Update HISTORY

Fixes GitHub #280

CONTRIBUTING.md

@@ -2,7 +2,7 @@ License and Contributions
 =========================
 
 `repmgr` is licensed under the GPL v3.  All of its code and documentation is
-Copyright 2010-2016, 2ndQuadrant Limited.  See the files COPYRIGHT and LICENSE for
+Copyright 2010-2017, 2ndQuadrant Limited.  See the files COPYRIGHT and LICENSE for
 details.
 
 The development of repmgr has primarily been sponsored by 2ndQuadrant customers.

COPYRIGHT

@@ -1,4 +1,4 @@
-Copyright (c) 2010-2016, 2ndQuadrant Limited
+Copyright (c) 2010-2017, 2ndQuadrant Limited
 All rights reserved.
 
 This program is free software: you can redistribute it and/or modify

HISTORY

@@ -1,3 +1,16 @@
+3.3.1   2017-03-
+        repmgrd: prevent invalid apply lag value being written to the
+          monitoring table (Ian)
+        repmgrd: fix error in XLogRecPtr conversion when calculating
+          monitoring statistics (Ian)
+        repmgr: if replication slots in use, where possible delete slot on old
+          upstream node after following new upstream (Ian)
+        repmgr: improve logging of rsync actions (Ian)
+        repmgr: improve `standby clone` when synchronous replication in use (Ian)
+        repmgr: stricter checking of allowed node id values
+        repmgr: enable `master register --force` when there is a foreign key
+          dependency from a standby node (Ian)
+
 3.3     2016-12-27
         repmgr: always log to STDERR even if log facility defined (Ian)
         repmgr: add --log-to-file to log repmgr output to the defined

Makefile

@@ -1,6 +1,6 @@
 #
 # Makefile
-# Copyright (c) 2ndQuadrant, 2010-2016
+# Copyright (c) 2ndQuadrant, 2010-2017
 
 HEADERS = $(wildcard *.h)
 

README.md

@@ -7,7 +7,7 @@ replication capabilities with utilities to set up standby servers, monitor
 replication, and perform administrative tasks such as failover or switchover
 operations.
 
-The current `repmgr` version, 3.2, supports all PostgreSQL versions from
+The current `repmgr` version (3.3) supports all PostgreSQL versions from
 9.3 to 9.6.
 
 Overview
@@ -333,10 +333,11 @@ The following replication settings may need to be adjusted:
     archive_command = '/bin/true'
 
     # If cloning using rsync, or you have configured `pg_basebackup_options`
-    # in `repmgr.conf` to include the setting `--xlog-method=fetch`, *and*
+    # in `repmgr.conf` to include the setting `--xlog-method=fetch` (from
-    # you have not set `restore_command` in `repmgr.conf`to fetch WAL files
+    # PostgreSQL 10 `--wal-method=fetch`), *and* you have not set
-    # from another source such as Barman, you'll need to set `wal_keep_segments`
+    # `restore_command` in `repmgr.conf`to fetch WAL files from another
-    # to a high enough value to ensure that all WAL files generated while
+    # source such as Barman, you'll need to set `wal_keep_segments` to a
+    # high enough value to ensure that all WAL files generated while
     # the standby is being cloned are retained until the standby starts up.
 
     # wal_keep_segments = 5000
@@ -390,7 +391,8 @@ least the following parameters:
 
 - `cluster`: an arbitrary name for the replication cluster; this must be identical
      on all nodes
-- `node`: a unique integer identifying the node
+- `node`: a unique integer identifying the node; note this must be a positive
+     32 bit signed integer between 1 and 2147483647
 - `node_name`: a unique string identifying the node; we recommend a name
      specific to the server (e.g. 'server_1'); avoid names indicating the
      current replication role like 'master' or 'standby' as the server's
@@ -501,7 +503,8 @@ 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.
+replication connections to be available (`repmgr` will verify sufficient
+connections are available before attempting to clonse).
 
 To override this behaviour, in `repmgr.conf` set `pg_basebackup`'s
 `--xlog-method` parameter to `fetch`:
@@ -513,6 +516,9 @@ See the `pg_basebackup` documentation for details:
 
     https://www.postgresql.org/docs/current/static/app-pgbasebackup.html
 
+> *NOTE*: From PostgreSQL 10, `pg_basebackup`'s `--xlog-method` parameter
+> has been renamed to `--wal-method`.
+
 Make any adjustments to the standby's PostgreSQL configuration files now,
 then start the server.
 
