On Wed, Jul 20, 2022 at 02:00:00PM +0900, Fujii Masao wrote: > I reported two trouble cases; they are the cases where BASE_BACKUP > is canceled and terminated, respectively. But you added the test > only for one of them. Is this intentional?
Nope. The one I have implemented was the fanciest case among the two, so I just focused on it. Adding an extra test to cover the second scenario is easier. So I have added one as of the attached, addressing your other comments while on it. I have also decided to add the tests at the bottom of 001_stream_rep.pl, as these are quicker than a node initialization. -- Michael
From 2aa841a3dfb643a14d28e6d595703f14e98ad919 Mon Sep 17 00:00:00 2001
From: Michael Paquier <mich...@paquier.xyz>
Date: Wed, 20 Jul 2022 15:48:27 +0900
Subject: [PATCH v2] Add more TAP tests with BASE_BACKUP
---
src/test/recovery/t/001_stream_rep.pl | 55 +++++++++++++++++++++++++++
1 file changed, 55 insertions(+)
diff --git a/src/test/recovery/t/001_stream_rep.pl b/src/test/recovery/t/001_stream_rep.pl
index 86864098f9..b15dd6b29a 100644
--- a/src/test/recovery/t/001_stream_rep.pl
+++ b/src/test/recovery/t/001_stream_rep.pl
@@ -531,4 +531,59 @@ my $primary_data = $node_primary->data_dir;
ok(!-f "$primary_data/pg_wal/$segment_removed",
"WAL segment $segment_removed recycled after physical slot advancing");
+note "testing pg_backup_start() followed by BASE_BACKUP";
+my $connstr = $node_primary->connstr('postgres') . " replication=database";
+
+# This test requires a replication connection with a database, as it mixes
+# a replication command and a SQL command.
+$node_primary->command_fails_like(
+ [
+ 'psql', '-c', "SELECT pg_backup_start('backup', true)",
+ '-c', 'BASE_BACKUP', '-d', $connstr
+ ],
+ qr/a backup is already in progress in this session/,
+ 'BASE_BACKUP cannot run in session already running backup');
+
+note "testing BASE_BACKUP cancellation";
+
+my $sigchld_bb_timeout =
+ IPC::Run::timer($PostgreSQL::Test::Utils::timeout_default);
+
+# This test requires a replication connection with a database, as it mixes
+# a replication command and a SQL command. The first BASE_BACKUP is throttled
+# to give enough room for the cancellation running below. The second command
+# for pg_backup_stop() should fail.
+my ($sigchld_bb_stdin, $sigchld_bb_stdout, $sigchld_bb_stderr) = ('', '', '');
+my $sigchld_bb = IPC::Run::start(
+ [
+ 'psql', '-X', '-c', "BASE_BACKUP (CHECKPOINT 'fast', MAX_RATE 32);",
+ '-c', 'SELECT pg_backup_stop()',
+ '-d', $connstr
+ ],
+ '<',
+ \$sigchld_bb_stdin,
+ '>',
+ \$sigchld_bb_stdout,
+ '2>',
+ \$sigchld_bb_stderr,
+ $sigchld_bb_timeout);
+
+# The cancellation is issued once the database files are streamed and
+# the checkpoint issued at backup start completes.
+is( $node_primary->poll_query_until(
+ 'postgres',
+ "SELECT pg_cancel_backend(a.pid) FROM "
+ . "pg_stat_activity a, pg_stat_progress_basebackup b WHERE "
+ . "a.pid = b.pid AND a.query ~ 'BASE_BACKUP' AND "
+ . "b.phase = 'streaming database files';"),
+ "1",
+ "WAL sender sending base backup killed");
+
+# The psql command should fail on pg_backup_stop().
+ok( pump_until(
+ $sigchld_bb, $sigchld_bb_timeout,
+ \$sigchld_bb_stderr, qr/backup is not in progress/),
+ 'base backup cleanly cancelled');
+$sigchld_bb->finish();
+
done_testing();
--
2.36.1
signature.asc
Description: PGP signature
