);
+			case 'V':
+				printf("%s %s (PostgreSQL %s)\n", progname(), REPMGR_VERSION, PG_VERSION);
+				exit(SUCCESS);
 			case 'f':
 				config_file = optarg;
 				break;
@@ -173,13 +200,9 @@ main(int argc, char **argv)
 			case 'p':
 				pid_file = optarg;
 				break;
-			case '?':
+
-				help();
-				exit(SUCCESS);
-			case 'V':
-				printf("%s %s (PostgreSQL %s)\n", progname(), REPMGR_VERSION, PG_VERSION);
-				exit(SUCCESS);
 			default:
+		unknown_option:
 				usage();
 				exit(ERR_BAD_CONFIG);
 		}
@@ -433,6 +456,7 @@ main(int argc, char **argv)
 					my_local_conn = establish_db_connection(local_options.conninfo, true);
 					update_registration();
 				}
+
 				/* Log startup event */
 				if (startup_event_logged == false)
 				{
@@ -639,7 +663,7 @@ witness_monitor(void)
 								 local_options.master_response_timeout) != 1)
 		return;
 
-	/* Get local xlog info */
+	/* Get timestamp for monitoring update */
 	sqlquery_snprintf(sqlquery, "SELECT CURRENT_TIMESTAMP");
 
 	res = PQexec(my_local_conn, sqlquery);
@@ -720,6 +744,8 @@ standby_monitor(void)
 	const char *upstream_node_type = NULL;
 
 	bool		receiving_streamed_wal = true;
+
+
 	/*
 	 * Verify that the local node is still available - if not there's
 	 * no point in doing much else anyway
@@ -741,15 +767,32 @@ standby_monitor(void)
 		goto continue_monitoring_standby;
 	}
 
-	upstream_conn = get_upstream_connection(my_local_conn,
+	/*
-											local_options.cluster_name,
+	 * Standby has `failover` set to manual and is disconnected from
-											local_options.node,
+	 * replication following a prior upstream node failure - we'll
-											&upstream_node_id,
+	 * find the master to be able to write monitoring information, if
-											upstream_conninfo);
+	 * required
+	 */
+	if (manual_mode_upstream_disconnected == true)
+	{
+		upstream_conn = get_master_connection(my_local_conn,
+												local_options.cluster_name,
+												&upstream_node_id,
+												upstream_conninfo);
+		upstream_node_type = "master";
+	}
+	else
+	{
+		upstream_conn = get_upstream_connection(my_local_conn,
+												local_options.cluster_name,
+												local_options.node,
+												&upstream_node_id,
+												upstream_conninfo);
 
-	upstream_node_type = (upstream_node_id == master_options.node)
+		upstream_node_type = (upstream_node_id == master_options.node)
-		? "master"
+			? "master"
-		: "upstream";
+			: "upstream";
+	}
 
 	/*
 	 * Check that the upstream node is still available
@@ -764,29 +807,52 @@ standby_monitor(void)
 
 	if (PQstatus(upstream_conn) != CONNECTION_OK)
 	{
+		int previous_master_node_id = master_options.node;
+
 		PQfinish(upstream_conn);
 		upstream_conn = NULL;
 
+		/*
+		 * When `failover=manual`, no actual failover will be performed, instead
+		 * the following happens:
+		 *  - find the new master
+		 *  - create an event notification `standby_disconnect_manual`
+		 *  - set a flag to indicate we're disconnected from replication,
+		 */
 		if (local_options.failover == MANUAL_FAILOVER)
 		{
 			log_err(_("Unable to reconnect to %s. Now checking if another node has been promoted.\n"), upstream_node_type);
 
+			/*
+			 * Set the location string in shared memory to indicate to other
+			 * repmgrd instances that we're *not* a promotion candidate and
+			 * that other repmgrd instance should not expect location updates
+			 * from us
+			 */
+
+			update_shared_memory(PASSIVE_NODE);
+
 			for (connection_retries = 0; connection_retries < local_options.reconnect_attempts; connection_retries++)
 			{
 				master_conn = get_master_connection(my_local_conn,
 					local_options.cluster_name, &master_options.node, NULL);
+
 				if (PQstatus(master_conn) == CONNECTION_OK)
 				{
 					/*
 					 * Connected, we can continue the process so break the
 					 * loop
 					 */
-					log_err(_("connected to node %d, continuing monitoring.\n"),
+					log_notice(_("connected to node %d, continuing monitoring.\n"),
 							master_options.node);
 					break;
 				}
 				else
 				{
+					/*
+					 * XXX this is the only place where `retry_promote_interval_secs`
+					 * is used - this parameter should be renamed or possibly be replaced
+					 */
 					log_err(
 					    _("no new master found, waiting %i seconds before retry...\n"),
 					    local_options.retry_promote_interval_secs
@@ -816,6 +882,36 @@ standby_monitor(void)
 