@@ -1153,11 +1159,10 @@ Additionally the following `repmgrd` options must be set in `repmgr.conf`:
     promote_command='repmgr standby promote -f /etc/repmgr.conf --log-to-file'
     follow_command='repmgr standby follow -f /etc/repmgr.conf --log-to-file'
 
-Note that the `--log-to-file` option will cause `repmgr` output to be logged to
+Note that the `--log-to-file` option will cause `repmgr`'s output to be logged to
-the destination configured to receive  log output `repmgrd`.
+the destination configured to receive log output for `repmgrd`.
 See `repmgr.conf.sample` for further `repmgrd`-specific settings
 
-
 When `failover` is set to `automatic`, upon detecting failure of the current
 master, `repmgrd` will execute one of `promote_command` or `follow_command`,
 depending on whether the current server is becoming the new master or
@@ -1461,7 +1466,7 @@ In general `repmgr` can be upgraded as-is without any further action required,
 however feature releases may require the `repmgr` database to be upgraded.
 An SQL script will be provided - please check the release notes for details:
 
-* http://repmgr.org/release-notes-3.2.1.html#UPGRADING
+* http://repmgr.org/release-notes-3.3.html#UPGRADING
 
 
 Distribution-specific configuration

check_dir.c

@@ -1,6 +1,6 @@
 /*
  * check_dir.c - Directories management functions
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

check_dir.h

@@ -1,6 +1,6 @@
 /*
  * check_dir.h
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

compat.c

@@ -1,10 +1,12 @@
 /*
  *
  * compat.c
- *	  Provide backports of various functions not publicly
+ *	  Provides a couple of useful string utility functions adapted
- *    exposed before PostgreSQL 9.6
+ *	  from the backend code, which are not publicly exposed. They're
+ *	  unlikely to change but it would be worth keeping an eye on them
+ *	  for any fixes/improvements
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * Portions Copyright (c) 1996-2013, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
@@ -24,8 +26,6 @@
  *
  */
 
-#if (PG_VERSION_NUM < 90600)
-
 #include "repmgr.h"
 #include "compat.h"
 
@@ -34,8 +34,8 @@
  * the string as a value, in a keyword/pair value in a libpq connection
  * string
  *
- * This function is copied from src/bin/pg_dump/dumputils.c
+ * This function is adapted from src/fe_utils/string_utils.c (before 9.6
- * as it is only publicly exposed from 9.6
+ * located in: src/bin/pg_dump/dumputils.c)
  */
 void
 appendConnStrVal(PQExpBuffer buf, const char *str)
@@ -79,8 +79,6 @@ appendConnStrVal(PQExpBuffer buf, const char *str)
 
 /*
  * Adapted from: src/fe_utils/string_utils.c
- *
- * Function not publicly available before PostgreSQL 9.6.
  */
 void
 appendShellString(PQExpBuffer buf, const char *str)
@@ -107,5 +105,3 @@ appendShellString(PQExpBuffer buf, const char *str)
 	appendPQExpBufferChar(buf, '\'');
 }
 
-
-#endif

compat.h

@@ -1,6 +1,6 @@
 /*
  * compat.h
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

config.c

@@ -1,7 +1,7 @@
 /*
  * config.c - Functions to parse the config file
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -448,6 +448,10 @@ _parse_config(t_configuration_options *options, ItemList *error_list)
 	{
 		item_list_append(error_list, _("\"node\": must be greater than zero"));
 	}
+	else if (options->node < 0)
+	{
+		item_list_append(error_list, _("\"node\": must be a positive signed 32 bit integer, i.e. 2147483647 or less"));
+	}
 
 	if (strlen(options->conninfo))
 	{

config.h

@@ -1,7 +1,7 @@
 /*
  * config.h
  *
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

dbutils.c

@@ -1,7 +1,7 @@
 /*
  * dbutils.c - Database connection/management functions
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -33,6 +33,15 @@ char repmgr_schema[MAXLEN] = "";
 char repmgr_schema_quoted[MAXLEN] = "";
 
 static int _get_node_record(PGconn *conn, char *cluster, char *sqlquery, t_node_info *node_info);
+static bool _set_config(PGconn *conn, const char *config_param, const char *sqlquery);
+
+/*
+ * _establish_db_connection()
+ *
+ * Connect to a database using a conninfo string.
+ *
+ * NOTE: *do not* use this for replication connections; use establish_db_connection_by_params() instead.
+ */
 
 PGconn *
 _establish_db_connection(const char *conninfo, const bool exit_on_error, const bool log_notice, const bool verbose_only)
@@ -77,6 +86,19 @@ _establish_db_connection(const char *conninfo, const bool exit_on_error, const b
 		}
 	}
 
+	/*
+	 * set "synchronous_commit" to "local" in case synchronous replication is in use
+	 */
+
+	else if (set_config(conn, "synchronous_commit", "local") == false)
+	{
+		if (exit_on_error)
+		{
+			PQfinish(conn);
+			exit(ERR_DB_CON);
+		}
+	}
+
 	return conn;
 }
 
@@ -116,8 +138,12 @@ PGconn *
 establish_db_connection_by_params(const char *keywords[], const char *values[],
 								  const bool exit_on_error)
 {
-	/* Make a connection to the database */
+	PGconn	   *conn;
-	PGconn	   *conn = PQconnectdbParams(keywords, values, true);
+	bool	    replication_connection = false;
+	int	   	    i;
+
+	/* Connect to the database using the provided parameters */
+	conn = PQconnectdbParams(keywords, values, true);
 
