Fix pg_rewind when rewinding new database with tables included

This fixes an issue introduced by 266b6ac, which has added filters to
exclude file patterns on the target and source data directories to
reduce the number of files transferred.  Filters get applied to both
the target and source data files, and include pg_internal.init which is
present for each database once relations are created on it.  However, if
the target differed from the source with at least one new database with
relations, the rewind would fail due to the exclusion filters applied on
the target files, causing pg_internal.init to still be present on the
target database folder, while its contents should have been completely
removed so as there is nothing remaining inside at the time of the
folder deletion.

Applying exclusion filters on the source files is fine, because this way
the amount of data copied from the source to the target is reduced.  And
actually, not applying the filters on the target is what pg_rewind
should do, because this causes such files to be automatically removed
during the rewind on the target.  Exclusion filters apply to paths which
are removed or recreated automatically at startup, so removing all those
files on the target during the rewind is a win.

The existing set of TAP tests already stresses the rewind of databases,
but it did not include any tables on those newly-created databases.
Creating extra tables in this case is enough to reproduce the failure,
so the existing tests are extended to close the gap.

Reported-by: Mithun Cy
Author: Michael Paquier
Discussion: https://postgr.es/m/CADq3xVYt6_pO7ZzmjOqPgY9HWsL=kLd-_tNyMtdfjKqEALDyTA@mail.gmail.com
Backpatch-through: 11

Author: michaelpq

src/bin/pg_rewind/filemap.c

@@ -147,7 +147,10 @@ process_source_file(const char *path, file_type_t type, size_t newsize,
 
 	Assert(map->array == NULL);
 
-	/* ignore any path matching the exclusion filters */
+	/*
+	 * Skip any files matching the exclusion filters. This has the effect to
+	 * remove all those files on the target.
+	 */
 	if (check_file_excluded(path, true))
 		return;
 
@@ -334,12 +337,10 @@ process_target_file(const char *path, file_type_t type, size_t oldsize,
 	file_entry_t *entry;
 
 	/*
-	 * Ignore any path matching the exclusion filters.  This is not actually
-	 * mandatory for target files, but this does not hurt and let's be
-	 * consistent with the source processing.
+	 * Do not apply any exclusion filters here.  This has advantage to remove
+	 * from the target data folder all paths which have been filtered out from
+	 * the source data folder when processing the source files.
 	 */
-	if (check_file_excluded(path, false))
-		return;
 
 	snprintf(localpath, sizeof(localpath), "%s/%s", datadir_target, path);
 	if (lstat(localpath, &statbuf) < 0)

src/bin/pg_rewind/t/002_databases.pl

@@ -15,19 +15,27 @@ sub run_test
 	RewindTest::setup_cluster($test_mode, ['-g']);
 	RewindTest::start_master();
 
-	# Create a database in master.
+	# Create a database in master with a table.
 	master_psql('CREATE DATABASE inmaster');
+	master_psql('CREATE TABLE inmaster_tab (a int)', 'inmaster');
 
 	RewindTest::create_standby($test_mode);
 
-	# Create another database, the creation is replicated to the standby
+	# Create another database with another table, the creation is
+	# replicated to the standby.
 	master_psql('CREATE DATABASE beforepromotion');
+	master_psql('CREATE TABLE beforepromotion_tab (a int)',
+		    'beforepromotion');
 
 	RewindTest::promote_standby();
 
 	# Create databases in the old master and the new promoted standby.
 	master_psql('CREATE DATABASE master_afterpromotion');
+	master_psql('CREATE TABLE master_promotion_tab (a int)',
+		    'master_afterpromotion');
 	standby_psql('CREATE DATABASE standby_afterpromotion');
+	standby_psql('CREATE TABLE standby_promotion_tab (a int)',
+		    'standby_afterpromotion');
 
 	# The clusters are now diverged.
 

src/bin/pg_rewind/t/RewindTest.pm

@@ -68,18 +68,20 @@ our $node_standby;
 sub master_psql
 {
 	my $cmd = shift;
+	my $dbname = shift || 'postgres';
 
 	system_or_bail 'psql', '-q', '--no-psqlrc', '-d',
-	  $node_master->connstr('postgres'), '-c', "$cmd";
+	  $node_master->connstr($dbname), '-c', "$cmd";
 	return;
 }
 
 sub standby_psql
 {
 	my $cmd = shift;
+	my $dbname = shift || 'postgres';
 
 	system_or_bail 'psql', '-q', '--no-psqlrc', '-d',
-	  $node_standby->connstr('postgres'), '-c', "$cmd";
+	  $node_standby->connstr($dbname), '-c', "$cmd";
 	return;
 }