 				terminate(ERR_DB_CON);
 			}
+
+			/*
+			 * connected to a master - is it the same as the former upstream?
+			 * if not:
+			 *  - create event standby_disconnect
+			 *  - set global "disconnected_manual_standby"
+			 */
+
+			if (previous_master_node_id != master_options.node)
+			{
+				PQExpBufferData errmsg;
+				initPQExpBuffer(&errmsg);
+
+				appendPQExpBuffer(&errmsg,
+								  _("node %i is in manual failover mode and is now disconnected from replication"),
+								  local_options.node);
+
+				log_verbose(LOG_DEBUG, "old master: %i; current: %i\n", previous_master_node_id, master_options.node);
+
+				manual_mode_upstream_disconnected = true;
+
+				create_event_record(master_conn,
+									&local_options,
+									local_options.node,
+									"standby_disconnect_manual",
+									/* here "true" indicates the action has occurred as expected */
+									true,
+									errmsg.data);
+
+			}
 		}
 		else if (local_options.failover == AUTOMATIC_FAILOVER)
 		{
@@ -916,8 +1012,8 @@ standby_monitor(void)
 		 * the stream. If we set the local standby node as failed and it's now running
 		 * and receiving replication data, we should activate it again.
 		 */
-	        set_local_node_status();
+		set_local_node_status();
-	        log_info(_("standby connection recovered!\n"));
+		log_info(_("standby connection recovered!\n"));
 	}
 
 	/* Fast path for the case where no history is requested */
@@ -929,6 +1025,7 @@ standby_monitor(void)
 	 * from the upstream node to write monitoring information
 	 */
 
+	/* XXX not used? */
 	upstream_node = get_node_info(my_local_conn, local_options.cluster_name, upstream_node_id);
 
 	sprintf(sqlquery,
@@ -983,12 +1080,19 @@ standby_monitor(void)
 		return;
 
 	/* Get local xlog info */
+
 	sqlquery_snprintf(sqlquery,
-					  "SELECT CURRENT_TIMESTAMP, "
+					  " SELECT ts, "
-					  "pg_catalog.pg_last_xlog_receive_location(), "
+					  "        receive_location, "
-					  "pg_catalog.pg_last_xlog_replay_location(), "
+					  "        replay_location, "
-					  "pg_catalog.pg_last_xact_replay_timestamp(), "
+					  "        replay_timestamp, "
-					  "pg_catalog.pg_last_xlog_receive_location() >= pg_catalog.pg_last_xlog_replay_location()");
+					  "        receive_location >= replay_location "
+					  "   FROM (SELECT CURRENT_TIMESTAMP AS ts, "
+					  "         pg_catalog.pg_last_xlog_receive_location() AS receive_location, "
+					  "         pg_catalog.pg_last_xlog_replay_location()  AS replay_location, "
+					  "         pg_catalog.pg_last_xact_replay_timestamp() AS replay_timestamp "
+					  "        ) q ");
+
 
 	res = PQexec(my_local_conn, sqlquery);
 	if (PQresultStatus(res) != PGRES_TUPLES_OK)
@@ -1073,10 +1177,12 @@ standby_monitor(void)
 	}
 	else
 	{
-		apply_lag = (long long unsigned int)lsn_last_xlog_receive_location - lsn_last_xlog_replay_location;
 		lsn_last_xlog_receive_location = lsn_to_xlogrecptr(last_xlog_receive_location, NULL);
+
+		apply_lag = (long long unsigned int)lsn_last_xlog_receive_location - lsn_last_xlog_replay_location;
 	}
 
+
 	/* Calculate replication lag */
 	if (lsn_master_current_xlog_location >= lsn_last_xlog_receive_location)
 	{
@@ -1121,7 +1227,6 @@ standby_monitor(void)
 					  last_xlog_receive_location,
 					  replication_lag,
 					  apply_lag);
-
 	/*
 	 * Execute the query asynchronously, but don't check for a result. We will
 	 * check the result next time we pause for a monitor step.
@@ -1158,8 +1263,6 @@ do_master_failover(void)
 	XLogRecPtr	xlog_recptr;
 	bool		lsn_format_ok;
 
-	char		last_xlog_replay_location[MAXLEN];
-
 	PGconn	   *node_conn = NULL;
 
 	/*
@@ -1340,8 +1443,8 @@ do_master_failover(void)
 				  " considered as new master and exit.\n"),
 				PQerrorMessage(my_local_conn));
 		PQclear(res);
-		sprintf(last_xlog_replay_location, "'%X/%X'", 0, 0);
+
-		update_shared_memory(last_xlog_replay_location);
+		update_shared_memory(LSN_QUERY_ERROR);
 		terminate(ERR_DB_QUERY);
 	}
 	/* write last location in shared memory */
@@ -1391,6 +1494,7 @@ do_master_failover(void)
 
 		while (!nodes[i].is_ready)
 		{
+			char location_value[MAXLEN];
 
 			sqlquery_snprintf(sqlquery,
 							  "SELECT %s.repmgr_get_last_standby_location()",
@@ -1406,7 +1510,11 @@ do_master_failover(void)
 				terminate(ERR_DB_QUERY);
 			}
 