 	/* Check to see that the backend connection was successfully made */
 	if ((PQstatus(conn) != CONNECTION_OK))
@@ -130,6 +156,28 @@ establish_db_connection_by_params(const char *keywords[], const char *values[],
 			exit(ERR_DB_CON);
 		}
 	}
+	else
+	{
+		/*
+		 * set "synchronous_commit" to "local" in case synchronous replication is in
+		 * use (provided this is not a replication connection)
+		 */
+
+		for (i = 0; keywords[i]; i++)
+		{
+			if (strcmp(keywords[i], "replication") == 0)
+				replication_connection = true;
+		}
+
+		if (replication_connection == false && set_config(conn, "synchronous_commit", "local") == false)
+		{
+			if (exit_on_error)
+			{
+				PQfinish(conn);
+				exit(ERR_DB_CON);
+			}
+		}
+	}
 
 	return conn;
 }
@@ -1150,19 +1198,12 @@ stop_backup(PGconn *conn, char *last_wal_segment)
 }
 
 
+
 bool
-set_config_bool(PGconn *conn, const char *config_param, bool state)
+_set_config(PGconn *conn, const char *config_param, const char *sqlquery)
 {
-	char		sqlquery[QUERY_STR_LEN];
 	PGresult   *res;
 
-	sqlquery_snprintf(sqlquery,
-					  "SET %s TO %s",
-					  config_param,
-					  state ? "TRUE" : "FALSE");
-
-	log_verbose(LOG_DEBUG, "set_config_bool():\n%s\n", sqlquery);
-
 	res = PQexec(conn, sqlquery);
 
 	if (PQresultStatus(res) != PGRES_COMMAND_OK)
@@ -1177,6 +1218,36 @@ set_config_bool(PGconn *conn, const char *config_param, bool state)
 	return true;
 }
 
+bool
+set_config(PGconn *conn, const char *config_param,  const char *config_value)
+{
+	char		sqlquery[QUERY_STR_LEN];
+
+	sqlquery_snprintf(sqlquery,
+					  "SET %s TO '%s'",
+					  config_param,
+					  config_value);
+
+	log_verbose(LOG_DEBUG, "set_config():\n%s\n", sqlquery);
+
+	return _set_config(conn, config_param, sqlquery);
+}
+
+bool
+set_config_bool(PGconn *conn, const char *config_param, bool state)
+{
+	char		sqlquery[QUERY_STR_LEN];
+
+	sqlquery_snprintf(sqlquery,
+					  "SET %s TO %s",
+					  config_param,
+					  state ? "TRUE" : "FALSE");
+
+	log_verbose(LOG_DEBUG, "set_config_bool():\n%s\n", sqlquery);
+
+	return _set_config(conn, config_param, sqlquery);
+}
+
 
 /*
  * witness_copy_node_records()
@@ -1872,7 +1943,16 @@ _get_node_record(PGconn *conn, char *cluster, char *sqlquery, t_node_info *node_
 
 	node_info->node_id = atoi(PQgetvalue(res, 0, 0));
 	node_info->type = parse_node_type(PQgetvalue(res, 0, 1));
-	node_info->upstream_node_id = atoi(PQgetvalue(res, 0, 2));
+
+	if (PQgetisnull(res, 0, 2))
+	{
+		node_info->upstream_node_id = NO_UPSTREAM_NODE;
+	}
+	else
+	{
+		node_info->upstream_node_id = atoi(PQgetvalue(res, 0, 2));
+	}
+
 	strncpy(node_info->name, PQgetvalue(res, 0, 3), MAXLEN);
 	strncpy(node_info->conninfo_str, PQgetvalue(res, 0, 4), MAXLEN);
 	strncpy(node_info->slot_name, PQgetvalue(res, 0, 5), MAXLEN);

dbutils.h

@@ -1,7 +1,7 @@
 /*
  * dbutils.h
  *
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -124,6 +124,7 @@ int			get_slot_record(PGconn *conn, char *slot_name, t_replication_slot *record)
 bool		drop_replication_slot(PGconn *conn, char *slot_name);
 bool		start_backup(PGconn *conn, char *first_wal_segment, bool fast_checkpoint);
 bool		stop_backup(PGconn *conn, char *last_wal_segment);
+bool		set_config(PGconn *conn, const char *config_param,  const char *config_value);
 bool		set_config_bool(PGconn *conn, const char *config_param, bool state);
 bool		witness_copy_node_records(PGconn *masterconn, PGconn *witnessconn, char *cluster_name);
 bool		create_node_record(PGconn *conn, char *action, int node, char *type, int upstream_node, char *cluster_name, char *node_name, char *conninfo, int priority, char *slot_name, bool active);

dirmod.c

@@ -3,7 +3,7 @@
  * dirmod.c
  *	  directory handling functions
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * Portions Copyright (c) 1996-2013, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California

dirmod.h

@@ -1,6 +1,6 @@
 /*
  * dirmod.h
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

errcode.h

@@ -1,6 +1,6 @@
 /*
  * errcode.h
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

log.c

@@ -1,6 +1,6 @@
 /*
  * log.c - Logging methods
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This module is a set of methods for logging (currently only syslog)
  *

log.h

@@ -1,6 +1,6 @@
 /*
  * log.h
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

repmgr.c

@@ -1,7 +1,7 @@
 /*
  * repmgr.c - Command interpreter for the repmgr package
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This module is a command-line utility to easily setup a cluster of
  * hot standby servers for an HA environment
@@ -62,13 +62,7 @@
 #include "check_dir.h"
 #include "strutil.h"
 #include "version.h"
-
-#if (PG_VERSION_NUM < 90600)
 #include "compat.h"
-#else
-#include "fe_utils/string_utils.h"
-#include "postgres_fe.h"
-#endif
 
 #ifndef RECOVERY_COMMAND_FILE
 #define RECOVERY_COMMAND_FILE "recovery.conf"
@@ -172,6 +166,8 @@ static void config_file_list_add(t_configfile_list *list, const char *file, cons
 static void matrix_set_node_status(t_node_matrix_rec **matrix_rec_list, int n, int node_id, int connection_node_id, int connection_status);
 static void cube_set_node_status(t_node_status_cube **cube, int n, int node_id, int matrix_node_id, int connection_node_id, int connection_status);
 
+static void drop_replication_slot_if_exists(PGconn *conn, int node_id, char *slot_name);
+
 /* Global variables */
 static PQconninfoOption *opts = NULL;
 
@@ -762,7 +758,7 @@ main(int argc, char **argv)
 	{
 		if (optind < argc)
 		{
-			if (runtime_options.host[0])
+			if (runtime_options.host_param_provided == true)
 			{
 				PQExpBufferData additional_host_arg;
 				initPQExpBuffer(&additional_host_arg);
@@ -946,7 +942,7 @@ main(int argc, char **argv)
 	 */
 	if (strcmp(repmgr_schema, DEFAULT_REPMGR_SCHEMA_PREFIX) == 0 && !runtime_options.force)
 	{
-		log_err(_("unable to determine cluster name - please provide a valid configuration file with -c/--config-file\n"));
+		log_err(_("unable to determine cluster name - please provide a valid configuration file with -f/--config-file\n"));
 		log_hint(_("Use -F/--force to continue anyway\n"));
 		exit(ERR_BAD_CONFIG);
 	}
@@ -1877,6 +1873,7 @@ do_master_register(void)
 	int		    primary_node_id = UNKNOWN_NODE_ID;
 
 	bool		record_created;
+	t_node_info node_info = T_NODE_INFO_INITIALIZER;
 
 	conn = establish_db_connection(options.conninfo, true);
 
@@ -1949,42 +1946,51 @@ do_master_register(void)
 		exit(ERR_BAD_CONFIG);
 	}
 
-	/* Delete any existing record for this node if --force set */
+	begin_transaction(conn);
-	if (runtime_options.force)
+
+	/*
+	 * Check whether there's an existing record for this node, and
+	 * update it if --force set
+	 */
+	if (get_node_record(conn, options.cluster_name, options.node, &node_info))
 	{
-		bool node_record_deleted;
+		if (!runtime_options.force)
-		t_node_info node_info = T_NODE_INFO_INITIALIZER;
-
-		begin_transaction(conn);
-
-		if (get_node_record(conn, options.cluster_name, options.node, &node_info))
 		{
-			log_notice(_("deleting existing master record with id %i\n"), options.node);
+			log_err(_("this node is already registered\n"));
-
+			log_hint(_("use -F/--force to overwrite the existing node record\n"));
-			node_record_deleted = delete_node_record(conn,
+			rollback_transaction(conn);
-													 options.node,
+			PQfinish(conn);
-													 "master register");
+			exit(ERR_BAD_CONFIG);
-			if (node_record_deleted == false)
-			{
-				rollback_transaction(conn);
-				PQfinish(conn);
-				exit(ERR_BAD_CONFIG);
-			}
 		}
-	}
 
-	/* Now register the master */
+		record_created = update_node_record(conn,
-	record_created = create_node_record(conn,
+											"master register",
-										"master register",
+											options.node,
-										options.node,
+											"master",
-										"master",
+											NO_UPSTREAM_NODE,
-										NO_UPSTREAM_NODE,
+											options.cluster_name,
-										options.cluster_name,
+											options.node_name,
-										options.node_name,
+											options.conninfo,
-										options.conninfo,
+											options.priority,
-										options.priority,
+											repmgr_slot_name_ptr,
-										repmgr_slot_name_ptr,
+											true);
-										true);
+
+	}
+	else
+	{
+		/* Now register the master */
+		record_created = create_node_record(conn,
+											"master register",
+											options.node,
+											"master",
+											NO_UPSTREAM_NODE,
+											options.cluster_name,
+											options.node_name,
+											options.conninfo,
+											options.priority,
+											repmgr_slot_name_ptr,
+											true);
+	}
 
 	if (record_created == false)
 	{
@@ -2846,11 +2852,18 @@ do_standby_clone(void)
 		maxlen_snprintf(local_repmgr_directory, "%s/repmgr",   local_data_directory  );
 		maxlen_snprintf(datadir_list_filename,  "%s/data.txt", local_repmgr_directory);
 