-			xlog_recptr = lsn_to_xlogrecptr(PQgetvalue(res, 0, 0), &lsn_format_ok);
+			/* Copy the returned value as we'll need to reference it a few times */
+			strncpy(location_value, PQgetvalue(res, 0, 0), MAXLEN);
+			PQclear(res);
+
+			xlog_recptr = lsn_to_xlogrecptr(location_value, &lsn_format_ok);
 
 			/* If position reported as "invalid", check for format error or
 			 * empty string; otherwise position is 0/0 and we need to continue
@@ -1414,10 +1522,36 @@ do_master_failover(void)
 			 */
 			if (xlog_recptr == InvalidXLogRecPtr)
 			{
+				bool continue_loop = true;
+
 				if (lsn_format_ok == false)
 				{
+
+					/*
+					 * The node is indicating it is not a promotion candidate -
+					 * in this case we can store its invalid LSN to ensure it
+					 * can't be a promotion candidate when comparing locations
+					 */
+					if (strcmp(location_value, PASSIVE_NODE) == 0)
+					{
+						log_debug("node %i is passive mode\n", nodes[i].node_id);
+						log_info(_("node %i will not be considered for promotion\n"), nodes[i].node_id);
+						nodes[i].xlog_location = InvalidXLogRecPtr;
+						continue_loop = false;
+					}
+					/*
+					 * This should probably never happen but if it does, rule the
+					 * node out as a promotion candidate
+					 */
+					else if (strcmp(location_value, LSN_QUERY_ERROR) == 0)
+					{
+						log_warning(_("node %i is unable to update its shared memory and will not be considered for promotion\n"), nodes[i].node_id);
+						nodes[i].xlog_location = InvalidXLogRecPtr;
+						continue_loop = false;
+					}
+
 					/* Unable to parse value returned by `repmgr_get_last_standby_location()` */
-					if (*PQgetvalue(res, 0, 0) == '\0')
+					else if (*location_value == '\0')
 					{
 						log_crit(
 							_("unable to obtain LSN from node %i"), nodes[i].node_id
@@ -1426,8 +1560,8 @@ do_master_failover(void)
 							_("please check that 'shared_preload_libraries=repmgr_funcs' is set in postgresql.conf\n")
 							);
 
-						PQclear(res);
 						PQfinish(node_conn);
+						/* XXX shouldn't we just ignore this node? */
 						exit(ERR_BAD_CONFIG);
 					}
 
@@ -1435,25 +1569,29 @@ do_master_failover(void)
 					 * Very unlikely to happen; in the absence of any better
 					 * strategy keep checking
 					 */
-					log_warning(_("unable to parse LSN \"%s\"\n"),
+					else {
-								PQgetvalue(res, 0, 0));
+						log_warning(_("unable to parse LSN \"%s\"\n"),
+									location_value);
+					}
 				}
 				else
 				{
 					log_debug(
 						_("invalid LSN returned from node %i: '%s'\n"),
 						nodes[i].node_id,
-						PQgetvalue(res, 0, 0)
+						location_value);
-						);
 				}
 
-				PQclear(res);
+				/*
-
+				 * If the node is still reporting an InvalidXLogRecPtr, it means
-				/* If position is 0/0, keep checking */
+				 * its repmgrd hasn't yet had time to update it (either with a valid
-				/* XXX we should add a timeout here to prevent infinite looping
+				 * XLogRecPtr or a message) so we continue looping.
+				 *
+				 * XXX we should add a timeout here to prevent infinite looping
 				 * if the other node's repmgrd is not up
 				 */
-				continue;
+				if (continue_loop == true)
+					continue;
 			}
 
 			if (nodes[i].xlog_location < xlog_recptr)
@@ -1461,8 +1599,7 @@ do_master_failover(void)
 				nodes[i].xlog_location = xlog_recptr;
 			}
 
-			log_debug(_("LSN of node %i is: %s\n"), nodes[i].node_id, PQgetvalue(res, 0, 0));
+			log_debug(_("LSN of node %i is: %s\n"), nodes[i].node_id, location_value);
-			PQclear(res);
 
 			ready_nodes++;
 			nodes[i].is_ready = true;
@@ -2138,7 +2275,7 @@ lsn_to_xlogrecptr(char *lsn, bool *format_ok)
 	{
 		if (format_ok != NULL)
 			*format_ok = false;
-		log_err(_("incorrect log location format: %s\n"), lsn);
+		log_warning(_("incorrect log location format: %s\n"), lsn);
 		return 0;
 	}
 

version.h

@@ -1,6 +1,6 @@
 #ifndef _VERSION_H_
 #define _VERSION_H_
 
-#define REPMGR_VERSION "3.2dev"
+#define REPMGR_VERSION "3.1.5"
 
 #endif