+		if (!create_pg_dir(local_data_directory, runtime_options.force))
+		{
+			log_err(_("unable to use directory %s ...\n"),
+					local_data_directory);
+			log_hint(_("use -F/--force option to force this directory to be overwritten\n"));
+			exit(ERR_BAD_CONFIG);
+		}
+
 		if (!create_pg_dir(local_repmgr_directory, runtime_options.force))
 		{
-			log_err(_("unable to use directory \"%s\" ...\n"),
+			log_err(_("unable to create directory \"%s\" ...\n"),
 					local_repmgr_directory);
-			log_hint(_("use -F/--force option to force this directory to be overwritten\n"));
 
 			exit(ERR_BAD_CONFIG);
 		}
@@ -2958,8 +2971,6 @@ do_standby_clone(void)
 				primary_conn = source_conn;
 			}
 
-
-
 			/*
 			 * Sanity-check that the master node has a repmgr schema - if not
 			 * present, fail with an error (unless -F/--force is used)
@@ -2979,7 +2990,6 @@ do_standby_clone(void)
 				log_warning(_("expected repmgr schema '%s' not found on master server\n"), get_repmgr_schema());
 			}
 
-
 			/* Fetch the source's data directory */
 			if (get_pg_setting(source_conn, "data_directory", master_data_directory) == false)
 			{
@@ -3001,8 +3011,6 @@ do_standby_clone(void)
 				log_hint(_("use -D/--data-dir to explicitly specify a data directory\n"));
 			}
 
-
-
 			/*
 			 * Copy the source connection so that we have some default values,
 			 * particularly stuff like passwords extracted from PGPASSFILE;
@@ -3587,6 +3595,9 @@ do_standby_clone(void)
 			/*
 			 * We must create some PGDATA subdirectories because they are
 			 * not included in the Barman backup.
+			 *
+			 * See class RsyncBackupExecutor in the Barman source (barman/backup_executor.py)
+			 * for a definitive list of excluded directories.
 			 */
 			{
 				const char* const dirs[] = {
@@ -3597,14 +3608,14 @@ do_standby_clone(void)
 					/* Only from 9.4 */
 					"pg_dynshmem", "pg_logical", "pg_logical/snapshots", "pg_logical/mappings", "pg_replslot",
 					/* Already in 9.3 */
-					"pg_serial", "pg_snapshots", "pg_stat", "pg_stat_tmp", "pg_tblspc",
+					"pg_notify", "pg_serial", "pg_snapshots", "pg_stat", "pg_stat_tmp", "pg_tblspc",
 					"pg_twophase", "pg_xlog", 0
 				};
 				const int vers[] = {
 					100000,
 					90500,
 					90400, 90400, 90400, 90400, 90400,
-					0, 0, 0, 0, 0,
+					0, 0, 0, 0, 0, 0,
 					0, -100000, 0
 				};
 				for (i = 0; dirs[i]; i++)
@@ -3614,7 +3625,7 @@ do_standby_clone(void)
 						continue;
 
 					/* directory existed in earlier versions than this server but has been removed/renamed - skip */
-					if (vers[1] < 0 && server_version_num >= abs(vers[i]))
+					if (vers[i] < 0 && server_version_num >= abs(vers[i]))
 						continue;
 
 					maxlen_snprintf(filename, "%s/%s", local_data_directory, dirs[i]);
@@ -3634,17 +3645,6 @@ do_standby_clone(void)
 				initPQExpBuffer(&tablespace_map);
 			}
 
-			/*
-			 * From 9.1 default is to wait for a sync standby to ack, avoid that by
-			 * turning off sync rep for this session
-			 */
-			if (set_config_bool(source_conn, "synchronous_commit", false) == false)
-			{
-				r = ERR_BAD_CONFIG;
-				retval = ERR_BAD_CONFIG;
-				goto stop_backup;
-			}
-
 			if (start_backup(source_conn, first_wal_segment, runtime_options.fast_checkpoint) == false)
 			{
 				r = ERR_BAD_BASEBACKUP;
@@ -3674,15 +3674,16 @@ do_standby_clone(void)
 								  master_data_directory, local_data_directory,
 								  true, server_version_num);
 			/*
-			  Exit code 0 means no error, but we want to ignore exit code 24 as well
+			 * Exit code 0 means no error, but we want to ignore exit code 24 as well
-			  as rsync returns that code on "Partial transfer due to vanished source files".
+			 * as rsync returns that code on "Partial transfer due to vanished source files".
-			  It's quite common for this to happen on the data directory, particularly
+			 * It's quite common for this to happen on the data directory, particularly
-			  with long running rsync on a busy server.
+			 * with long running rsync on a busy server.
-			*/
+			 */
-			if (!WIFEXITED(r) && WEXITSTATUS(r) != 24)
+			if (WIFEXITED(r) && WEXITSTATUS(r) && WEXITSTATUS(r) != 24)
 			{
-				log_warning(_("standby clone: failed copying master data directory '%s'\n"),
+				log_err(_("standby clone: failed copying master data directory '%s'\n"),
 							master_data_directory);
+				r = retval = ERR_BAD_RSYNC;
 				goto stop_backup;
 			}
 
@@ -3766,15 +3767,16 @@ do_standby_clone(void)
 									  true, server_version_num);
 
 				/*
-				  Exit code 0 means no error, but we want to ignore exit code 24 as well
+				 * Exit code 0 means no error, but we want to ignore exit code 24 as well
-				  as rsync returns that code on "Partial transfer due to vanished source files".
+				 * as rsync returns that code on "Partial transfer due to vanished source files".
-				  It's quite common for this to happen on the data directory, particularly
+				 * It's quite common for this to happen on the data directory, particularly
-				  with long running rsync on a busy server.
+				 * with long running rsync on a busy server.
-				*/
+				 */
-				if (!WIFEXITED(r) && WEXITSTATUS(r) != 24)
+				if (WIFEXITED(r) && WEXITSTATUS(r) && WEXITSTATUS(r) != 24)
 				{
-					log_warning(_("standby clone: failed copying tablespace directory '%s'\n"),
+					log_err(_("standby clone: failed copying tablespace directory '%s'\n"),
-					            cell_t->location);
+							cell_t->location);
+					r = retval = ERR_BAD_RSYNC;
 					goto stop_backup;
 				}
 			}
@@ -3782,7 +3784,7 @@ do_standby_clone(void)
 			/*
 			 * If a valid mapping was provide for this tablespace, arrange for it to
 			 * be remapped
-			 * (if no tablespace mappings was provided, the link will be copied as-is
+			 * (if no tablespace mapping was provided, the link will be copied as-is
 			 * by pg_basebackup or rsync and no action is required)
 			 */
 			if (mapping_found == true || mode == barman)
@@ -3941,7 +3943,7 @@ do_standby_clone(void)
 
 				r = copy_remote_files(runtime_options.host, runtime_options.remote_user,
 									  file->filepath, dest_path, false, server_version_num);
-				if (r != 0)
+				if (WEXITSTATUS(r))
 				{
 					log_err(_("standby clone: unable to copy config file '%s'\n"),
 							file->filename);
@@ -3975,7 +3977,7 @@ do_standby_clone(void)
 		r = copy_remote_files(runtime_options.host, runtime_options.remote_user,
 							  master_control_file, local_control_file,
 							  false, server_version_num);
-		if (r != 0)
+		if (WEXITSTATUS(r))
 		{
 			log_warning(_("standby clone: failed copying master control file '%s'\n"),
 						master_control_file);
@@ -4533,7 +4535,8 @@ do_standby_follow(void)
 	char		script[MAXLEN];
 	char		master_conninfo[MAXLEN];
 	PGconn	   *master_conn;
-	int			master_id = 0;
+	int			master_id = UNKNOWN_NODE_ID,
+				original_upstream_node_id = UNKNOWN_NODE_ID;
 
 	int			r,
 				retval;
@@ -4680,7 +4683,11 @@ do_standby_follow(void)
 	}
 
 
-	/* Fetch our node record so we can write application_name, if set */
+	/*
+	 * Fetch our node record so we can write application_name, if set,
+	 * and to get the upstream node ID, which we'll need to know if
+	 * replication slots are in use and we want to delete the old slot.
+	 */
 	query_result = get_node_record(master_conn,
 								   options.cluster_name,
 								   options.node,
@@ -4713,6 +4720,15 @@ do_standby_follow(void)
 			if (application_name != NULL && strlen(application_name))
 				param_set(&recovery_conninfo, "application_name", application_name);
 		}
+
+		if (local_node_record.upstream_node_id != UNKNOWN_NODE_ID)
+		{
+			original_upstream_node_id = local_node_record.upstream_node_id;
+		}
+		else
+		{
+			original_upstream_node_id = master_id;
+		}
 	}
 
 	log_info(_("changing standby's master to node %i\n"), master_id);
@@ -4742,6 +4758,54 @@ do_standby_follow(void)
 		exit(ERR_NO_RESTART);
 	}
 
+
+	/*
+	 * If replication slots are in use, and an inactive one for this node
+	 * (a former standby) exists on the former upstream, drop it.
+	 */
+
+	if (options.use_replication_slots && original_upstream_node_id != UNKNOWN_NODE_ID)
+	{
+		t_node_info upstream_node_record  = T_NODE_INFO_INITIALIZER;
+		int			upstream_query_result;
+
+		log_verbose(LOG_INFO, "attempting to remove replication slot from old upstream node %i\n",
+					original_upstream_node_id);
+
+		/* XXX should we poll for server restart? */
+		conn = establish_db_connection(options.conninfo, true);
+
+		upstream_query_result = get_node_record(conn,
+												options.cluster_name,
+												original_upstream_node_id,
+												&upstream_node_record);
+
+		PQfinish(conn);
+
+
+		if (upstream_query_result != 1)
+		{
+			log_warning(_("unable to retrieve node record for old upstream node %i"),
+						original_upstream_node_id);
+		}
+		else
+		{
+			conn = establish_db_connection_quiet(upstream_node_record.conninfo_str);
+			if (PQstatus(conn) != CONNECTION_OK)
+			{
+				log_info(_("unable to connect to old upstream node %i to remove replication slot\n"),
+						 original_upstream_node_id);
+				log_hint(_("if reusing this node, you should manually remove any inactive replication slots\n"));
+			}
+			else
+			{
+				drop_replication_slot_if_exists(conn,
+												original_upstream_node_id,
+												local_node_record.slot_name);
+			}
+		}
+	}
+
 	/*
 	 * It's possible this node was an inactive primary - update the
 	 * relevant fields to ensure it's marked as an active standby
@@ -4759,6 +4823,8 @@ do_standby_follow(void)
 		exit(ERR_BAD_CONFIG);
 	}
 
+	log_notice(_("STANDBY FOLLOW successful\n"));
+
 	create_event_record(master_conn,
 						&options,
 						options.node,
@@ -4794,8 +4860,6 @@ do_standby_follow(void)
  *
  * TODO:
  *  - make connection test timeouts/intervals configurable (see below)
- *  - add command line option --remote_pg_bindir or similar to
- *    optionally handle cases where the remote pg_bindir is different
  */
 
 static void
@@ -5529,6 +5593,8 @@ do_standby_switchover(void)
 
 	/* Check for entry in the new master's pg_stat_replication */
 
+	local_conn = establish_db_connection(options.conninfo, true);
+
 	{
 		int			i,
 			replication_check_timeout  = 60,
@@ -5538,8 +5604,6 @@ do_standby_switchover(void)
 
 		initPQExpBuffer(&event_details);
 
-		local_conn = establish_db_connection(options.conninfo, true);
-
 		i = 0;
 		for (;;)
 		{
@@ -5635,30 +5699,9 @@ do_standby_switchover(void)
 		}
 		else
 		{
-			t_replication_slot  slot_info;
+			drop_replication_slot_if_exists(remote_conn,
-			int					query_res;
+											remote_node_id,
-
+											local_node_record.slot_name);
-			query_res = get_slot_record(remote_conn, local_node_record.slot_name, &slot_info);
-
-			if (query_res)
-			{
-				if (slot_info.active == false)
-				{
-					if (drop_replication_slot(remote_conn, local_node_record.slot_name) == true)
-					{
-						log_notice(_("replication slot \"%s\" deleted on former master\n"), local_node_record.slot_name);
-					}
-					else
-					{
-						log_err(_("unable to delete replication slot \"%s\" on former master\n"), local_node_record.slot_name);
-					}
-				}
-				/* if active replication slot exists, call Houston as we have a problem */
-				else
-				{
-					log_err(_("replication slot \"%s\" is still active on former master\n"), local_node_record.slot_name);
-				}
-			}
 		}
 
 		PQfinish(remote_conn);
@@ -6168,7 +6211,7 @@ do_witness_create(void)
 
 	r = copy_remote_files(runtime_options.host, runtime_options.remote_user,
 						  master_hba_file, runtime_options.dest_dir, false, -1);
-	if (r != 0)
+	if (WEXITSTATUS(r))
 	{
 		char *errmsg = _("Unable to copy pg_hba.conf from master");
 		log_err("%s\n", errmsg);
@@ -6934,9 +6977,11 @@ copy_remote_files(char *host, char *remote_user, char *remote_path,
 
 	r = system(script);
 
-	if (r != 0)
+	log_debug("copy_remote_files(): r = %i; WIFEXITED: %i; WEXITSTATUS: %i\n", r, WIFEXITED(r), WEXITSTATUS(r));
-		log_err(_("unable to rsync from remote host (%s:%s)\n"),
+
-				host_string, remote_path);
+	/* exit code 24 indicates vanished files, which isn't a problem for us */
+	if (WEXITSTATUS(r) && WEXITSTATUS(r) != 24)
+		log_verbose(LOG_WARNING, "copy_remote_files(): rsync returned unexpected exit status %i \n", WEXITSTATUS(r));
 
 	return r;
 }
@@ -8760,3 +8805,33 @@ config_file_list_add(t_configfile_list *list, const char *file, const char *file
 
 	list->entries ++;
 }
+
+
+static void
+drop_replication_slot_if_exists(PGconn *conn, int node_id, char *slot_name)
+{
+	t_replication_slot  slot_info;
+	int					query_res;
+
+	query_res = get_slot_record(conn,slot_name, &slot_info);
+
+	if (query_res)
+	{
+		if (slot_info.active == false)
+		{
+			if (drop_replication_slot(conn, slot_name) == true)
+			{
+				log_notice(_("replication slot \"%s\" deleted on node %i\n"), slot_name, node_id);
+			}
+			else
+			{
+				log_err(_("unable to delete replication slot \"%s\" on node %i\n"), slot_name, node_id);
+			}
+		}
+		/* if active replication slot exists, call Houston as we have a problem */
+		else
+		{
+			log_err(_("replication slot \"%s\" is still active on node %i\n"), slot_name, node_id);
+		}
+	}
+}

repmgr.conf.sample

@@ -80,12 +80,6 @@
 #
 #logfile='/var/log/repmgr/repmgr.log'
 
-# By default only repmgrd log output will be written to a file,
-# if defined in "logfile"
-# enable this to restore old behaviour where output from the repmgr
-# client will be written to the logfile too
-#log_repmgr_to_file = 0
-
 # event notifications can be passed to an arbitrary external program
 # together with the following parameters:
 #
@@ -149,8 +143,15 @@
 # external command arguments. Values shown are examples.
 
 #pg_ctl_options='-s'
-#pg_basebackup_options='--xlog-method=s'
+#pg_basebackup_options='--label=repmgr_backup'
 
+# This is the host name of the barman server, which is used for connecting over
+# to the barman server (passwordless ssh keys should be in place)
+#barman_server='backup_server'
+# If you are placing the barman.conf file in a non-standard path, or using
+# a name other than barman.conf, use this parameter to specify the path and
+# name of the barman configuration file.
+#barman_config='/path/to/barman.conf'
 
 # Standby clone settings
 # ----------------------

repmgr.h

@@ -1,6 +1,6 @@
 /*
  * repmgr.h
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

repmgr.sql

@@ -1,7 +1,7 @@
 /*
  * repmgr.sql
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  */
 

repmgrd.c

@@ -1,7 +1,7 @@
 /*
  * repmgrd.c - Replication manager daemon
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This module connects to the nodes of a replication cluster and monitors
  * how far are they from master
@@ -1187,10 +1187,22 @@ standby_monitor(void)
 	PQclear(res);
 
 	lsn_master_current_xlog_location = lsn_to_xlogrecptr(last_wal_primary_location, NULL);
-	lsn_last_xlog_replay_location = lsn_to_xlogrecptr(last_xlog_replay_location, NULL);
 	lsn_last_xlog_receive_location = lsn_to_xlogrecptr(last_xlog_receive_location, NULL);
+	lsn_last_xlog_replay_location = lsn_to_xlogrecptr(last_xlog_replay_location, NULL);
+
+	if (lsn_last_xlog_receive_location >= lsn_last_xlog_replay_location)
+	{
+		apply_lag = (long long unsigned int)lsn_last_xlog_receive_location - lsn_last_xlog_replay_location;
+	}
+	else
+	{
+		/* This should never happen, but in case it does set apply lag to zero */
+		log_warning("Standby receive (%s) location appears less than standby replay location (%s)\n",
+					last_xlog_receive_location,
+					last_xlog_replay_location);
+		apply_lag = 0;
+	}
 
-	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)
@@ -1199,7 +1211,7 @@ standby_monitor(void)
 	}
 	else
 	{
-		/* This should never happen, but in case it does set lag to zero */
+		/* This should never happen, but in case it does set replication lag to zero */
 		log_warning("Master xlog (%s) location appears less than standby receive location (%s)\n",
 					last_wal_primary_location,
 					last_xlog_receive_location);
@@ -2327,13 +2339,13 @@ lsn_to_xlogrecptr(char *lsn, bool *format_ok)
 	if (format_ok != NULL)
 		*format_ok = true;
 
-	return (((XLogRecPtr) xlogid * 16 * 1024 * 1024 * 255) + xrecoff);
+	return (XLogRecPtr) ((uint64) xlogid) << 32 | (uint64) xrecoff;
 }
 
 void
 usage(void)
 {
-	log_err(_("%s: Replicator manager daemon \n"), progname());
+	log_err(_("%s: replication management daemon for PostgreSQL\n"), progname());
 	log_err(_("Try \"%s --help\" for more information.\n"), progname());
 }
 

sql/Makefile

@@ -1,7 +1,7 @@
 #
 # Makefile
 #
-# Copyright (c) 2ndQuadrant, 2010-2016
+# Copyright (c) 2ndQuadrant, 2010-2017
 #
 
 MODULE_big = repmgr_funcs

sql/repmgr_funcs.sql.in

@@ -1,6 +1,6 @@
 /*
  * repmgr_function.sql
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  */
 

sql/uninstall_repmgr_funcs.sql

@@ -1,6 +1,6 @@
 /*
  * uninstall_repmgr_funcs.sql
- * Copyright (c) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  */
 

strutil.c

@@ -1,7 +1,7 @@
 /*
  * strutil.c
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by

strutil.h

@@ -1,6 +1,6 @@
 /*
  * strutil.h
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  *
  * This program is free software: you can redistribute it and/or modify

uninstall_repmgr.sql

@@ -1,7 +1,7 @@
 /*
  * uninstall_repmgr.sql
  *
- * Copyright (C) 2ndQuadrant, 2010-2016
+ * Copyright (c) 2ndQuadrant, 2010-2017
  *
  */
 

version.h

@@ -1,6 +1,6 @@
 #ifndef _VERSION_H_
 #define _VERSION_H_
 
-#define REPMGR_VERSION "3.3"
+#define REPMGR_VERSION "3.3.1"
 
 #endif